DP Tool Club 17/CD_ASCQ_17_101194.iso/vrac/vester_e.zip/TVINFO.TXT

Text File | 1994-08-01 | 113KB | 3,427 lines

          Please stop right now and make a backup copy of these programs

     on a floppy disk. And activate the Write Protection on that disk !!



                SSTV/FAX480/WEFAX SYSTEM for IBM and CLONES

                     Copyrights 1993, 1994 Ben Vester



          This system includes a receiving and transmitting program  for

     Amateur  Radio  Color Slow-Scan TV and FAX480 and  a  receive  only

     program for high resolution weatherfax. The so-called SSTV as being

     used  today is really nothing more than color facsimile, so all  of

     these systems are alike in that only one image is transmitted at  a

     time and they all use 1500 hz to 2300 hz as limits for the  modula-

     tion  audio. The system is basically software-based, with the  only

     hardware  required  being  a very simple  audio  clipper  interface

     between  the receiver and the computer's serial COM port. The  com-

     puter  facsimile  input and output to the  amateur  transceiver  is

     transformer isolated to avoid grounding and noise problems. Details

     of the interface are contained in Sept. 1991 "QST" and in Jan. 1994

     "QST".  There was an error in the schematic in the Jan. 1994  arti-

     cle, however. The Transmit audio comes out of the Serial  interface

     on  the  RTS pin, NOT the CTS pin. That is, the  Transmit  coupling

     capacitor  connects to pin 7 on a 9-pin connector, NOT pin  8.  RTS

     comes  out  on pin 4 on a 25-pin connector. An early draft  of  the

     1994 article is contained on this disk under filename,  ARTICLE.TXT

     and should be read for other needed information.



          Doing  all of the fax demodulation in the computer requires  a







     80286 or higher class of computer with higher clock speeds  helping

     the  quality of the images being transmitted and received.  All  of

     the  programs  except  the Wefax program could be  handled  in  the

     minimal  640 Kb. of RAM but to get any reasonable speed in  manipu-

     lating the images after reception a minimum of 1 megabyte of RAM is

     needed.  For  copying Wefax, 4 Mb. is needed for easy  copying  and

     manipulation. 2 Mb. is  almost enough to copy a full pix. A RAMDISK

     should  be created on your computer before trying to run  the  pro-

     grams. Just add a line to your CONFIG.SYS file as follows:



     For 1 Mb. system:    device=c:\dos\ramdrive.sys 368 512 64 /e



     For 4 Mb. system:    device=c:\dos\ramdrive.sys 3072 512 64/e

          If  you must coexist with Windows (uses 2 mb.) 3072  is  2048.

     With  less  memory, you can use batch files  (described  below)  to

     remove Windows from CONFIG.SYS and AUTOEXEC.BAT when you are  using

     SSTV.



     When you reboot the system take note of what letter designation the

     computer  assigns to the RAMDISK as you will need to tell  the  fax

     program  what that is. Now if you only have 640 Kb of RAM  you  can

     use  the hard disk "c" as the designated drive but it will be  very

     slow  in  manipulating  the color images and won't work at  all  on

     Wefax. Even a floppy drive can be designated but it will be ridicu-

     lously  slow.  High resolution color pix's with 3 bytes  per  pixel

     require  a lot of memory, so the old PC's with minimal memory  just

     won't hack it.









          Timing  is extremely critical in decoding the fax  modulation.

     We use the computer's TIME clock and since it is needed   full-time

     during  the actual copying, you may have to sacrifice  other  time-

     related  functions while using  the programs.  Also, since we  need

     both  the TIME clock and most of the memory for our  copying  task,

     the program may not work properly if your system is loaded up  with

     TSR resident programs, particularly ones that use the TIME timer or

     occupy  any  appreciable  memory . This is a  common  problem  with

     programs using SVGA graphics and I notice that my 8 year old grand-

     son  has already learned how to unload the resident  programs  from

     his  computer in order to get his favorite game programs  to  play.

     Some  resident programs may look like small memory users  but  they

     often  use the TIME clock continuosly, so they may  interfere  with

     the  timing functions in the program. This depends on how fast  the

     computer  is. If it's very fast, it may accommodate both  functions

     without interference. For example, with my 33 mhz 80386 computer, I

     can  accommodate  the excellent screen saver,  "BLANK-IT",  without

     interference.  With a slower computer, that might not be  possible.

     When I use my Computer Eyes/RT frame-grabber, however, I even  have

     to remove the screen saver. These resident programs can be  located

     in  your  AUTOEXEC.BAT file for your perusal. The  easiest  way  to

     remove them and add them back is with two batch files,  STARTTV.BAT

     and  STOPTV.BAT.  First, you copy the present AUTOEXEC.BAT  into  a

     AUTOEXEC.REG file. That is:

                         c:>copy autoexec.bat autoexec.reg



     Then you create an AUTOEXEC.TV file that only contains the absolute







     minimum (ie, NO resident programs). The STARTTV.BAT will consist of

     one line:              copy c:\autoexec.tv c:\autoexec.bat

     After  you  run  STARTTV.BAT, you reboot and are ready  to  go.  To

     restore  the  computer to normal, use STOPTV.BAT with  this  single

     line:           copy  c:\ autoexec.reg c:\autoexec.bat

     You,  of  course, could just type the single line  each  time,  but

     using the batch file minimizes your chances of losing your  present

     AUTOEXEC.BAT  file information. If you  have Windows installed  and

     are working with limited RAM, you can add the CONFIG.SYS changes to

     eliminate Windows temporarily to these same START/STOP batch files.

     Once  you  have the system working, you might want to  add  back  a

     screen  saver for use with TV and see if it messes up your  system.

     Another  DOS resident program that is extremely useful to  speed-up

     the  continuous file manipulations in this program is SMARTDRV,  so

     try to squeeze it in. SMARTDRV, unfortunately, does interfere  with

     the  two  big memory programs that copy data directly to  the  RAM-

     DRIVE,  namely, Wefax and Mode 96 ColorFAX, so that has to be  con-

     sidered..

          A  new source of troublesome "TSR" programs have shown  up  in

     some  of the newer 486 machines. These are actually "wired  in"  to

     the  system  (in ROM, I presume) and are turned ON or  OFF  in  the

     selection panel for CMOS. In particular, the transmit function  has

     been affected by some of these. While the receive program only uses

     the  TIME clock, the precision needed for the transmit can only  be

     obtained  using  the processing speed. We calibrate  the  processor

     speed  against the TIME clock and then use the processing speed  as

     the reference in the actual generation of the transmit output. This







     then  becomes susceptible to varying processor speeds due to  proc-

     essor  cache, etc., so  these functions need to be turned  off.  We

     have made a transmit program which only uses  the TIME clock but it

     has much more jitter than the mechanization we picked. Almost every

     function  that typically goes in as a "TSR" uses the TIME clock  to

     trigger it's periodic status check. Even the common EMM386 has been

     reported  as causing a problem. Once you eliminate  these  "jitter"

     sources   by  turning  them off, however, you  will  have  transmit

     stability  that will match anything available.  See  Experimenter's

     Info below for further details on more obscure transmit problems.

          The  programs  were developed for use  by  the  experimentally

     inclined radio amateur. It uses GWBASIC to interface with the user,

     so you will need a copy of GWBASIC.EXE (note:some have tried  using

     GWBASIC  that was used on Tandy machines and it did not work.  I've

     used  versions 3.22 and  3.23, both created for true IBM  clones.).

     The guts of the program are contained in machine language  programs

     which are called up by the BASIC program as needed. To  accommodate

     the  wide  variety of modes, the system parameters are  changed  by

     inserting  POKE's in the BASIC program, a different set  of  POKE's

     being  used  for each mode. This allows experimenters with  only  a

     modest  amount of skill in BASIC to create new modes, or  adapt  to

     additional  modes  that others have created. Even if  you  have  no

     previous  experience with BASIC, I recommend you LIST out the  pro-

     grams  and  study them to find how easy it is to make mods.  I  have

     even included some motivations--like getting your own call sign  on

     the pix you transmit, instead of my call sign--to encourage you  to

     learn  a  little  about hexidecimal numbers, about  2  color  image

     formation, and about changing the program.







          To call up the frequently used programs directly from the  DOS

     line you can use simple batch files. For example, to LOAD in GWBAS-

     IC, then LOAD in VU.BAS, and then RUN the program, you can create a

     batch file, VU.BAT, with a single line:

                         gwbasic.exe vu.bas

     Now  you  just  type VU on the DOS line to access  the  program.  A

     number of these batch programs are included on this disk.

          Each  of  the programs has a fairly good description  of  it's

     operation on the opening screen, so you should read both that  data

     and the data below to fill out your understanding. Also each  BASIC

     program is supported by one or several machine language files.  You

     can identify these by looking through the BASIC listing to find the

     BLOAD  command   followed by the name of the .ASM file.  These  ASM

     files must be in the same directory as the BASIC program so it  can

     find them. I recommend you make a separate directory (ie; c:>MD  TV

     and then c:>CD\TV) and copy all the files into that directory.

          In the different programs described below, you are often asked

     to  hit certain keys. The program assumes you are keying the  lower

     case  letter,  so make sure that you don't have the CAPS  LOCK  on!

     They are capitalized in the writeup strictly for  emphasis.



     VU.BAS

          You need to find out how to see pictures on the screen  before

     you can do anything, so we'll start with the viewing program. First

     activate  BASIC;  ie, c:\tv>gwbasic.exe  and then  when  the  BASIC

     screen  comes  up  LOAD"VU.  You actually  hit  F3  and  then  type

     VU---right here is where you can go astray and why you need  backup







     copies.  If  you  inadvertently  hit F4  instead  of  F3  you  will

     SAVE"VU---which  means you will save a blank sheet under the  title

     VU.BAS, writing right over the real VU.BAS program---another reason

     it  helps to use batch programs for normal use. Now hit F2  to  RUN

     the  program.  Read the first screen and then select the  mode  and

     give the file name of the sample picture. Note that the mode suffix

     to  the  filename is added by the program. If you add it  too,  the

     computer  won't be able to find a file with a double suffix, so  it

     will just return you to BASIC (you'll get Ok on the screen when you

     return to BASIC) when you answer all the questions. The program  is

     still  loaded, so you only need to hit F2 again and try  to  follow

     the instructions this time.

          The Video Card is the next selection to be made. If your  card

     doesn't  appear in the first 6 choices, then you'll make  choice  7

     and go find yourself a VESA driver for your card. If it's a  recent

     card  it  may  even meet the VESA standard without  a  driver.  The

     drivers are a product of the board manufacturer and you may already

     have  it  on a Utility disk that came with the card.  I  understand

     that  BBS: [phone] has a collection of drivers if  you  can't

     get  one from the manufacturer. The driver gets loaded  before  the

     program, so you can add it as a first line in the VU.BAT or  RT.BAT

     or AUTOEXEC.TV batch file for convenient operation. Be sure to  try

     all  7 choices before looking for a driver since card  names  don't

     always  reflect  the name of the SVGA microchip used in  it.  Also,

     there are different microchips in the same name boards, so  there's

     no  guarantee  the  most common one I used for  this  program  will

     necessarily  have  the  same software commands.  Some  recent  SVGA

     programs  even include a collection of drivers to select from.  The







     card type is often shown in the first message when you boot up  the

     computer or you can go to the DOS directory and type:

                       C:\DOS>debug                  then type in a dump

                       -dC000:0000

     You may need to hit d again to dump the whole message on the right,

     but  this  will often tell you the origin of the  video  card.  Try

     dumping at D000:0000 and E000:0000 if C000:0000 doesn't work.

