earth.cs.bath.ac.uk

home *** CD-ROM | disk | FTP | other *** search

/ earth.cs.bath.ac.uk / earth.cs.bath.ac.uk.zip / earth.cs.bath.ac.uk / pub / dream / README.Csound < prev next >

Wrap

Text File | 1999-02-15 | 42KB | 1,225 lines

***********NEW: Log of changes introduced from 3.15.10********************** CSOUND for MSDOS CSOUND for ATARI ST John Fitch School of Mathematical Sciences University of Bath Bath BA2 7AY England Tel: +44-1225-826820 FAX: +44-1225-826492 E-mail: jpff@maths.bath.ac.uk or J.P.Fitch@bath.ac.uk (also Codemist Ltd, Tel/FAX: +44-1225-837430) csound_286.zip (old) csound_fpt_286.zip (old) csound_new.zip csound_win.zip csound_src.zip These files are the executables for CSound for 286/386/486 machines running MS-DOS or Windows. The "new" version is supposed to be self adapting to any hardware over 386. Also, the files *.ttp are Csound for the Atari ST. See below for ATARI notes There is a mailing list kept for this version; to join send mail to majordomo@bath.ac.uk with a single line "subscribe pcsound". There is also a Csound mailing list for discussion of any aspect of the system, (joined my mailing majordomo@maths.ex.ac.uk) and a WWW page at http://www.leeds.ac.uk/music/Man/c_front.html Local Changes: ============= It attempts to do graphics. It is supposed to adjust to your graphics system. See below for notes on how to set screen types explicitly. I also have graphics in PVANAL and LPANAL with a -g option. In CSOUND itself there is a pause before and after each graph. This can be turned off if the environment variable CSNOSTOP is set to YES. To ensure you get WAV sound files either use the -W option, or set an environment variable SFOUTYP to WAV in your AUTOEXEC.BAT On Windows95 the output device doc is recognised as immediate output. One needs a fast machine though I have introduced a local version number; currently I have v3.46. The major number refers to MIT's current version (3 is the beta), the 46 is my sequence number for the PC. There are three new utilities to scale for amplitude, a mixer for mixing sound files together, and a mkgraph program to write envelopes for these two. See separate release files for versions 3.44-3.46 Version 3.36: Fixed bug in mixer and a bug in grain Version 3.35: Added granule unit generator, and relaxed restrictions on memory files. Also MAC improvements to dialogs. Version 3.34: Revised WAV support and added 8bit unsigned type Version 3.33: Added a "Play on Exit" option for Windows. Print version number always. On Windows use the registry system to remember parameters, and tidy the code. Decide not to contine the Zortech version. Version 3.31: Various additions, now based on WatcomC (rather than Zortech). Blaster support temporarily withdrawn. Incorporates some Windows code, and consolidation from Unix system. Version 3.28.12: Added convolution operations; see documentation below. Also deferred loads of tables. Version 3.26.12: Fixed bug in granular synthesis Version 3.25.12: Internal only -- not distributed Version 3.24.12: Restored MIDI changes as they all failed to work. Correction to reverb2 made. Another attempt to correct Soundblaster support. The 8-bit seems to work for me; 16-bit hangs still. High speed seems OK. Have not checked stereo. Some shortening of messages, and usage message removed in 286 versions (as Data segment is full). Versions 3.21-3.23 internal; version 3.x.11 was disaster. Version 3.20.10: A number of changes in the MIDI area. A number of new generators added, including Butterworth filters, vdelay, multitap and reverb2. Also granular synthesis generator and stochastic generators. There is an envelope-following generator as well. See Appendix 7 for details of these generators. These are largely the work of Paris Smaragdis. Version 3.19.10 and Version 3.18.10: Not released Version 3.17.10: Created new utility mkgraph which creates envelope files using the mouse for drawing. These files can be used by mixer and scale to provide more flexible gain control on sound files. New utility ENVEXT for create an envelope file from a sound file. Version 3.16.10: Mixer and scale can take envelope files Version 3.15.10: Experimentally I have attempted to read the device sbmidi as a MIDI input for use with the -M option. I have no idea if this works. When using -o devaudio (also can use -o dac or -o sblst) it will force the format to be -c or -s. The mixer can now take varying numbers of channels as input and can include some or all channels, and can direct input channel n to output channel m. As the scaling can be negative as well as positive this incorporates removal of information as well. The syntax is not good, but inspiration is not with me this weekend. Copyright: ========= The systems are the product of the MIT Media Laboratory, and this is their copyright notice: ------------------------------------------------------------------------ Copyright 1986, 1987 by the Massachusetts Institute of Technology. All rights reserved. Developed by Barry L. Vercoe at the Experimental Music Studio, Media Laboratory, M.I.T., Cambridge, Massachusetts, with partial support from the System Development Foundation, and from NSF Grant IRI-8704665. Permission to use, copy, or modify these programs and their documentation for educational and research purposes only and without fee is hereby granted, provided that this copyright and permission notice appear on all copies and supporting documentation. For any other uses of this software, in original or modified form, including but not limited to distribution in whole or in part, specific prior permission from M.I.T. must be obtained. M.I.T. makes no representations about the suitability of this software for any purpose. It is provided "as is" without express or implied warranty. ------------------------------------------------------------------------ The mixer, graphing and SoundBlaster support are probably copyright John ffitch, and I hereby give permission to use, copy, or modify this code for any purpose whatsoever. I would like my name to remain in there, but I do not insist. ------------------------------------------------------------------------ Modifications and new code provided by Robin Whittle are provided without warranty etc. etc. They are copyright 1995 Robin Whittle. Use of this code for non-commercial software products is permitted, provided the source code is made freely available according to the principles of the Free Sofware Foundation, with this copyright notice attached - and those of Barry Vercoe and John Fitch. ------------------------------------------------------------------------ Other parts of Csound have other ownerships. In particular there is copyright code of Codemist (cwin.cpp and elsewhere), Boulanger Music (various opcodes), Allan S C Lee (granule). Others may have claims. Interested parties should note that CSound is a system for creation of sound, and is not a MIDI sequencer. The systems built are described briefly below. CSOUND EXE digital audio processing and sound synthesis csound [flags] orchfile scorefile Csound is an environment in which a "scorefile" or external event sequence can invoke arbitrarily complex signal-processing "instruments" to produce sound. Audio may be displayed during its creation, and the resulting sound sent to an on-line audio device or to an intermediate soundfile for later playback. Flags include -C use Cscore processing of scorefile -I I-time only orch run -n no sound onto disk -i fnam sound input filename -o fnam sound output filename (if fnam is devaudio, dac or sblst use directly) -b N sample frames (or -kprds) per software sound I/O buffer -B N samples per hardware sound I/O buffer -A create an AIFF format output soundfile -W create a WAV format output soundfile -h no header on output soundfile -c 8-bit signed_char sound samples -a alaw sound samples -u ulaw sound samples -s short_int sound samples -l long_int sound samples -f float sound samples -r N orchestra srate override -v verbose orch translation -m N tty message level. Sum of: 1=note amps, 2=out-of-range msg, 4=warnings 8=SB mesages -d suppress all displays -g suppress graphics, use ascii displays -S score is in Scot format -t N use uninterpreted beats of the score, initially at tempo N -L dnam read Line-oriented realtime score events from device 'dnam' -M dnam read MIDI realtime events from device 'dnam' (must be sbmidi) -F fnam read MIDIfile event stream from file 'fnam' -P N MIDI