          A  more direct way to find out the card parameters is to  call

     the  computer  manufacturer's software assist line. The  card  must

     support  640x480 with 256 colors (requires 512K video  memory)  al-

     though  a  limited  capability is possible  with  640x400x256  (see

     Experimenter's   Info  below). Pix MODES.M1 exercises  the  640x480

     capability while SAMPLE2.94 only needs 640x400 so you can use these

     to  explore  your card's parameters. Trident cards with  1  Mb.  of

     video memory are available mail-order for less than $50 so you  can

     easily  upgrade. To use the full capability of the  Trucolor  (16.7

     million  colors) capability now incorporated into this system,  you

     should go ahead and upgrade to a Trucolor. At revision  D time  you

     can  get  640x480x16.7 million colors in a  Diamond  Speedstar  PRO

     ($100) or Cirrus Logic ($70) boards which are fully VESA compatible

     and available in both VLB and  ISA (to fit the older systems).

          A  couple of people have reported trouble getting  their  VESA

     cards  or drivers to paint a full picture. I ran into this  problem

     myself with a recently acquired Diamond Speedstar Pro which used  a

     Cirrus  Logic video processor chip. We have added a  Bank-Switching

     parameter,  BS, to  the SYSTEM CONFIGURATION which   gives  another

     dimension of control. If the picture paints only part way down  the







     screen,  increase BS. If it goes all the way to the bottom  of  the

     screen and starts re-painting, decrease BS.

          When you select the wrong video card, the program will  faith-

     fully  load the picture into the video buffer but no  picture  will

     result.  After  a short wait for the picture loading,  hit  Q  (for

     Quit) and  you should get the BASIC Ok  displayed. Hit F2 to re-RUN

     the program and try the next video card selection. While the  video

     cards  in my two computers are well behaved and don't hang  up  the

     computer  when sent the wrong choice, this may not be true for  all

     cards.

          Once  you get a picture, play with the scroll commands to  see

     what  happens. Moving up and/or left initially will move  the  file

     pointer outside the file and give peculiar results, so move down or

     right  first.  The  arrows work for slow scroll and R  and  L  give

     larger scroll increments (in the special case of the Robot 36 mode,

     the  R stands for Reverse and will reverse the blue and red  colors

     if  you happen to start copying on the wrong line--it  actually  is

     scrolling  but  the  scroll increment is chosen to  be  one  line).

     Scroll through the color frames with R and see what  an out-of-sync

     picture will look like. You can scroll it right back in  sync.  Now

     try the color registration keys, C and X. to see what they do. When

     you  finish, go on to the next program, or maybe at this point  you

     want  to  change the Default choice for the video card to  be  your

     card.  LOAD the program and LIST 2000-3000  to see the SYSTEM  CON-

     FIGURATION  list.  The  first item sets the video  card  choice  to

     either be automatically set to whatever you choose in the next line

     (2030)  or  you can leave it in manual so you answer  the  question

     about  the  card every time you view a picture. The  BR  brightness







     parameter  is like the Color level control on a TV, and it and  the

     TU tuning parameter are only applicable to the Robot modes. Setting

     TU to 1 will give you retune capability in 10 cps increments  which

     allows best resolution in setting the color balance (or hue)  using

     the  F and G keys (note that the F stands for Frequency tuning  and

     the  G stands for Greener). The next  item is the picture  trimming

     parameter,  PT, which you may want to change occasionally.  If  you

     have  a  pix with trash still showing on the right  edge,  you  can

     raise  PT to trim more off the edge. You can change  the  parameter

     and  RUN the program without SAVEing it or you can make the  change

     permanent by SAVEing it. The final item, EX,  allows you to  expand

     the  pictures that have less bytes per line, such as Martin  2  and

     Scottie  2, to fill the screen. This aggravates any copying  anoma-

     lies but gives the more correct perspective for these modes.  Since

     these  modes send only about 1/2 the bytes and use 1/2 the time  of

     Scottie 1 and Martin 1, their fidelity is inherently worse, partic-

     ularly in less than optimum conditions.

          Note  that the card parameters you have determined for  VU  as

     described  above  will be transferred to most of the  programs  de-

     scribed below.



     RT.BAS

          This  will be  the most used program, since you will  spend  a

     lot   more  time receiving than anything else. To view  the  SYSTEM

     CONFIGURATION for the program activate BASIC as described above and

     LOAD"RT.  Now LIST (hit F1) 3000-3500. Move the cursor up  to  line

     3030  where  the video card selection is made and type  in  the  VC







     selection  you found worked in VU.BAS. Hit ENTER and then  move  the

     cursor back to the bottom of the screen.

          The RAMDISK designation that is stated on the screen when  you

     boot the system will be inserted in line 3050. Most likely it  will

     be "D", unless you have extra disk drives. The COM port you  select

     will be noted in line 3170. Set TC=0 in line 3250 and RE=0  in line

     3280  if you only have a 256 color board. After making  the  selec-

     tions,  move the cursor into the clear and LIST 3000-3500  to  make

     sure you got what you wanted. Note that when you change anything in

     a  line on the screen, you must hit ENTER to actually change it  in

     the  program. Also be sure to move the cursor in the  clear  before

     hitting  any of the Function keys. Now SAVE the program as  RT.BAS.

     Some of the other CONFIGURATION items may need to be changed later.

          You  are  now ready to check out the interface box  with   the

     program.  Plug in the interface to the selected COM port  and  feed

     audio  from the receiver into it--noise will be OK. Hit F2  to  run

     the program and select mode 6 and ENTER. Normally the program  will

     wait  until it receives a header before it begins to copy.  But  if

     you miss the header, you can bypass that step by hitting  SPACEBAR.

     In this case, with noise or voice audio, you'll need to bypass  the

     header start, so hit SPACEBAR. If you get a tone with a tick super-

     imposed  on  top  of it, you know that the system  is  copying  the

     incoming signal. If nothing happens or you get a steady tone  with-

     out any tick, the fax audio isn't getting into the serial port. You

     could have plugged into  the wrong COM port (surprizing how confus-

     ing COM port numbering is!) or the audio  may not be getting to the

     interface.  You'll  have to reboot the computer if  you  can't  get

     audio  in to start the copying.  (A few people have gotten to  this







     point and got a single interrupted tone and the computer hangs  up.

     This  occurs  when  you try to use a Tandy GWBASIC.)  The  tick  is

     fairly  brisk,   about 2 per second. If it is very slow,  you  have

     some  resident program interfering or you may be trying to run  the

     program  from Windows. Windows slows processing  down  drastically,

     so  the timing-critical   transmit and receive programs  cannot  be

     run  from Windows.  The most common problem with the  interface  is

     wiring  errors. The pin numbers on the opamp seem to get  mixed  up

     with  the pin numbers on the COM port plug. The Wefax  choice  (99)

     will tick away even if the interface is not connected, so don't let

     that fool you. If the Wefax doesn't tick but gives a constant tone,

     you  forgot to create the RAMDRIVE and insert it's  designation  in

     line 3050.

          Once you get the machine ticking, let it run for 30 seconds or

     so  and  then hit H (for Halt) to stop the copying.  The  tone-tick

     should stop and after a brief delay, you should get an image on the

     screen---in this case it will be multi-colored noise. After looking

     at  the image, you hit Q  (for Quit) to exit the image. You can now

     choose  to SAVE the image, retransmit  the image, transmit  another

     file,  or  return to BASIC where you can hit F2 to  setup  to  copy

     another  picture.  The major keys to remember  are  SPACEBAR  which

     bypasses the frame sync and starts the ticking, then H which  halts

     the  ticking and advances you to a picture, and then Q which  exits

     the picture. You should now be ready to copy real pictures.

          Tune to 14230, the most used SSTV frequency and carefully tune

     in the voice of a transmitting SSTV station to make it sound  natu-

     ral. If you have pass band tuning or other methods for shifting the







     received  passband,  move the passband to attenuate the  lows.  The

     lowest  frequency  used is 1100 hz. When the station  announces  it

     will  send  a  certain mode, select that mode  quickly  and  ENTER.

     Different  modes  have different  header lengths---AVT  and  FAX480

     headers  are quite long before they start the picture. The  picture

     copying  should start properly synchronized. On everthing but  AVT,

     this   should be almost immediately after the fax audio begins.  If

     no  ticking starts (give it 6 full seconds for the long  header  on

     AVT pictures), you missed the header so hit SPACEBAR to bypass. Let

     the copying run until the machine turns itself off (or if you tuned

     in  late you may need to Halt when the picture ends). It then  will

     display what it copied. The pictures are intentionally offset  just

     a  little to facilitate easy scrolling to position the pix  exactly

     if  you choose to SAVE it or retransmit it. Use the right arrow  to

     drop  the trash off the left edge of the picture and then  use  the

     fine step left arrow to just bring the trash back to the left  edge

     (look  at the right edge to see what you are doing). The  trash  on

     the right will be trimmed off by the VUing program. If it is a  256

     line  picture  (Scottie, Martin, etc.), it will have  16  lines  of

     shades  of gray display so you'll want to scroll down  to see  what

     fell off the bottom of the screen (there are only  240 lines avail-

     able  on the screen). The transmit program adds the 16 line  shades

     of  "gray" display to whatever file it is asked to send, so if  you

     don't  scroll down before saving or retransmitting you'll  transmit

     two such displays and less picture. (Actually, we transmit a shades

     of red and shades of green display since the gray looks incongruous

     with  all the other pretty colors. This was promptly reported as  a

     "fault" by some folks. C'est la vie !). To facilitate scrolling the







     shades of gray out of the picture, the key D (for Down) will  scroll

     you down at 8 lines per hit.

          If  the  picture mode was a Robot mode the  three  frames  are

     different  widths  (the  chroma frames are 1/2 the  B/W  frame)  so

     scrolling  affects  the color registration--you scroll to  get  the

     proper color registration. Then you use keys F and G to retune  the

     signal in 10 hz increments to optimise the color balance. The Robot

     modes  are  critical to tuning and deteriorate more  rapidly  under

     poor  conditions so on average don't expect to see  equivalence  to

     Scottie modes.

          When  you SAVE a pix or retransmit it, DO carefully scroll  it

     so the trash on the left edge of the picture is JUST BARELY off the

     left  edge  of  the screen. The transmit program is  based  on  you

     having done that.

          Occasionally,  some  one will send a pix in a  mode  different

     from  what  was announced. Let it go ahead and copy  and  after  it

     shuts  off you can go to BASIC, LOAD VU.BAS and click  through  the

     mode  choices using the d:a.p default file name to find  the  right

     mode.  The picture probably will be skewed slightly, but you  don't

     lose the whole thing.

          Make  a  practice  of listening to the fax  signals  and  soon

     you'll be able to tell what mode it is with the tick rate being the

     most  recognizable  feature.  The AVT's  have  a  very  distinctive

     "triple  chirp" beginning (VIS code repeated 3 times)  with a  pro-

     longed  count down to start and finally the pix starts with a  more

     irregular  sound.  It doesn't appear justified to me  for  the  AVT

     modes  to spend six seconds on the header. As the number of  people







     using color fax grows, you can expect some attrition of modes  that

     do not use on-the-air time efficiently. The AVT's don't waste  time

     on line sync's but then they blow it on the header.