sustain pedal threshold (0 - 128) -R continually rewrite header while writing soundfile (WAV/AIFF) -H print a heartbeat character at each soundfile write -N notify (ring the bell) when score or miditrack is done -T terminate the performance when miditrack is done flag defaults: csound -s -otest -b1024 -B1024 -m7 -P128 EXTRACT.EXE Program for extracting parts of a work. Not tested in PC version HETRO.EXE hetrodyne filter analysis for Csound adsyn module hetro [flags] [fundamental] [filename] hetro takes as input a file containing amplitude samples of some sound over time (it is assumed that the samples are evenly spaced in time) and decomposes that sound into a set of harmonically related sine waves with time varying amplitude and phase. LPCANAL.EXE Paul Lansky's software for linear predictive analysis and pitch tracking, adapted for Csound. lpcanal [-p<n> -i<n> -s<t> -d<t> -o<file> -C<str> -P<frq> -Q<frq>] soundfile lpcanal is the new experimental combination of the old anallpc and ptrack. It performs linear predictive analysis and pitch tracking on monaural 16bit fixed point soundfiles. If a -g flag is used then a graphical display is given of some of the output as it is being computed. PVANAL.EXE Fourier analysis module for Csound PVOC unit generator pvanal [-n frame-size] [-o overlap | -i increment] \ inputSoundFile outputFFTFile pvanal converts a playable sample (a time-domain representation) into a series of short-time Fourier transform (STFT) frames centred at regular points throughout the file (a frequency-domain representation). The output file can then be used as the data for the PVOC unit generator in Csound to generate notes based on the original sample, but with their timescales and pitches arbitrarily and dynamically modified. If a -g flag is used then a graphical display is given of some of the output as it is being computed. SCSORT.EXE Stand-alone sorting of sound files Not tested on PC SNDINFO.EXE Reads the header of a sound file to identify type, duration etc sndinfo soundfile SCALE.EXE As well as doing the same as SNDINFO this utility reports on the maximum amplitude, and can generate a new soundfile with the amplitude scaled by a floating point value. It can also change formats. scale [-flags] soundfile Legal flags are: -o fnam sound output filename -A create an AIFF format output soundfile -W create a WAV format output soundfile -h no header on output soundfile -c 8-bit signed_char sound samples -a alaw sound samples -u ulaw sound samples -s short_int sound samples -l long_int sound samples -f float sound samples -F fpnum amount to scale amplitude -F fname envelope file for scaling -R continually rewrite header while writing soundfile (WAV/AIFF) -H print a heartbeat character at each soundfile write -N notify (ring the bell) when score or miditrack is done flag defaults: scale -s -otest -F 0.0 If scale is 0.0 then reports maximum possible scaling; otherwise scale and generate a new soundfile MIXER.EXE This utility can mix together a number of sound files (up to 20 at present) with different starting times and with scaling on each file. mixer [-flags] soundfile [-flags] soundfile ... Legal flags are: -o fnam sound output filename -A create an AIFF format output soundfile -W create a WAV format output soundfile -h no header on output soundfile -c 8-bit signed_char sound samples -a alaw sound samples -u ulaw sound samples -s short_int sound samples -l long_int sound samples -f float sound samples -F fpnum amount to scale amplitude of next sound file -F fname an envelope file for scaling -R continually rewrite header while writing soundfile (WAV/AIFF) -H print a heartbeat character at each soundfile write -N notify (ring the bell) when score or miditrack is done -S int Sample at which to insert next sound file -T fpnum Time at which to insert next sound file -1 -2 -3 -4 include named channel -^ n m include channel n and output as channel m Defaults are: mixer -s -otest -F1.0 -S0 MIXER can also be used for some echo effects. MKGRAPH.EXE A small utility to creat envelope files for MIXER and SCALE. Type ? when the program is running to get all the controls. mkgraph [-v] [envfile] [-o outname] Default is mkgraph -o newgraph. If an envfile is given it is loaded and can be editted. envfile and outname can be the same. ENVEXT.EXE Given a sound file it creates an envelope file with an approximation to the envelope of the sound file. envext [-w time] [-o file] soundfile Defaults are envext -w 0.25 -o newenv ------------------------------------------------------------------------ Graphics: ======== The graphics is just for the display of waveforms. The full specification of the graphics used says that it tries auto-determining the graphics on the machine. This can sometimes fail, so it reads the environment variable FG_DISPLAY, and if set as below it uses that kind of graphics. Value Type ===== ==== GCAHIRES GCA 640 x 200 x 2 GCAMEDRES GCA 320 x 200 x 4 EGACOLOR EGA 640 x 200 x 16 EGAECD Enhanced EGA 640 x 350 x 16 EGALOWRES EGAMONO EVGAHIRES Everest EVGA board HERCFULL Hercules 2 pages 2 colour HERCHALF Hercules 1 page 2 colour ORCHIDPROHIRES VGA type PARADISEHIRES VGA type TOSHIBA Toshiba 3100 -- 640 x 400 x 2 TRIDENTHIRES Trident 800 x 600 x 16 VEGAVGAHIRES Video 7 vega VGA board VESA6A VESA mode 0x6a VESA2 VESA mode 0x102 VGA11 IBM VGA mode 0x11 VGA12 IBM VGA mode 0x12 VGA13 IBM VGA mode 0x13 8514A IBM 8514A display adapter The csound_new version uses a different graphics system which seems more reliable. Virtual Memory: ============== The system uses virtual memory on 386/486. The limits on memory size are the minimum of 1. Free disk space + code size 2. 256 times your extended memory 3. 3.5Gbytes (!) You should set up the environment variable TMP or TEMP to the disk to use for swap space. If this is not set it looks at disks C:, D:, ... looking for the largest free space. That gives the limit of space. This is by the Flashtek Dos extender. The csound_new version used the DOS4GW DOS extender, which I do not think supports virtual memory. Reporting Bugs: ============== Please mail (or possibly FAX) me reports on any bugs and shortcomings of the PC version. I will endeavour to fix or assist, but it is only fair to warn you that this is not either of my jobs, and so it may be lower in priorities. But I am interested in widening the availability of CSound. The csound_new version is built with WatcomC v10. 80286 - version =============== The files in the 286 versions also need the program ZPM.EXE, which is provided, in your search path. Windows ======= The version in csound_win.zip is a Windows version with suitable dialog boxes. It is built from the same sources. It needs WindowsNT, Windows95 or the Win32s system. John Fitch School of Mathematical Sciences Codemist Ltd University of Bath "Alta", Horsecombe Vale Bath BA2 7AY Combe Down, Bath BA2 5QR United Kingdom United Kingdom Tel: +44-1225-826820 Tel: +44-1225-837430 FAX: +44-1225-826492 FAX: +44-1225-837430 ------------------------------------------------------------------------ Appendix 7 : Newest Csound opcodes by Paris Smaragdis Berklee College of Music This appendix describes recent additions to Csound. These additions include a granular synthesis synthesizer, a new set of filters, a new variable delay, a multitap delay, a new reverb, an envelope follower, various noise generators, a power function generator and two gen routines, GEN20 and GEN21. 1) Granular synthesizer. ar grain xamp, xpitch, xdens, kampoff, kpitchoff, kgdur, igfn, iwfn, imgdur Generates granular synthesis textures. INITIALIZATION igfn, igdur - igfn is the ftable number of the grain waveform. This can be just a sine wave or a sampled sound of any length. Each grain will start from a random table position and sustain for igdur seconds. iwfn - Ftable number of the amplitude envelope used for the grains (see also GEN20). imgdur - Maximum grain duration in seconds. This the biggest value to be assigned on kgdur. PERFORMANCE xamp - Total amplitude of the sound. xpitch - Grain frequency in cps. xdens - Density of grains measured in grains per second. If this is constant then the output is synchronous granular synthesis, very similar to fof. If xdens has a random element (like added noise), then the result is more like asynchronous granular synthesis. kampoff - Maximum amplitude deviation from kamp. This means that the maximum amplitude a grain can have is kamp + kampoff and the minimum is kamp. If kampoff is set to zero then there is no random amplitude for each grain. kpitchoff - Maximum pitch deviation from kpitch in cps. Similar to kampoff. kgdur - Grain duration in seconds. The maximum value for this should be declared in imgdur. If kgdur at any point becomes greater than imgdur, it will be truncated to imgdur. 