          This program was in the development cycle when WB8DQT came out

     with his FAX480. It was easy to add this mode to the software, so I

     did.  He only had 16 shades of gray--my 64 shades processing  gives

     truly  photographic  quality to this mode. Unfortunately,  he  suc-

     cumbed  to the "header syndrome" and loaded up what is a  very  good

     mode with another excessive header (based on unattended  commercial

     fax  machines). I rebelled at this and decided not to make  another

     header  decoder but to just wait until most  of the 20 empty  lines

     in  the header were finished to hit SPACEBAR to start it  manually,

     so that's how you have to manually acquire the FAX480 pictures. The

     empty  lines are readily recognizable by ear. If he had just put  a

     VIS  header  like Scottie on the front end, and  called it  a  high

     resolution B/W SSTV mode, it would have worked great and would have

     been  more welcome on 14230. The obsession with starting "in  sync"

     every  time comes from paper type fax systems. With a  CRT  system,

     the occasional out of sync condition just requires a little scroll-

     ing. As described below in ANOTHER MODE,  we have a VIS version  of

     this mode now.

          When  you're finished with the picture, hit Q to go back to  a

     selection screen. You can retransmit the picture directly from  the

     d:a.p file where it is stored. Or you can SAVE it.

          You'll be very happy about now if your computer is at least as

     fast as mine, and it's TIME clock loses exactly as many seconds per

     day as mine does. If that is not the case, you will have seen  some

     slant to the sides of  the picture and maybe even some  break-locks







     on  the sync (looks like someone cut the photo and slid the  bottom

     half to the left). The parameters which affect these things are  on

     the SYSTEM CONFIGURATION list (LIST 3000-3500). If you have a  slow

     computer  it may have bad breakup---if that is the case,  start  by

     doubling  the  sync  window (SW). If there  aren't  too  many  sync

     breaks,  ignore them for now and concentrate on adjusting the  line

     timing, LT, first. Since the Navy Wefax is on the air all the  time

     (on  USB on frequencies 3354.9, 8077.9, 10862.9 and 16407.9  Khz.),

     that might be a good choice to set first. Choose mode 99 and answer

     the questions  or accept the default choices. This is the only mode

     that  offers a choice of manual start. If you choose  AUTO-SYNC  in

     this mode as in all other modes you can bypass it by hitting SPACE-

     BAR  if  the SYNC doesn't  show---which happens more often  in  the

     Wefax transmissions. Scroll the image to put some vertical line  in

     the picture against the edge of the screen. If the top leans right,

     lower LT99 and vice versa. You can converge on the right value more

     quickly  if you use a ruler to measure and record  the  out-of-true

     across  the top of the screen for each value of LT you try. If your

     screen  aspect is the same as mine, the formula for the  correction

     is  1125 TAN X, where X is the angle off vertical. Or if you  count

     the pixels you are off by at  the bottom of the screen, the correc-

     tion is about 1125/240 times the pixel offset. That is, measure the

     pixel offset and multiply it by 4.7.

          The most popular mode on 14230 seems to be Scottie 1, so  that

     will  be a good SSTV mode to align first. PLEASE DON'T  ask  people

     to  just transmit for your alignment use. Take whatever gets  sent.

     Once  you determine what the exact correction for  one mode is  you







     can  calculate the correction for all the other modes. The  correc-

     tions will tend to scale with line lengths (ie, the bytes/line)  so

     you  can get very close by taking the correction for one  mode  and

     adjusting  it for other modes. For example, the Scottie 1 mode  has

     878  bytes/line. If we found that it's line timer, LT3, had  to  be

     increased  by 25 counts, then the Martin 1 (915 bytes/line)  timer,

     LT7, would be increased by 915/878 times 25 = 26 counts. For refer-

     ence  the bytes/line for each mode is: (1)306 (2)614 (3)878  (4)568

     (5)768  (6)960 (7)915 (77)465 (8)745 (9)546 (99)1024. You may  need

     to  do  a little more trimming on these calculated values  for  LT.

     Reports  from other users indicates the calculated corrections  are

     usually  right  on target, however. PLEASE DO take the  trouble  to

     make  the pictures exactly square on receive as any skew you  leave

     in will be transmitted to the world every time you send a  picture.

     In  Revision D we have added a simple program, LINETIME.BAS,  which

     will calculate the exact line timing numbers for all modes when  it

     is fed the corrected timing for the Scottie 1 mode, LT3. Fortunate-

     ly, the exact same timing numbers are used for transmit, so you can

     get your transmit alignment done without even transmitting. Believe

     me, the people on 14230 will appreciate that courtesy.

          I have read a suggestion on one of the BBS that you use a tape

     recorder to copy a picture off the air and then play that back  for

     alignment.  While most audio recorders aren't stable enough,  using

     the audio channel on a reasonable quality VCR is useful. While most

     people  operating on the non-Robot modes have pretty accurate  line

     timing, the Robot modes as mechanized on different systems vary all

     over  depending on how many updates a particular station  has.  You

     can  use the non-synchronous modes, (11) for Robot 36 and (22)  for







     Robot  72,  on receive to avoid having to align the Robot  LT's  but

     eventually  you'll want to retransmit a Robot pix and it  would  be

     desirable  to have determined the  synchronous  (ie,crystal-locked)

     alignment---on  modes  (1)  and  (2)---with  an  up-to-date   Robot

     equipped  station like W5ZR or N4OBQ. Note that LT1 is exactly  1/2

     of LT2 so you only need to experimentally align one. The calculated

     value  for this correction on Robot will be far better than  you'll

     get  by trying to align to just any station you happen to  hear.  I

     heard  a lengthy exercise on 14230 recently which  illustrated  how

     NOT to do this. A station with this system asked W7--- to send  him

     a Robot picture so he  could align his  LT (which is a no-no).  The

     W7  was  using a Pasokon which has different line timing  from  the

     Robot  systems (on the Robot modes and the Wraase 96), so the  fre-

     quency  was tied up for an extended period of time aligning to  the

     wrong standard. Remember the SSTV frequencies are like giant party-

     line phones with many people listening and waiting to send or  copy

     pictures. Move your extended chit-chat to other frequencies.  Since

     almost  everyone  has resigned themselves to  copying  Robot  modes

     using Line-Sync due to the wide variety of  timings, the   benefits

     of crystal locked  reception of the Robot will only be available to

     those who bother to set their timing to the Robot standard, and  is

     communicating  with  a station who has done the  same  thing.  When

     making these LT adjustments, note that if you pick anything but the

     BASIC choice when you exit the picture, you will lose any temporary

     changes  you  have  made to the CONFIGURATION list.  When  you  get

     perfectly  square  pictures, SAVE the program with  the  latest  LT

     value. The program will only accept values for LT up to 65,535. You







     start over with 0  if you need greater numbers.

          After aligning LT, if you still have any sync breaks, you will

     need  to  lower the Gross Timer. The process here is  identical  to

     trying to fit tiles into a fixed size room. We've squared the  room

     with  LT. We have a few hundred tiles  (pixels) to squeeze  into  a

     room  which has a fixed size (LT) and if the tiles are  too   wide,

     the  last tile will  run into the wall  and break sync. The  number

     for GT represents the width of each tile  (pixel). Decreasing it by

     1  will slice one microinch (microsecond) off of every   tile.  You

     can slice several microinches off each tile  (maybe decrease GT  by

     5)  so you're sure it  will fit but that may leave an ugly  gap  at

     the  edge  of the floor. You can cover that over with trim  at  the

     edge  of  the floor, but if you copy the picture  out-of-sync  that

     ugly gap moves into the middle of the floor (picture). So you  want

     to  make GT low enough to eliminate any occasional break  locks  but

     not so low as to leave an ugly gap which will show up when you miss

     sync  on a picture. If your computer is very fast, the  measurement

     accuracy  for  each pixel (tile) will be more precise, so  you  can

     stack the pixels with less gap at the edge without having the  risk

     of  running into the "wall" set by LT. If you have sync  breaklocks

     that  seem  independent  of the GT value, you need  a  larger  Sync

     Window  (SW). The SW represents the amount of time in  microseconds

     it takes the computer to decide if  it's at the LT "wall". Obvious-

     ly, a slower computer needs more time.

          You  may ask why we chose to use only one pixel  (tile)  size,

     GT,  to fit all these different modes (rooms) with  different  line

     timing  (widths).  It  was just a matter of  convenience.  You  can

     further  optimise  the  system by using a  different  GT  for  each







     mode---in fact, with the deterioration you might get from a  slower

     computer, this could be quite worthwhile. You will note that we  did

     use a different GT (GT99) in the Wefax mode.  Remember, the way  to

     see this effect is to copy a picture out-of-sync and look carefully

     at the results at the parting line and on either side of that  line

     look  at the color registration. Since the AVT modes have  no  line

     sync period in which to hide this "gap", the effect is more notice-

     able there.

          People  who have not seen SSTV pictures are usually struck  by

     the  fact that many of them are noisy--in fact, anything  you  hear

     (noise,  splatter, QRM, etc.)will be painted into the  picture.  So

     high S/N ratios are needed for flawless pictures. Narrowing the  BW

     will  help  some, but at some point will smear the  picture  bytes.

     Another  source  of noise was found by KY1S in a  serial  interface

     board  (made by Data Technology Corporation) which he  changed  out

     for  another board he had which was clean. I have heard of  several

     other  cases  since then which were solved the same  way.  Opinions

     have been heard that this is more common on the newer, more compli-

     cated,  boards. The effect varies from a few flecks in the  pix  to

     very noisy pix even with very strong signals. These I/O boards  are

     very  cheap  ($10 to $15) but finding which ones work  may  take  a

     little  research on the air. NF4K reports PT-606 as being  a  quiet

     card  and M10-16 and UN-1072 as being noisy. AA7GU reports  UN-1072

     as  being  noisy. He replaced it with a very  clean  board,  D.F.I.

     model MIO-400 KF Rev. F. This board also got rid of a computer  RFI

     problem  he had with his mouse connected. K2HTO reports  his  Tandy

     2100, SX25, installed I/O board had much noise. A card from a local







     supplier (G.C.S. tel:[phone]---board is stock no. C9806)  was

     completely  clean.  General indications are that  the  simpler  I/O

     cards work best--cheaper is better!

          Signal modulation can also come from your receiver. My  TS930S

     is  completely clean while an old ICOM-720 which I use on the  boat

     gives  noticeable hum bars in the pictures even though there is  no

     detectable  hum in the audio. Clearly, the 120 cycle  is  frequency

     modulating  the high frequency phase-lock loop in the IC720 but  it

     was undetectable until I copied SSTV. With such a sensitive way  of

     seeing  it,  I  intend to fix it; and I won't be  surprised  if  it

     improves  my  SSB voice signal. Both of these cases may be  due  to

     design  faults or just to the particular piece of  hardware.  Point

     is, SSTV shows up shortfalls in hardware very clearly!

          You need at least 512 kb. of memory on the video board to copy

     the SSTV modes. The Wefax copying optimally will need 1 mb. But the

     provisions  are  made  in the VUing program to look  at  Wefax  pix

     through  a  640x480  window. The program  provides  1024x768  Wefax

     viewing  on video card choices Trident, Tseng, and VESA  only.  The

     normal Wefax image has 1024 pixels per line and 1536 lines, so  you

     only  see a portion of the total image at a time. The keys  S  (for

     Shrink)  and X (for eXpand) can be used to see more or less on  the

     screen. Also, if you are using the 1024x768 screen, you can get  an

     expanded view of a portion of the pix by switching to 640x480.  Try

     copying  maps in the Wefax mode. You'll be able to read the  really

     small print ! If you want to print a copy of the map, position  the

     place on the map where you'd like the printer to begin to the upper

     left  corner of the screen and hit P to start the printer.  Hitting

     any key will stop the printer.







          The  Wefax images initially appear in 64 gray shades. You  can

     choose  color  palettes 1 through 5 for various  effects.  This  is

     useful for seeing ocean thermal currents like the Gulf stream on IR

     images (you can identify IR images as the black sky will be  white;

     ie,cold). Cold is white and warm is black  in IR images. Or you can

     take  a  vertical profile of the clouds by first hitting G  to  get

     Gray,  then hit C to Color the top (coldest) layer. Each  time  you

     add another C, you color the next lower level. A total of 64 levels

     are  available. This is useful to identify the more severe  weather

     which  generally boils up to higher altitudes. When you use one  of

     the  color palettes, C can also be used to step through  the  color

     palettes  to highlight visibility of small temperature  differences

     at sea level to see things like ocean currents. These currents  are

     of particular interest to fishermen.

          A portion of a Wefax pix may be saved in FAX480 format. You're

     seldom  interested in the whole image and it takes a lot of  memory

     to save the whole thing. I use date/time/mode filenames for  these.

     28MY123V.FAX  denotes  1230Z on May 28 in Visual mode.  Scroll  the

     Wefax image to put the portion to be saved in the upper left corner

     of  the screen (you save 536x480 pixels--I use pencil marks  on  my

     CRT  to  mark these borders). When you or someone you  send  it  to

     views  this  pix with VU, you select the Wefax mode (99)  and  when

     queried  about screen size take the choice FAX480 PIX.  This  gives

     you the colorizing choices of Wefax described above for viewing  IR

     pixes. Or just VU it as a B/W FAX480 pix.

          One  last item on RT.BAS is the parameter, DE, on  the  SYSTEM

     CONFIG  list. I made this delay DE=0 to suit my own practice.  This







     value  starts the picture copying a little early, so you'll  always

     need to hit the right arrow first to scroll  into picture left edge

     coincidence.  You  may  prefer to set it to start  at  the  picture

     beginning--accepting  the risk that QRM will occasionally  make  it

     start  late. If it starts late, you will move the "edge  of  floor"

     gap  into the active picture region instead of hiding it under  the

     SYNC period. You must set DE to numbers less than 256. Generally  it

     will be less than 20.

          When you manually select the mode, the frame SYNC for the non-

     AVT modes is fairly simple and can be spooked  by any 1200 hz  tone

     of adequate length. This has not proved to be a problem except  for

     a very few stations. For some reason, the Robot (?) systems have  a

     capability  to send your call sign in CW at the beginning  (instead

     of  at  the end ?) of picture transmission and they  chose  to  use

     keyed  1200  hz. Most people have turned off this  feature  (giving

     your  call sign verbally, immediately followed by giving it in  CW,

     followed  by  a picture with your call sign on it seems a  tad  too

     often) since it uses valuable on-the-air time unnecessarily. A  few

     gentle reminders may fix this problem. With a little dexterity, you

     can hit the mode ENTER key on the last character of the CW and  get

     a normal start--or just wait until start and bypass with the SPACE-

     BAR for any stations which continue to use this feature.

          Since the above paragraph was written, I  have added automatic

     VIS recognition to the program. You need to be within 50 cps of the

     transmitting  station's frequency for this to work well  since  the

     VIS  frequencies  are so close. Unfortunately, people  have  gotten

     careless  about  being right  on  frequency. When you  select  this

     option, the mode name is displayed as soon as it is recognised. You







     can  break out of this mode by hitting any key, so if it  fails  to

     grab  and  start copying, you can hit the proper mode key  and  hit

     ENTER  twice  to  start manually. This detection is  made  to  have

     almost  a zero chance of false alarm, and it also only  works  with

     fairly  good   signals.  Since I wanted to use  it  for  unattended

     operation to copy only fairly noise free pictures, this fits  fine.

     It's  sensitivity  vs. false alarm rate can be changed  by  playing

     with  the parameters in lines 2605, 2606, 2607 of the  RT  program.

     You'll  probably end up about where it is now. Rock  the  frequency

     around  + or -100 cps.  to get a feel for where the  VIS  detection

     centers.

          I  have made several versions of the VIS  recognition  machine

     language  .ASM module; namely VISCODE.ASM, VISCODEA.ASM and  VISCO-

     DEB.ASM.  As  of Revision D time, VISCODEB is the clear  choice.  I

     have  precisely measured the VIS code frequencies (1200  cps,  1300

     cps and 1100 cps) being transmitted by various stations and  except

     for  the Robot systems, there is considerable variation in how  far

     above  and below 1200 cps the systems transmit. One popular  system

     goes  up to 1300 cps OK but only down to 1180 cps. So if  you  have

     trouble with certain stations, record his VIS with an AVT mode, and

     look  at the frequencies using the new TUNE program to see what  he

     is transmitting. While absolute frequency measurement is not possi-

     ble  due to the vagaries of whether you are on the same  frequency,

     the differential frequency measurement is very precise.

          Another mode added provides Auto-recognition plus Save to give

     you unattended capability. It ignores weak signals. You can  select

     any  drive\path   to  Save the pix in. I use the RAMDISK  so  if  I







     forget it's turned on I won't fill up my hard disk. When I return I

     use  d:>ss (ie, Slidesho) to view what was received to decide  what

     to save more permanently. You can break out of AUTO+SAVE by hitting

     F5 to goto Basic, then F10 to return to DOS.

          In revision C of RT, we have added the capability when  saving

     pictures  of  doing it not only in full form but also  directly  in

     compressed format using a batch file, PK.BAT, which must be  placed

     in  the  same  directory  as SLIDESHO, TRUSHO,  etc.  (ie,  DOS  or

     TOOLS).  In revision D we use PKZIP so you need a copy of that.  If

     you  like  some other compression program, you can  change  PK.BAT,

     PKV.BAT, and PKX.BAT to incorporate it.

          If you copy (or REName) a file in the RAMDISK to d:a.p you can

     then  LOAD RT: make a temporary change to RT by typing a  new  line

     880-----880  GOTO 890----and then RUN the program. Select the  mode

     the  picture was copied in and ENTER. The pix will appear as if  it

     was just copied and you can rescroll it if needed and then save  it

     permanently. This is another non-obvious use for RT.

          If  you  don't  want the tone  tick  (measures  the  program's

     "heart-beat"),  Rev C has added a "heart" symbol on the  screen  to

     show when the tone is ticking. This is contained in  lines 775  and

     776  in  Rev C in case you want to remove it. The  easiest  way  to

     remove  a  line  but keep it's content for later  reference  is  to

     insert  an apostrophe as the first character in  the  line--turning

     the line into a Comment.

          Since AVT modes have both  a VIS code and a complicated  frame

     SYNC  sequence,  you can get hung up by passing the  VIS  test  but

     missing the frame SYNC test  (it's capture window is narrower  than

     the  VIS), so I just bypassed the frame SYNC and copy them  out  of







     SYNC. You can eliminate line 724 if you want to keep the frame SYNC.

     Don't  forget to clear out the RAMDISK for future copying. You  can

     tell  when  the RAMDISK gets full when you copy a pix and  it  only

     fills part of the screen.

          In my experience with the previous wefax programs, I have only

     seen one computer (a name brand laptop) which didn't have a proper-

     ly  operating crystal-controlled clock. I have heard of one  other.

     The  sympton  is the lines in the pix wander all  over  the  place.

     You'll have to fix the clock to crystal-controlled stability to use

     these  programs  if you run into this. We have seen one  case  like

     this  which only broke out of crystal-lock in Transmit. This was  a

     486  machine  and K3OWX solved it by going out of the  turbo  mode.

     There  is some general trend we have noted that the crystal  clocks

     are not being held to as stringent tolerances as before.



     REVISION D  ADDITIONS TO RT.BAS

          KY1S  added a picture menu screen to his copy of  our  program

     using some fancy Basic programming. It seemed neat, and not expect-

     ing  very many would want to go to that trouble, we decided to  add

     the capability for you to add your own customized picture menu just

     by  making a file. That is, take your own favorite picture and  use

     PCXLABEL   to  add  the  menu  items.  The  menu  pcx   files   are

     MODELIST.PCX,  FCNKEYS.PCX, K3BCSSTV.PCX, and PICKMODE.PCX.  You've

     already  seen the one I made. You'll probably want to put your  own

     call sign at the top. The menu file name goes in line 3240 and  the

     mode number goes in line 3245 of the RT SYSTEM CONFIGURATION.  Line

     3230  gives you a selection of menu choices. Try all 3 to see  what







     the differences are. Also, to minimize the key strokes when operat-

     ing   from  the  picture  menu,  we  added  a  small  ML   program,

     SELECTMO.ASM,  which gives single key stroke mode command  (ie.  no

     ENTER  required). On modes like M1, mode 77, which require two  key

     strokes,  a 1 second delay is allowed for the second  stroke.  This

     delay can be changed in line 1352 of RT.

          We  have  included the Trucolor viewing into the  RT  program.

     There  are a number of possible choices of picture size  and  color

     resolution  on both the menu screen and the received  pictures,  so

     try  all the combinations to see what you like; lines  3250,  3260,

     and 3270 contains the switches. We have used fractal techniques  to

     take  the pixellization out of the full screen  Trucolor  pictures,

     as described below in TRUSHO.



     REVISION E ADDITIONS TO RT.BAS

          We have added REAL TIME copying capability to RT for those  of

     you with TRUCOLOR video cards and fast enough computers (386-25 mhz

     equivalent  speed approximately). It is switched IN or OUT in  line

     3280  of RT. Robot modes are not covered since they take  too  much

     processing to display quickly. The real time pix is raw as received

     with  no processing or resizing. In fact it gives you a  very  good

     illustration of the effectiveness of our poor man's digital filter-

     ing  which  replaces bytes that lie outside the 1500  to  2300  cps

     range.  In  AUTO modes you'll need to be quick to  catch  the  mode

     names. Also, the six second wait for the AVT header to be  received

     seems  aggravated since you are looking at the screen just waiting.

     Be  sure to get your Trucolor card all integrated using TRUSHO  and

     then  transfer  the card's SYSTEM CONFIG parameters  to  RT  before







     trying this.

          The Wefax mode will also copy in REAL  TIME. While none of the

     real  time screens can be scrolled while copying, the Wefax  has  a

     skip control to jump the picture to the left on the screen approxi-

     mately 1/7 of a screen every time you hit S (for skip). This  actu-

     ally changes the SYNC point on the incoming data, so you'll  gener-

     ally only use it early in the picture. You can quickly  see how far

     out of SYNC picture is and estimate how many times to hit S.

          We have added a TUNE screen to facilitate getting on the other

     guy's frequency and to give you some insight into what the  equiva-

     lent  analog signals look like. It will give you snapshots of  what

     the signal looks like as it comes in (at the expense of  interrupt-

     ing  the copying--the computer is strained to it's  capacity  while

     copying,  so  interleaving  a Tune display  without  degrading  the

     copying isn't practical) or after the picture is received, you  can

     scroll  the TUNE screen throughout the pix to see what's going  on.