2) Butterworth filters. ar butterhp asig, kfreq ar butterlp asig, kfreq ar butterbp asig, kfreq, kband ar butterbr asig, kfreq, kband Implementations of second-order hipass, lopass, bandpass and bandreject Butterworth filters. PERFORMANCE These new filters are butterworth second-order IIR filters. They are slightly slower than the original filters in Csound, but they offer an almost flat passband and very good precision and stopband attenuation. asig - Input signal to be filtered. kfreq - Cuttoff or center frequency for each of the filters. kband - Bandwidth of the bandpass and bandreject filters. EXAMPLE asig rand 10000 ; White noise signal alpf butterlp asig, 1000 ; cutting frequencies above1K ahpf butterhp asig, 500 ; passing frequencies above 500Hz abpf butterbp asig, 2000, 100 ; passing only 1950 to 2050 Hz abrf butterbr asig, 4500, 200 ; cutting only 4400 to 4600 Hz 3) Vdelay ar vdelay asig, adel, imaxdel This is an interpolating variable time delay, it is not very different from the existing implementation (deltapi), it is only easier to use. INITIALIZATION imaxdel - Maximum value of delay in samples. If adel gains a value greater than imaxdel it is folded around imaxdel. This should not happen. PERFORMANCE With this unit generator it is possible to do Doppler effects or chorusing and flanging. asig - Input signal. adel - Current value of delay in samples. Note that linear functions have no pitch change effects. Fast changing values of adel will cause discontinuities in the waveform resulting noise. Example f1 0 8192 10 1 ims = 100 ; Maximum delay time in msec a1 oscil 10000, 1737, 1 ; Make a signal a2 oscil ims/2, 1/p3, 1 ; Make an LFO a2 = a2 + ims/2 ; Offset the LFO so that it is positive a3 vdelay a1, a2, ims ; Use the LFO to control delay time out a3 Two important points here. First, the delay time must be always positive. And second, even though the delay time can be controlled in k-rate, it is not advised to do so, since sudden time changes will create clicks. 4) Multitap delay ar multitap asig, itime1, igain1, itime2, igain2 . . . Multitap delay line implementation. INITIALIZATION The arguments itime and igain set the position and gain of each tap. The delay line is fed by asig. Example: a1 oscil 1000, 100, 1 a2 multitap a1, 1.2, .5, 1.4, .2 out a2 This results in two delays, one with length of 1.2 and gain of .5, and one with length of 1.4 and gain of .2. 5) Reverb2 ar reverb2 asig, ktime, khdif This is a reverberator consisting of 6 parallel comb-lowpass filters being fed into a series of 5 allpass filters. PERFORMANCE The input signal asig is reverberated for ktime seconds. The parameter khdif controls the high frequency diffusion amount. The values of khdif should be from 0 to 1. If khdif is set to 0 the all the frequencies decay with the same speed. If khdif is 1, high frequencies decay faster that lower ones. Example: a1 oscil 10000, 100, 1 a2 reverb2 a1, 2.5, .3 out a1 + a2 * .2 This results in a 2.5 sec reverb with faster high frequency attenuation. 6) Envelope follower kr follow asig, idt Envelope follower unit generator. INITIALIZATION idt - This is the period, in seconds, that the average amplitude of asig is reported. If the frequency of asig is low then idt must be large (more than half the period of asig ) PERFORMANCE asig - This is the signal from which to extract the envelope. Example k1 line 0, p3, 30000 ; Make k1 a simple envelope a1 oscil k1, 1000, 1 ; Make a simple signal using k1 ak1 follow a1, .02 ; ak1 is now like k1 a2 oscil ak1, 1000, 1 ; Make a simple signal using ak1 out a2 ; Both a1 and a2 are the same To avoid zipper noise, by discontinuities produced from complex envelope tracking, a lowpass filter could be used, to smooth the estimated envelope. 