     The same scroll commands used when viewing the picture are used. To

     view the line SYNC in modes that have one, you scroll down one line

     and  then hit the left arrow several times to bring the SYNC  pulse

     into  view. Now you can scroll down one line at the time  with  the

     down  arrow and see successive SYNC's and what their average   fre-

     quency  is.  Or you can scroll along a whole line  using  the  fast

     scroll  R  key. If pix was an AVT copied in AUTO+SAVE, it  has  the

     header  captured and you can look at the 1600, 1900, and  2200  cps

     header  patterns. You can even go back to old files, copy  them  to

     d:a.p,  change line 880 as described above, and view their  tuning.

     Note that the scale on the TUNE screen is not linear.







          When using the TUNE screen while copying, just hit T, and  the

     last  400  or  so pixels are displayed. Hit T again  for  each  new

     snapshot, We used snapshots because displaying continuously is  too

     confusing to use. When you want to resume copying, hit G; or if you

     want to quit, hit Q. The snapshots cannot be scrolled. If you  want

     to look at another stations VIS code frequencies, you can use  mode

     6  and  SPACEBAR to start copying just before the  picture  starts.

     After  copying a few lines, you can halt and go to the TUNE  screen

     to  scroll through his VIS code frequencies which should  be  1900,

     1200,  and then switch between 1100 and 1300. This is probably  the

     quickest  way  to get on frequency. Or if he sends  out  a  tune-up

     tone, the T snapshot in mode 6 would work well.

          If  you have a precision audio oscillator available,  you  can

     calibrate  the TUNE screen more precisely by copying the tone  from

     it.  Calibration for the snapshot is in line 480 and for  the  file

     perusal TUNE screen is in line 2805. The same correction POKE  goes

     in both places.





     TX.BAS

          [NOTE:  The TX program has been superceded by the VT.BAS  pro-

     gram below, so it is not being kept updated with new changes. Since

     there  may be someone with a very slow computer who might  want  to

     still use TX instead of VT, it is still included.]

          The  transmit  program  uses  the  same  SYSTEM  CONFIGURATION

     choices  as the receive (RT.BAS) program, so the first step  is  to

     transfer  those  choices.  LIST 1000-2000 to  see  the  choices  on

     TX.BAS. Be sure to put the cursor in the clear and re-LIST to check







     things before SAVEing the modified program. Note that if you  don't

     put the cursor in the clear before hitting a function key, it  will

     type  the function key word right where the cursor is,  over-writing

     the program and giving some weird results.

          Running TX.BAS is very simple. Put  the transmitter in VOX  if

     it  isn't already there. You select the file you want to  send  and

     note  what  mode suffix it has. The mode suffix and file  name  are

     given when the program asks for them and after the file loads,  you

     just  hit  G  to start transmitting. If  you  give  a  non-existing

     path/filename, the program will return you to the BASIC "Ok"  where

     you can hit F2 to try again. In this case ignore the Picture Trans-

     mitted message---it should be obvious that nothing has been  trans-

     mitted. When you key G to start transmitting, the VOX  will come on

     and  the header is transmitted. The tone-tick audio cue will  start

     when  the actual picture starts being transmitted. If you  want  to

     terminate  the sending before the picture is finished, key  H  (for

     Halt).  Listen to the ticks to see that they are very  regular.  If

     you  hear a skip, someone on frequency will probably tell you  that

     the picture had a break-sync. Although it didn't happen on the  two

     computers  I  have, it's conceivable that a different GT  might  be

     needed  for TX.BAS since the overhead processing done on each  line

     is different from the receive program. If that is the case, use the

     lower of the two values for GT in both programs. A better  alterna-

     tive  to lowering GT is available if the offending mode is  NOT  AN

     AVT  MODE. You can LIST 240-370 in the TX.BAS program   and  locate

     the  line that has the data for the offending mode (ie, 290 IF  I=4

     THEN LL=568 etc.etc. is the line for Mode 4). That line will have a







     value  called  out for a parameter SC, which is the number  of  1/2

     cycles of 1200 hz line sync that will be sent on each line.  Lower-

     ing  this value will fix your problem (equivalent to  removing  one

     "tile"). It's very unlikely you'll need to do either. Incidentally,

     don't  try  to send a picture in a mode other than the  one  it  is

     formatted  for. Hopefully I can find time to include a mode  refor-

     matting program in future revisions.

          People seem to be using rather long sub-directory names  these

     days, so if your PATH to the file to be transmitted is long, it may

     not  fit  in the space allocated in the program. You can  copy  the

     file  to  the  RAMDISK as A.P and use the  program  default  choice

     instead if you run into this:ie, if the program rejects a perfectly

     good Path/Filename.

          Before  you  actually start transmitting pictures,  hook  your

     headphones  across the audio transformer output to the  transmitter

     and  listen to the signal while sending a pix. Compare it  to  what

     you hear when receiving. If the frequencies sound quite  different,

     you'll save time by adjusting the header frequency as described  in

     VT.BAS below before proceeding. If the frequencies are very differ-

     ent, you may be breaking out of SYNC on many lines and transmitting

     a mess. We've seen one 486DX2-66 machine  where this was the  case.

     His  frequencies were quite low. In using this "headphone" test  on

     my  laptop's transmit audio output on the RTS pin, I found  a  1000

     cps  tone there continuously. This was undoubtedly  the  alternator

     frequency  on the + and -12 vdc supplies which were not  adequately

     filtered.  While not fully confirmed, I'm fairly sure this  is  the

     source  of some wiggle (beat note effect) in the vertical lines  in

     receive and transmit with  this machine.







          You  can  do  the initial checking of   this  program  without

     actually  transmitting,  of  course. But as soon as  you  do  start

     transmitting,  BE SURE to check that  the transmitter is  operating

     at a low enough level so it won't overheat. Most transmitters  will

     not handle a 100% duty cycle signal  like this and still operate at

     maximum peak power. I failed to take my own advice on this and  had

     to make a multi-transistor repair on my TS-930. If your transmitter

     has an AM mode, set the output power in TV mode to the same  output

     power  level  as AM. This should be plenty safe.  The  audio  level

     resistor,  R_x in the schematic, is set so that you can  leave  the

     transmitter mike gain at the same level whether you are talking  or

     are sending fax.

          A  useful  feature in BASIC is the ability to list  all  files

     contained  in the current default directory. For example,  to  list

     all Scottie 1 files, just type:    FILES"*.S1    and RETURN       .

     TX, VU, and SLIDESHO all can use this, so I have included an insert

     to  make key F9 type the words FILES for convenience. FILES is  the

     BASIC  equivalent  of DIR in DOS format, so the use  of  wildcards,

     etc.  apply.  Note that keys F9 and F10 are not operative  for  our

     program specific-use until at least one program has been RUN.



     VT.BAS

          This  is  identical to the TX program except the  file  to  be

     transmitted  is put on the screen to see if it is the one  desired.

     It  stays on the screen after transmitting so you can discuss  it's

     details.  Q will exit the picture. If you call up a  non-file,  the

     screen   will  be blank---you'll need Q to get back to  BASIC.  The







     SYSTEM CONFIGURATION for the VUing part of this program is seen  by

     LIST  2000-3000 and is copied from VU.BAS. The transmitting  SYSTEM

     CONFIGURATION is seen by LIST 4000-5000 and is copied from RT.BAS.

          I have added a couple of features to this program as a  result

     of  inputs from KY1S. He found it convenient to have listed on  the

     screen  all of the files in the current directory which were  saved

     in  the mode being used every time he picked a mode. That is,  when

     you  pick  mode 3 in the VT program, all Scottie 1 files  would  be

     instantly  listed for you to select  from. With the BASIC   format,

     this is very easy to insert and is a perfect example of how you can

     adapt  to your own operating methods. I thought others  might  like

     this  capability, so I added it so you can switch it ON or  OFF  in

     line 2092.

          Dennis  also  found that his LT numbers had to be  changed  by

     about 400. Other  486 users have also reported a wider variation in

     the  frequency  of the TIME-OF-DAY clock than we had  seen  before.

     While this has a rather small effect on the color registration, the

     headers used  for automatic recognition of modes have somewhat more

     stringent  tolerances. If you get reports of people receiving  your

     pictures  having  to manually start their reception, or you  had  a

     large change in LT or a particularly slow computer, you may want to

     measure  and possibly trim the header frequencies. We have added  a

     test  mode in VT.BAS which is turned on by a switch,  TE,  in  line

     4173  in  SYSTEM  CONFIGURATION. Switch TE to a 1200  or  1900  cps

     position and then hook up a counter or some other method of measur-

     ing  frequency to the output of the interface which normally  feeds

     fax audio to the transmitter. If you use a counter  you will proba-

     bly  need to reduce  the signal down by 50 or more to  avoid  over-







     driving the counter. Load in a Scottie 1 picture, and when you  hit

     G(o),  a prolonged steady tone at the selected frequency will  pre-

     cede  the  picture.  Parameter HF in line 4171 can  be  changed  to

     center the frequency. If the frequency is 3% low, raise HF by 3%. A

     potentiometer  to  change the level into the counter is  useful  to

     avoid  the  erroneous results you get by overdriving  the  counter.

     After finding the best value for HF, SAVE it in VT.BAS and put  the

     same value in line 3210 of RT.BAS and SAVE it. While this parameter

     is called "header" frequency, it actually recalibrates both  header

     and picture frequencies with the pokes in line 3011 of VT.BAS  (for

     reference  in case you wanted to change only one of these, 2E9  and

     2EA  are  the location for the header and 1FE and 1FF are  for  the

     picture  data frequency). Many SSTV stations can give you  a  rough

     measurement  of these frequencies over the air.[NOTE: This  feature

     was  added  in Rev A; then in Rev B we  inadvertently disabled  it.

     While we fixed the error fairly quickly on the ARRL BBS, some early

     downloads had this error. It is OK in Rev C.]

          In revision E we have added a TUNE screen to RT.BAS so if  you

     have  a VCR, you can record the Transmit audio output on the  VCR's

     audio  channel and play it back into RT to look at the  frequencies

     with the TUNE screen. While the VCR will add noise and jitter  (the

     VCR  mechanical "noises" are in the audio range), you can  set  the

     value  for HF with this. Or if you want to compare  picture  shades

     for  different HF values, make recordings of the same pix  file  at

     several settings for HF. The received image files are labelled with

     the HF number and saved  or copied into the RAMDISK along with  the

     original  file. Now you can use TRUSHO in it's 4-picture  mode  and







     make shade comparisons between these to get a shade match with  the

     original  over the entire  image. This is amazingly  accurate  when

     the  match is made on the same screen. When you get a  match,  Quit

     from TRUSHO on this pix and by keying Yes on the next question, the

     file name value for HF will be displayed.

          Another  feature  added to VT.BAS in revision E  is  a  tuning

     reference  tone.  After  loading a picture to  be  transmitted  but

     before keying G to start transmitting, you can key T (for Tone) and

     a  pulsing  1200 cps tone will be transmitted as long as  you  hold

     down  the  key. This could also be used in setting HF via  the  VCR

     technique as described above.

          Another  problem  which  I and others have  encountered  is  a

     radical  shift  in frequencies both received  and  transmitted--ie;

     overly  light  or overly dark pictures. This has been  isolated  to

     residual  settings  of computer parameters which are  left  in  the

     computer  after using such programs as JVFAX, HAMCOM,  and  others.

     They  can be removed by re-booting the computer. In Revision  B,  I

     think we have isolated the offending parameters and modified the RT

     and VT programs so it won't be necessary to re-boot. Feedback  will

     be appreciated.

          Another  problem which a few others have experienced which  is

     similar to the above (but only affects transmit) has been fixed  by

     taking  some of the flexibility out of VT  and manually fixing  the

     processor  speed calibration number to fit a specific machine.  The

     trouble  symptom is a transmit function which intermittently  works

     properly  but then reverts to sending trash. The  transmit  program

     uses the processing speed as it's frequency determinant. This speed

     is calibrated each time the program is run. This makes the  program







     work  independent of whether you are using a slow 286 or  a  super-

     fast 486 and also corrects in case you turn the turbo on or off. It

     uses the 1/18 sec TIME clock ticks to calibrate itself. Apparently,

     in  the  computer  we fixed, either these TIME  clicks  were  being

     intermittently  changed by some internal software, or there  was  a

     hardware  intermittent. You can bypass the calibration and set to  a

     fixed value by adding the following line to VT.BAS:



     3019 poke &h165,144:poke &h166,184:poke &h167,128:poke &h168,6



     The number poked into &h167 is the LSB and &h168 is the MSB for the

     calibration number. The numbers shown are for my 386 at 16 mhz. You

     can determine the correct numbers for your machine by using the  TE

     switch  and  adjusting the values in &h167/&h168 to  get  the  test

     tones  on  frequency. Or when the machine works properly,  you  can

     read the correct values by peeking into memory after you return  to

     BASIC. The value for &h167 prints out by typing PRINT  PEEK(&H5A3).

     The value for &h168 comes from PRINT PEEK(&H5A4). My 386 at 33  mhz

     gives a MSB in &h5A4 of 13, so you can scale from there to  get  in

     the ballpark.



          In  Rev. B, we have reconfigured the headers on both  the  AVT

     modes  and the FAX480 mode. Not having found anyone with  a  FAX480

     receiving  capability  before the original release,  it  never  was

     tested. It had an error in the software. While fixing that, I  also

     brought  out an adjustment so you can tune the initial 244  cps  in

     the  header to  be precisely on frequency. This is located in  line







     3015  of VT.BAS. In the AVT headers, while they worked,  there  was

     some diversion in the SYNC point depending on which type of  system

     was  doing the receiving. Since all the systems seem to  work  well

     with the headers from Bert's  (W5ZR) AVT pictures, I just  scrapped

     my  header generator and made a digital recording of Bert's  header

     and  included  it ahead of the picture. In retrospect,  this  would

     have been a much easier way to generate  all of the headers.

          In  revision D we have added the capability to both  view  and

     transmit compressed files. You can collect compressed files into  a

     single file, but it takes a lot of memory and extra time to extract

     a  single  one  of these out. I prefer to collect  mine  as  single

     compressed files on floppies  or in subdirectories named after  the

     mode  used; ie, in a S1 or M1 or 90 (for AVT90), etc  subdirectory.

     That  is,  the suffix for the uncompressed files is  used  for  the

     subdirectory  name.  Since  the  compressed  files  all  have   the

     same  .ARC suffix, it is useful to collect them together  by  mode.

     The VT program can handle single compressed files; it just takes  a

     little longer for the program to list the file name and mode so you

     know which mode to call up. You can use this program to view  indi-

     vidual files, however, if you want to have a slide-show, it's  more

     convenient  to  decompress a series of files into  the  RAMDISK.  A

     simple  batch file can be written using wild cards (pkxarc  a:*.arc

     d:) for this.



     SLIDESHO.BAS

          SLIDESHO  works just like VU except that it has an  additional

     capability  to  step through a series of pictures  without  knowing

     what  the  filenames  are. You pick a mode, and if  you  want  some







     constraints in the filenames viewed, enter that. For example, if you

     wanted  to look at only Scottie 1 pix's starting  with the  letters

     BE, enter 3 for the mode and enter BE for the constraint. The first

     BE*.S1 pix in the directory will paint. To see the next BE*.S1 pix,

     hit N (for Next). It will exit to BASIC when it runs out of  files.

     You can use this program to look at single files also. This program

     (and it's associated files, SS.BAT. PK.BAT, and 64SHO.ASM) is  con

     figured  to  go into your DOS directory, so it will  be  accessible

     from any drive or any directory where you might have picture files.

     If  you  want  it somewhere else, you'll need to  change  the  PATH

     callout  in line 30 of SLIDESHO.BAS and in SS.BAT. They  should  in

     any  case  be placed in a directory which is listed  in  your  path

     statement in AUTOEXEC.BAT. I find it convenient to store pix  files

     in  a compressed .ARC file (using PKARC) grouped together by  mode.

     When  I  want  to view them, I decompress the .ARC  file  into  the

     RAMDISK  (using  PKXARC---the  format for an  archived  file  named

     S1PIXES.ARC  is: PKXARC S1PIXES d:). After the files load,  with  a

     quick d:>ss I can start the show. Incidentally, this program is not

     configured to work with the Robot modes or the Wefax mode.

          Another  feature  added  since the original  release  is   the

     capability to make the SLIDES advance automatically. The number  of

     seconds  you want to dwell on each slide can be set in SYSTEM  CON-

     FIGURATION, line 2050 as DL. Instead of N(ext), you key A(utomatic)

     to  start the show. It will go through every pix in  the  directory

     and  then start over again indefinitely. Q will exit  sequence.  If

     you get large numbers of pix in the same mode in one directory, you

     may  have  to raise the "files=" reservation in CONFIG.SYS  to  use







     AUTO mode.

          Revision C of the software adds a new capability which is very

     useful if you use the AUTOMATIC + SAVE mode. Typically you store  a

     number of pictures in the RAMDRIVE while the machine is unattended.

     When  you return you use SS (SLIDESHO) or TSS (TRUSHO) to view  the

     pictures in sequence. Often you see ones which you want to save, so

     the capability was added to both scroll, rename, and SAVE in either

     full or compressed formats any picture that you like. When you view

     one  you  like, use the normal scrolls, if needed, and hit  Q.  The

     program  then walks you thru the steps from there. If you  use  the

     compression, it requires the presence of PK.BAT in the same  direc-

     tory where you store SLIDESHO and TRUSHO (typically in DOS or TOOLS

     directory).

          When you're all finished remember to clear out the RAMDISK  if

     you  used  it for temporary storage and you're going to  return  to

     copying  pictures; particularly Wefax which uses a lot  of  RAMDISK

     memory.



     TRUSHO.BAS

          This is a version of SLIDESHO that supports TRUE COLOR (ie; 16

     million color) video boards with a 640x480x16.7 million colors mode

     capability.  The display is brighter than is possible with the  256

     color  video board. Also, in Automatic, you can choose between  one

     picture  at  a time filling the whole screen or  4  pictures,  each

     filling  1/4  of the screen but still at full resolution.  You  can

     S(top) the automatic sequence at any time and restart it by hitting

     G(o). I haven't yet found out what mode number is being used by the

     different  card manufacturers for this mode, so you'll have to  use







     the  VESA  choice  unless you have the same card  I  have  (Diamond

     Speedstar  Pro).  If you can find out the mode number,  it  can  be

     poked into memory location &H113 with the card driver sequences  in

     the BASIC program (the one with your card name on it). Interesting-

     ly,  Diamond's  instruction  manual said it was mode  &H72  but  it

     actually was mode &H71.

          In  Rev  C,  we have added capability to  accomodate  the  ATI

     TRUCOLOR  card in  the VESA card choice. I thought the VESA  stand-

     ards  would actually settle on such things as  bytes/line--no  such

     luck! ATI uses the same byte storage sequence as all previous  SVGA

     modes  where  the lines are  stored seamlessly so there is  no  gap

     between  lines. Diamond (or maybe it was Cirrus, whose  chipset  is

     used  in the PRO board) decided to jump to the next 2K boundary  in

     memory at the end of each line. I can't think of any other rational

     choice,  so maybe future cards we see will fit in one of these  two

     categories. The parameter BL is included in the SYSTEM CONFIG  list

     of  each program to set this. If someone comes up with  some  other

     oddball  choice, just assign the value to BL. The wrong BL gives  a

     greatly slanted picture.

          This  program gives very high quality pictures which can  com-

     pare favorably with any other  SSTV system. The pictures themselves

     are  captured, like the Robot 1200C, at 262,144 colors.  While  our

     basic  program structure could be easily modified to send  and  re-

     ceive  at the full 16.7 million colors, it's  questionable  whether

     you could tell the difference. You can tell the difference  between

     32,768 and 262,144, however.

          Be  sure  to read the section on the Bank-Shift  parameter  in







     VU.BAS  above  when setting the SYSTEM CONFIGURATION  on  this  new

     program. This program, like  SLIDESHO, allows you to save  pictures

     directly.

          In  revision D we  have made a significant improvement in  the

     full  screen pictures. Using fractal techniques, we managed to  get

     rid  of the "square pixel" effect which is characteristic of  lower

     resolution  pictures  which are expanded to fill the  screen.  Line

     2080 has a 3 position switch for choices on the even lower  resolu-

     tion Martin 2 and Scottie 2. Try all  three. I like choice 2 as  it

     has the proper perspective and the pictures are not over-expanded.



     TIFCONV.BAS

          If  you  want to create some picture files to  transmit,  this

     program will convert from the TIFF 640x480, 24 bit color format  to

     the SSTV/FAX480 formats. We chose the TIFF format because it is the

     most  common  format used for  transferring  high  color/resolution

     pictures  between  programs and is an available  format  from  most

     frame  grabbers. The typical frame grabber will accept images  from

     any video source. I have tried this with both the Computer  Eyes/RT

     and  S.S.C.  grabbers.  For some reason,  the  picture  information

     started  at a different offset into the picture file in  these  two

     cases.  LIST  205 in this program will show you where to  make  the

     change  for this. You'll know when the offset is wrong as  the  re-

     sulting  picture will have the colors scrambled. Change the  offset

     by  one and try again. Note that the TIF file must be 640x480  with

     24-bit color so the  file size is over 920,000 bytes. If it doesn't

     have  that many bytes, it isn't an acceptable format. KY1S  reports

     that  he  has used TIFCONV with the TEMPRA program and  with  PHOTO







     FINISH  3.  WB7PAP found a Graphics Workshop 7.0  on  Rimmer's  BBS

     (tel:[phone])  which converts other formats very well.  It  is

     called GRAFWK70.ZIP. N9ARX reports success using Print Shop Pro for

     Windows.  The  last two are shareware. I believe all  of  these  and

     other  programs accept pictures in lower resolution  formats  which

     can then be converted to the highest resolution format that we use.

     The TIF file must be in a non-compressed format. I use the  RAMDISK

     as temporary storage for this monster file and immediately  convert

     it to a SSTV format for permanent storage.

          Since this program includes the VU program to show you the end

     result,  you must transfer the SYSTEM CONFIGURATION from VU.BAS  to

     lines 2000-3000 in this program.

          Revision  D adds the capability to directly save  any  picture

     generated by this program in either full or compressed formats.