7) Noise generators All of the following opcodes operate in i-, k- and a-rate. The output rate depends on the first letter of the opcode, a for a-rate, k for k- rate and i for i-rate. xlinrand krange - Linear distribution random number generator. krange is the range of the random numbers [0 - krange). Outputs only positive numbers. xtrirand krange - Same as above only ouputs both negative and positive numbers. xexprand krange - Exponential distribution random number generator. krange is the range of the random numbers [0 - krange). Outputs only positive numbers. xbexprnd krange - Same as above, only extends to negative numbers too with an exponential distribution. xcauchy kalpha - Cauchy distribution random number generator. kalpha controls the spread from zero (big kalpha => big spread). Outputs both positive and negative numbers. xpcauchy kalpha - Same as above, ouputs positive numbers only. xpoisson klambda - Poisson distribution random number generator. klambda is the mean of the distribution. Outputs only positive numbers. xgauss krange - Gaussian distribution random number generator. krange is the range of the random numbers (-krange - 0 - krange). Outputs both positive and negative numbers. xweibull ksigma, ktau - Weibull distribution random number generator. ksigma scales the spread of the distribution and ktau, if greater than one numbers near ksigma are favored, if smaller than one small values are favored and if t equals 1 the distribution is exponential. Outputs only positive numbers. xbeta krange, kalpha, kbeta - Beta distribution random number generator. krange is the range of the random numbers [0 - krange). If kalpha is smaller than one, smaller values favor values near 0. If kbeta is smaller than one, smaller values favor values near krange. If both kalpha and kbeta equal one we have uniform distribution. If both kalpha and kbeta are greater than one we have a sort of gaussian distribution. Outputs only positive numbers. For more detailed explanation of these distributions, see: 1. C. Dodge - T.A. Jerse 1985. Computer music. Schirmer books. pp.265 - 286 2. D. Lorrain. "A panoply of stochastic cannons". In C. Roads, ed. 1989. Music machine . Cambridge, Massachusetts: MIT press, pp. 351 - 379. Examples: a1 atrirand 32000 ; Audio noise with triangle distribution k1 kcauchy 10000 ; Control noise with Cauchy dist. i1 ibetarand 30000, .5, .5 ; i-time random value, beta dist. 8) Power functions ir ipow iarg, kpow kr kpow karg, kpow, [inorm] ar apow aarg, kpow, [inorm] Computes xarg to the power of kpow and scales the result by inorm. INITIALIZATION inorm - The number to divide the result (default to 1). This is especially useful if you are doing powers of a- or k- signals where samples out of range are extremely common! iarg - If you are using ipow this is the base. PERFORMANCE karg - If you are using kpow this is the base. aarg - If you are using apow this is the base. EXAMPLES: 1. i2t2 ipow 2,2 ; Computes 2^2. 2. kline line 0, 1, 4 kexp kpow kline, 2, 4 This feeds a linear function to kpow and scales that to the line's peak value. The output will be an exponential curve with the same range as the input line. 3. iamp ipow 10, 2 a1 oscil iamp, 100, 1 a2 apow a1, 2, iamp out a2 This will output a sine with its negative part folded over the amp axis. The peak value will be iamp = 10^2 = 100. 9) GEN20 This subroutine generates functions of different windows. These windows are usually used for spectrum analysis or for grain envelopes. f # time size 20 window max opt size - number of points in the table. Must be a power of 2 ( + 1). window - Type of window to generate. 1 - Hamming 2 - Hanning 3 - Bartlett ( triangle) 4 - Blackman ( 3 - term) 5 - Blackman - Harris ( 4 - term) 6 - Gaussian 7 - Kaiser 8 - Rectangle 9 - Sinc max - For negative p4 this will be the absolute value at window peak point. If p4 is positive or p4 is negative and p6 is missing the table will be post-rescaled to a maximum value of 1. opt - Optional argument required by the Kaiser window. Examples: f 1 0 1024 20 5 This creates a function which contains a 4 - term Blackman - Harris window with maximum value of 1. f 1 0 1024 -20 2 456 This creates a function that contains a Hanning window with a maximum value of 456. f 1 0 1024 -20 1 This creates a function that contains a Hamming window with a maximum value of 1. f 1 0 1024 20 7 1 2 This creates a function that contains a Kaiser window with a maximum value of 1. The extra argument specifies how `open' the window is, for example a value of 0 results in a rectangular window and a value of 10 in a Hamming like window. 