     TGACONV.BAS

          I  found a collection of CD-ROM pictures at Dayton which  were

     of  high quality, 24-bit color, but in a TGA format.  This  program

     converts  any 24-bit color TGA file of any line length equal to  or

     greater  than 640 pixels to SSTV files. Using a simple batch  file,

     STGA.BAT,  you can call up any image on the CD-ROM, convert  it  to

     any  SSTV format and then transmit it in a seamless operation.  The

     images  are filed on the CD-ROM as img2018.tga kind of  files.  You

     merely type: C:\TV\>stga 2018 and the picture is transferred to the

     RAMDISK, the TGACONV program is activated, where you pick the  mode

     and  take the default choices. When you see the picture,  you  exit

     with  Q and then F10. The batch program then moves you into the  VT







     program where you answer the questions and transmit. I suggest  you

     study  the STGA.BAT as you can put together  sequential  operations

     using several of the programs. You could eliminate answering  ques-

     tions  even  by making special versions of say  TGACONV  where  the

     question  lines  are  replaced with the answers you  want  for  the

     particular  sequence,  and it will just run straight  through.  The

     special  versions would be given different names, of  course.  Read

     the  DOS book on batch files. You can do some powerful things  with

     a few little lines.

          TGACONV also has direct SAVE capability.



     TIFLABEL.BAS and PCXLABEL.BAS

          You may have one of the fancier software programs to take  the

     original  TIFF  file and add text titles and labels  to  it  before

     converting  it to the SSTV format. I don't, so a LABEL program  was

     written to give some limited capability in this area. Not having  a

     library of fonts and of different background patterns to paint  the

     letters with, we made the program general purpose enough to use any

     font and any background pattern you can take a camcorder picture of

     or can obtain in a TIFF file format or in a SSTV format. There  are

     a number of interesting patterns right in the room you are  sitting

     in which will add interest to your labels---wicker chair back, tile

     floor,  rug, venetian blinds, etc., etc. These two  program  accept

     TIF  or PCX 1 bit (ie, TWO COLOR) files which might typically  come

     from  a handscanner or one of the many PAINT programs  of  whatever

     font  letters you may choose. Typically you can resize the  letters

     to  fit. The TIF or PCX file  is used as a mask and where ever  the

     file  picture  is black, the LABEL program inserts either  a  color







     which you can choose, or a cutout from any background file  picture

     which  you might have available. You can even use a pix you  copied

     on  the air as a background file. The effect is as if you  cut  the

     letters  out  of the background photograph and pasted  it  on  your

     picture.  You manually select where you paste it on. You can get  a

     three dimension effect by pasting on a color mask, then going  back

     through  LABEL  again, moving the mask a few   pixels  and  pasting

     another  color  or  background file on top of  what's  there.  With

     multiple  passes,  you can get a variety of effects. You  also  can

     make  a square or circular mask and place it to transfer a  face  or

     some  other  feature of one file to paste it over another.  A  mask

     file, K3BC.PCX which is 64 pixels wide, is included for you to  try

     out this program on the SAMPLE.M1 pix. You can use SAMPLE2.94 as  a

     background file. Or rename SAMPLE2 with a different mode suffix for

     a  different background file, etc. I use the RAMDISK to store  suc-

     cessive  passes  thru  LABEL (under PATH/NAMES of  d:a,  d:b,  d:c,

     etc.);  then with SS (SLIDESHO) in drive d: you can view  the  pro-

     gressive results.

          Since  these programs use VU for VUing, you must transfer  the

     SYSTEM CONFIGURATION you settled on in VU.BAS to lines 2000-3000 in

     these  programs.  Some TIF 2-color files don't seem  to  work  with

     these,  but  I have not found a PCX 2-color file  source  yet  that

     doesn't  work (16 color or 256 color files DO NOT work!).  Put  the

     text in the upper left corner of the PAINT screen when you generate

     the  mask. PCXLABEL includes the direct SAVE function in both  full

     and compressed formats.









     3D.BAS

          This is a new program added in  Rev C that creates 3-dimension

     picture  files in our SSTV formats. It is not a synthetic  3D  pro-

     gram,  although  you can get some synthetic depth of  stuff  pasted

     over the picture files with the LABEL program. It uses the original

     "two  camera" approach for 3D. You use your video camera  and  frame

     grabber  to  capture two pictures of a scene, a RIGHT eye  pix  and

     after  moving  the camera a distance about like your  eye  spacing,

     capture a LEFT eye pix. It's hard to keep the pointing angle right,

     so pick some vertical feature in the scene--note where it is in the

     frame  grabber screen w.r.t. the letters below the screen.  On  the

     second pix, adjust the camera pointing to keep the vertical feature

     only a few pixels from where it was on the first pix. The magnitude

     and  direction of this offset can be estimated by closing each  eye

     alternately  and  noting what the apparent shift is from  Right  to

     Left  eye. The final product picture should only show a few  pixels

     of "red shift" as seen without the 3d glasses. If the glasses color

     bandpass matched the phosphor colors of your C.R.T., then the  "red

     shift" would be of no consequence, since you couldn't see it. Until

     color matched glasses become generally available, assume the person

     receiving  your  pix will not be color matched and limit  the  "red

     shift".

          I  store  both of these TIF files in the  RAMDRIVE,  then  use

     TIFCONV  to  obtain two picture files in the  desired  mode,  using

     RIGHT  and LEFT as file names. These feed directly into the  3D.BAS

     program. Use your 3D glasses to see if the result is  satisfactory.

     Slow-scanners  seem enamored of synthetically derived 3D where  you

     take a series of single images and by displacing color frames a few







     pixels  can get discrete depths for the different images used.  You

     can get the same effects using the LABEL programs, but it's labori-

     ous. When you put your call  sign on these 3D pix, you can get  the

     same  effect by pasting the same label on both RIGHT and  LEFT  pix

     but  displaced a few pixels from each other. You can find out  more

     about  this by playing with it. You'll get lots of advice from  the

     synthetic enthusiasts---not all of which is applicable to the clas-

     sic two displaced camera approach which this program is closest to.

          If  you're recording 3D pictures in the field with  your  cam-

     corder, be sure to voice record whether you are in a RIGHT or  LEFT

     eye  position. KB9MC has located a source for 3-D glasses in  Cali-

     fornia; Reel 3-D Enterprises--tel:[phone].



     NEW MODE

          I  considered  adding the Scottie DX mode, but after  studying  it,

     decided it didn't use the extra redundant pixels as efficiently  as

     it might. It basically sends the same pixel value three times in  a

     row  and  then combines these three for a more  noise  free  single

     pixel. This is great for decreasing white noise which is not corre-

     lated  in the three samples. But you only need look at a  few  pic-

     tures  to  see that many of the "noises"  which pollute  our  pixel

     values  (QRM, QRN, multipath, carriers, etc.) are often  correlated

     over 3 successive pixels or more. So I began looking at putting the

     redundancy with time separation between samples--like repeating the

     same  line 3 times in a row. Each pixel would have 3 samples  sepa-

     rated  by one line length in time. The good stability of our  crys-

     tal-locked  receiving   system  would allow us  to  combine  pixels







     separated  in time quite precisely (a system which line  SYNC's  on

     each line using the 1200 cps sync would not combine so precisely!).

     This  would decorrelate the samples against more than  white  noise

     and give better results.

          While  it is generally agreed that there is  more  information

     content in the luminance (B/W) of a picture than in the Chrominance

     (color),  that depends to some extent on a value judgement  between

     spatial  resolution and color content. The engineers would  usually

     opt  for spatial resolution while the artists would probably  value

     color  content.  The more I thought about it, the more  foolish  it

     seemed  to  use completely redundant pixels.  Four  pixels  grouped

     together  in  a square could be exactly the same value and  give  a

     redundancy improvement against noise in preserving that value.  The

     same four pixels could be used to improve the spatial resolution by

     a factor of 2 in each axis, and if conditions were good, you  would

     see  the  resolution improvement. If conditions are bad,  your  eye

     would average those same 4 pixels and get at least as much informa-

     tion  as if they were all sent with the same value. So if  you  are

     going to send more bytes to improve poor condition reception, it  is

     better to use the bytes for better resolution, and let your eyes do

     the  averaging.  Since the signals fade  up and down, you  get  the

     better  resolution when it's up. Also, since most  SSTV  operations

     have  people at different skip distances, some will get  extraordi-

     narily  good high resolution pictures, while the weaker  ones  will

     get  less.  This whole process is illustrated quite well  when  you

     look  at the differences in Martin 1 and Martin 2. Here  the  extra

     resolution is in one axis only, but the difference in results as to

     how  fast  they  deteriorate under worsening  conditions  is  quite







     striking.

          Another  illustration of this is in the FAX480 where the  dou-

     bled resolution in both axes gives very  durable info content under

     quite bad conditions  compared to SSTV modes.

          So  we  decided  to make a full color mode  with  the  doubled

     resolution  in both axes. To avoid start-up problems  with  getting

     everybody  "tuned" to the same line timing, we decided to  use  the

     same  timing as the FAX480. Three successive 546 byte  "lines"  are

     used  for  the Red, Green, and Blue. In fact, it's  nothing  but  a

     double size Scottie or Martin 1 mode--double in both axes.

          It's called mode 96, has a VIS code of &h6A, and uses the same

     LT9 already set for FAX480. In fact you can copy FAX480 pictures in

     this  mode  to  align it (the picture will only  fill  1/3  of  the

     screen). Displaying the extra information in this new mode requires

     a  full 640x480 TRUCOLOR video card---but these are down  to  about

     $70  now (Cirrus ISA boards) and hopefully you have already  bought

     one  to   use TRUSHO in the previous revision. There is  already  a

     large population of hams with the TRUCOLOR cards.

          Some will  question the length of time to send such a picture,

     but  it's  only  a bit longer than the Scottie DX  mode.  Like  the

     Scottie  DX mode, it should only be used sparingly,  and  certainly

     not  often on busy frequencies like 14230. But there are plenty  of

     often unused frequencies on 15 and 10 meters and 75 meters is often

     vacant  during the daylight hours. Also a number of our  users  are

     operating  on VHF bands for local picture trading. When KY1S and  I

     were  evaluating it initially, we shared a frequency with a  strong

     broadcast station on 40 meters, and got surprizingly good  results.







     Even  today we shift  down to Scottie 2 and Martin 2 when the  fre-

     quency  gets busy and go back up when the  conditions  deteriorate.

     This mode is just one more step in that progression. It should be a

     good  match  with the Kodak CD-ROM photo system--and  I'll  make  a

     conversion program choice for Kodak as soon as I can find out their

     format.



     ANOTHER MODE

          We  have  added  a new high resolution  Black  and  White  FAX

     mode---well,  a little bit new. It uses the exact same line  timing

     as  FAX480  and mode 96 Colorfax, and differs only from  FAX480  in

     having  a VIS code attached instead of the lengthy header. We  have

     renamed FAX480 as mode 95, and assigned mode 9 to the B/W FAX.  The

     VIS  code assigned is &HAA (decimal 170). In RT the VIS code   will

     be  recognized  and AUTO operation will be just like all  the  SSTV

     modes. You can still acquire this mode and FAX480 pictures manually

     in RT. In VT  you can choose the VIS-coded mode 9 or use mode 95 to

     put  a  FAX480  header on the picture.  Files  are  interchangeable

     between  these modes and use the same .FAX suffix. When you find  a

     pix which has the detail to justify a mode 96 Colorfax file, make a

     mode 9 B/W fax file so you can send it to folks without Trucolor to

     show  what kind of detail is possible. As described  before,  these

     modes have the 64 gray shades necessary for photo-like quality. The

     resolution is 536x480.