10) GEN21 This generates tables of different random distributions. (see also noise generators, above). f # time size 21 type lvl arg1 arg2 Time and size are the usual Gen function arguments. Type defines the distribution to be used. 1 - Uniform 2 - Linear 3 - Triangular 4 - Exponential 5 - Biexponential 6 - Gaussian 7 - Cauchy 8 - Positive Cauchy 9 - Beta 10 - Weibull 11 - Poison Of all these cases only 9 (Beta) and 10 (Weibull) need extra arguments. Beta needs two arguments and Weibull one. Examples: f1 0 1024 21 1 ; Uniform (white noise) f1 0 1024 21 6 ; Gaussian f1 0 1024 21 9 1 1 2 ; Beta (note that level precedes arguments) f1 0 1024 21 10 1 2 ; Weibull All of the above additions were designed by the author between May and December 1994, under the supervision of Dr. Richard Boulanger. This appendix was written on 20 December 1994 by Paris Smaragdis, Berklee College of Music. Internet: psmaragdis@aol.com ------------------------------------------------------------------------ AUTHOR: Greg Sullivan, sullivan@aussie.enet.dec.com (Based on algorithm given in 'Elements Of Computer Music', by F. Richard Moore. CVANAL - Impulse Response Fourier Analysis for CONVOLVE operator csound -U cvanal [flags] infilename outfilename cvanal converts a soundfile into a single Fourier transform frame. The output file can be used by the CONVOLVE operator to perform Fast Convolution between an input signal and the original impulse response. Analysis is conditioned by the flags below. A space is optional between the flag and its argument. -s<rate> sampling rate of the audio input file. This will over-ride the srate of the soundfile header, which otherwise applies. If neither is present, the default is 10000. -c<channel> channel number sought. If omitted, the default is to process all channels. If a value is given, only the selected channel will be processed. -b<begin> beginning time (in seconds) of the audio segment to be analysed. The default is 0.0 -d<duration> duration (in seconds) of the audio segment to be analysed. The default of 0.0 means to the end of the file. EXAMPLE cvanal asound cvfile will analyse the soundfile "asound" to produce the file "cvfile" for the use with CONVOLVE. HINT: To use data that is not already contained in a soundfile, a soundfile converter that accepts text files may be used to create a standard audio file. E.g, the .DAT format for SOX. This is useful for implementing FIR filters. FILES The output file has a special CONVOLVE header, containing details of the source audio file. The analysis data is stored as 'float', in rectangular (real/imaginary) form. ***NOTE***: The analysis file is NOT system independent! Ensure that the original impulse recording/data is retained. If/when required, the analysis file can be recreated. AUTHOR: Greg Sullivan, sullivan@aussie.enet.dec.com (Based on algorithm given in 'Elements Of Computer Music', by F. Richard Moore. ------------------------------------------------------------------------------ CONVOLVE unit generator: ar1[,ar2[,ar3[,ar4]]] convolve ain,ifilcod,channel Output is the convolution of signal ain and the impulse response contained in ifilcod. Note that it is considerably more efficient to use one instance of the operator when processing a mono input to create stereo, or quad, outputs. INITIALISATION ifilcod - integer or character-string denoting an impulse response data file. An integer denotes the suffix of a file convolve.m; a character string (in double quotes) gives a filename, optionally a full pathname. If not a fullpath, the file is sought first in the the current directory, then in the one given by the environment variable SADIR (if defined). The data file contains the Fourier transform of an impulse response. Memory usage depends on the size of the data file, which is read and held entirely in memory during computation, but which is shared by multiple calls. channel - integer denoting the channel of the impulse response to be used for the convolution. 