     USER ADDED FEATURES

          Some users of this system have taken advantage of it's  flexi-

     bility to customize it for their own use. The most common modifica-







     tion  is to change the initial screen to remove all those words and

     use the space for a more readable list of the mode choices. Of  the

     ones  I am aware of, KY1S has the most imaginative  one---including

     picture icons, introductory tunes, novel uses of the function keys,

     etc., so some of you may want to get a copy from him to use direct-

     ly  or  to give you ideas for your  own  creations.[At  revision  E

     time,  he  has chained our conversion programs and  VT.BAS  with  a

     shareware  format conversion program called Alchemy to go  directly

     from pix files in the more common GIF, PCX, etc. formats to picture

     transmission in the SSTV format chosen.]

          To make these adders portable, so they can be easily put  into

     later revisions without risking overwriting lines, it would be wise

     to write them with line numbers above 6000. Then you can SAVE  them

     as a BASIC ASCII file, and use the MERGE command to add them to any

     future revision. Incidentally, "all those words" were an attempt to

     force  some instructions on those who never read the  instructions.

     They  are  easily relegated to a HELP screen with a  few  lines  of

     BASIC.  In revision D we've offered a picture menu described in  RT

     above.



     REAL TIME DISPLAY

          Some  have  asked  for real time paint of the  pix  as  it  is

     received. I personally prefer not having it as I use the  receiving

     time  to read the paper, and the incoming pix is  distracting.  But

     many  others  have been used to it. There are tradeoffs.  When  you

     take  the time to paint a line right after it comes in, it adds  to

     the  "edge  of the room" gap we described above (in  RT).  We  have







     chosen to go for maximum resolution (pixels/line) so that  increases

     the gap time over a lower resolution choice. This gap can be hidden

     in  all  but the AVT modes, but only if you copy in SYNC  from  the

     start. We went to considerable trouble to minimize this gap so  you

     can take almost any late start or out of sync picture and scroll it

     into  position with negligible picture degradation. To  our  knowl-

     edge,  no other system offers that capability. Also, we  wanted  to

     accomodate  the  lower speed 286 machines, so the "Store  and  then

     Look"  choice  won the tradeoff. Of  course,  the  state-of-the-art

     moves on, and with such a large population of 486's out there  now,

     we  plan to revisit that tradeoff. Also, the newer  TRUCOLOR  cards

     take  less time to paint a pixel. So if you really want  real-time,

     be  patient. It will only be added, however, if it doesn't  degrade

     the picture quality or the system flexibility. And it clearly  will

     require the faster  computers.

          Revision  E adds this feature---and I just can't get my  paper

     read anymore. Can't resist looking. I have not been able to  detect

     any  deterioration  with the interleaving technique used.  I  would

     welcome  someone on VHF, where there is no noise, copying the  same

     picture  (a  clean, original file) with and  without  the  REALTIME

     enabled, so I can make a more detailed analysis. Please state  what

     computer  processor/speed was used. My VCR is not clean  enough  to

     run a meaningful test.



     EXPERIMENTER'S INFO

          To  simplify  the use and modification of  these  programs  by

     experimeters, some of the conventions used are detailed here. First

     of  all, the file bytes are the raw data  as received. If you  sub-







     tract the number 138 from each of them and XOR the result with  63,

     you  will  get the actual color amplitude for that byte  in  6  bit

     format. Conversely, you can take a picture in 3x6=18 bit format and

     convert each of it's bytes  to the SSTV format by XORing each  byte

     with 63 and adding 138 to the result.

          The  symbol,  LL,  refers to total line  length  (measured  in

     bytes/line) in a picture. This includes sync bytes, picture  bytes,

     etc.

          The symbol L refers to the byte count in a single color  frame

     line. This also is the number of pixels in the color picture. Since

     there  are  3 colors, 3xL would equal to the line  length,  LL,  if

     there were no sync bytes. In the AVT modes, there is no line  sync,

     so there 3xL=LL.

          The  symbol  I is the mode chosen. Or in one  case  where  two

     modes are used in a program, J is used for the second mode.

          The symbol FS is used for frame sync. If this is called out in

     a mode, then a routine is inserted which holds the program  captive

     until  a 1200 hz sync is received before it allows the frame to  be

     copied.  Parameter  DE allows you to adjust the delay  after  frame

     sync before the picture copying begins.

          LT denotes line timing with LT3 being line timing for mode 3.

          D% represents the number of lines in the picture mode.

          DR$ is the letter assigned to the RAMDISK (with quotes, "d").

          SW   is the line sync  window in TIME clock  counts  (approxi-

     mately in microseconds).

          EA  is  the number of bytes used by the line sync when  it  is

     present.  In  the viewing program, this same number is  POKEd  into







     memory  location,  &H3EA, which is where it got it's name.  In  the

     transmit  program it is the bytes thrown away in each line  of  the

     file being transmitted and replaced by sync. It is identical to the

     above except in the non-synchronously copy modes , (11) and (22).

          VC is the video card choice.

          S%  is  the number of bytes a picture was scrolled to  get  it

     aligned  for SAVEing. It is returned as an output when you run  the

     VUing  program.  It  gets converted to it's 4 byte form  as  S!  in

     RT.BAS.

          SC  is used in the transmit program and represents the  number

     of 1/2 cycles of the sync frequency (1200 hz) transmitted  on  each

     line.

          LI  is  used  as the number of picture lines  in  the  TIFCONV

     program. This program also uses Z as the byte value to be  inserted

     in blank bytes. Z=138 is white; Z=249 is black.

          The  files as SAVEd have the picture color bytes placed  in  3

     successive groups of bytes, each group being L bytes long. Then the

     line  sync  bytes (if they are used in the  particular   mode)  are

     tacked on to the end of each line.

          The  DEF  SEG statement in each of these  programs  tells  the

     program  where in memory to place the program. All of the  programs

     are  relocatable,  so  if you find interference  between  some  TSR

     resident program and one of these programs, you may be able to move

     it  up to keep from overwriting the  TSR (typically, this shows  up

     as a computer hang-up which can only be cleared by re-booting). You

     still need enough memory between DEF SEG and &HA000:0000 to  accom-

     modate the pictures you want to copy.

          If  you  find all  of your Robot pictures  are  initially  too







     green or too purple, you can change the number 10 in line 1371  (in

     both  memory  locations)  of RT.BAS to better center  it  for  your

     particular tuning habits.

          If you find that you want to reset the default setting for the

     color  registration  on a mode, you can find out what  the  desired

     values  are  by first using C and X to get what you like  with  the

     VU.BAS  program.  Then when you return to BASIC,  type  PRINT  PEEK

     (&H3EA).  This  will be the new value to poke into &H3EA  for  this

     mode.  Then  PRINT  PEEK(&h15D) and call the value  A.  Then  PRINT

     PEEK(&H15E) and call the value B. Then PRINT PT. The new value  for

     L, the color frame width, then is L = A + 256*B + PT. These  values

     are added back into the program and SAVE'ed. This capability is not

     available in the Robot modes.

          The  color  shades display at the beginning of  the  256  line

     modes  has my call sign (K3BC) superimposed over it. In  order  for

     you  to  change it to your call sign, you'll need to  learn  a  bit

     about how two-color (or 1 bit) images are formed. If, for  example,

     in  the  call sign array of bytes you wanted to make  the  first  8

     pixels overwrite white, you would make the first byte equal to  255

     or in hexidecimal, &hFF (which is 1111,1111 in binary). To paint the

     first  4 pixels out of 8, the first byte would be &hF0  (in  binary

     equals 1111,0000). Look at the lines starting at 780 in the  TX.BAS

     program,  at  line 3430 in VT.BAS, and at line 2400 in  the  RT.BAS

     program to see how to figure out what DATA bytes  to include to get

     your  call sign. Once you determine the DATA array for  your   call

     sign  insert it in all three programs. If you have 6 characters  in

     your call sign it'll be a bit crowded---time for an upgrade!  Don't







     try  to  change  the number of DATA bytes. If you're  too  lazy  to

     figure  this out, you can signal that to the world by changing  all

     the DATA line bytes to 0, which removes the call sign.

          If  you  belatedly decide to save a picture  and  haven't  yet

     overwritten it in D:a.p then you can recover by LOAD "RT" and  then

     typing  a temporary line---- 880 GOTO 890. Now execute the  program

     choosing  the same mode as was used to capture d:a.p.  The  picture

     will  show up and you can scroll it to the proper position and SAVE

     it. You can move any file to d:a.p and re-scroll it this way if you

     don't like the initial choice you made when you saved the file.  Or

     if you want to transfer a part of one file to another using  LABEL,

     use this method to scroll the transfer image to where you want  it,

     and use the new file as a background file.

          In  the  transmit program, each mode has it's VIS  code  poked

     into memory location  &h3A5. The delay (in pixels) between the  end

     of  the VIS code and the beginning of the red frame is  poked  into

     memory  locations  &h523  (LSB)  and &h510  (MSB).  This  delay  is

     switched in by the poke, poke &h52B,227, and is used in S1, S2, and

     WR96. The pokes into &h373, &h374, and  &h375 are to switch a  path

     inside  the  program for modes which have the 16  line  gray  shade

     header.

          A  few  of  you  may have a  SVGA  card  which  only  supports

     640x400x256. You can still view up to 200 lines of picture, but the

     perspective stretches the picture vertically by a factor of 480/400

     in the 240 and 256 line modes. The programs are constructed to  use

     the 640x400 mode for the AVT94 (mode 6) only. There are a number of

     lines  which  state:   IF I=6 THEN POKE &h113 or  &H398  with  some

     number.  If  you take these lines and eliminate the  "IF  I=6"  con-







     straint  you will use the 640x400 for all modes:  specifically  you

     delete the words "IF I=6 THEN"  in each line and leave the POKE. In

     the RT.BAS program, the lines affected are 1765, 1815, 1875,  1925,

     1995,   and   2045.  In  the  VU.BAS,  VT.BAS,   TIFCONV.BAS,   and

     SLIDESHO.BAS  programs  the affected lines are  1590,  1635,  1685,

     1735,  1787, and 1835. In PCXLABEL.BAS and TIFLABEL.BAS  the  lines

     affected  are  1540,  1610, 1680, 1740, 1820,  and  1890.  Strictly

     speaking, you only need to eliminate the constraint for the partic-

     ular  Video Card choice you are using. To my surprise,  a few  sys-

     tems have shown up with 640x480 but not 640x400. If you have  this,

     just  eliminate all of the above lines, and mode 6 will be  painted

     in the 640x480 screen (the easiest way  to eliminate them is to put

     an apostrophe before the first word in the line, leaving the  words

     there in case you get another video card at a later date).



     COPYRIGHT NOTICE

          These programs were developed for the free use of other  Radio

     Amateurs  but  since they might have other commercial use,  I  have

     registered them with the U.S. Copyright Office. Radio Amateurs  are

     free to make copies for other amateurs provided no monetary compen-

     sation  is involved. Any other use for commercial purposes is  for-

     bidden  without specific written permission from the  author.  Note

     that the individual core machine language files are copyrighted  so

     they may not be copied and sold as part of any commercial  venture.

     Shareware  vendors may not reproduce these programs.  Also,  copies

     may not be made to ship with any package for which monetary compen-

     sation is charged. Please help me police this!







                                                       Have fun,

                                                       Ben Vester,    K3BC

     8/1/94