0 (the default) means to use all channels. For multi-channel output, the number of channels in the impulse response must match the number of output signals. PERFORMANCE CONVOLVE implements Fast Convolution. The output of this operator is delayed with respect to the input. The following formulas should be used to calculate the delay: For (1/kr) <= IRdur: Delay = ceil(IRdur * kr) / kr For (1/kr) > IRdur: Delay = IRdur * ceil(1/(kr*IRdur)) Where: kr = Csound control rate IRdur = duration, in seconds, of impulse response ceil(n) = smallest integer not smaller than n One should be careful to also take into account the initial delay, if any, of the impulse response. For example, if an impulse response is created from a recording, the soundfile may not have the initial delay included. Thus, one should either ensure that the soundfile has the correct amount of zero padding at the start, or, preferably, compensate for this delay in the orchestra. (the latter method is more efficient). To compensate for the delay in the orchestra, _subtract_ the initial delay from the result calculated using the above formula(s), when calculating the required delay to introduce into the 'dry' audio path. For typical applications, such as reverb, the delay will be in the order of 0.5 to 1.5 seconds, or even longer. This renders the current implementation unsuitable for real time applications. It could conceivably be used for real time filtering however, if the number of taps is is small enough. Example: - Create frequency domain impulse response file: c:\> csound -Ucvanal l1_44.wav l1_44.cv - Determine duration of impulse response. For high accuracy, determine the number of sample frames in the impulse response soundfile, and then compute the duration with: duration = (sample frames)/(sample rate of soundfile) This is due to the fact that the SNDINFO utility only reports the duration to the nearest 10ms. If you have a utility that reports the duration to the required accuracy, then you can simply use the reported value directly. c:\> sndinfo l1_44.wav length = 60822 samples, sample rate = 44100 Duration = 60822/44100 = 1.379s. - Determine initial delay, if any, of impulse response. If the impulse response has not had the initial delay removed, then you can skip this step. If it has been removed, then the only way you will know the initial delay is if the information has been provided separately. For this example, let's assume that the initial delay is 60ms. (0.06s) - Determine the required delay to apply to the dry signal, to align it with the convolved signal: If kr = 441: 1/kr = 0.0023, which is <= IRdur (1.379s), so: Delay1 = ceil(IRdur * kr) / kr = ceil(608.14) / 441 = 609/441 = 1.38s Accounting for the initial delay: delay2 = 0.06s Total delay = delay1 - delay2 = 1.38 - 0.06 = 1.32s - Create .orc file, e.g: ----CUT---- ; Simple demonstration of CONVOLVE operator, to apply reverb. sr = 44100 kr = 441 ksmps = 100 nchnls = 2 instr 1 imix = 0.22 ; Wet/dry mix. Vary as desired. ; NB: 'Small' reverbs often require a much higher ; percentage of wet signal to sound interesting. 'Large' ; reverbs seem require less. Experiment! The wet/dry mix is ; very important - a small change can make a large difference. ivol = 0.9 ; Overall volume level of reverb. May need to adjust ; when wet/dry mix is changed, to avoid clipping. idel = 1.32 ; Required delay to align dry audio with output of convolve. ; This can be automatically calculated within the orc file, ; if desired. adry soundin "anechoic.wav" ; input (dry) audio awet1,awet2 convolve adry,"l1_44.cv" ; stereo convolved (wet) audio adrydel delay (1-imix)*adry,idel ; Delay dry signal, to align it with ; convolved signal. Apply level ; adjustment here too. outs ivol*(adrydel+imix*awet1),ivol*(adrydel+imix*awet2) ; Mix wet & dry signals, and output endin ---CUT---- The granule unit generator (Allan Lee) is more complex than grain (above), but does add new possibilities. This is a shorten manual, without the pictures, NAME granule - Granular synthesis unit generator for Csound. SYNOPSIS granule xamp ivoice iratio imode ithd ifn ipshift igskip \ igskip_os ilength kgap igap_os kgsize igsize_os iatt idec [iseed] \ [ipitch1] [ipitch2] [ipitch3] [ipitch4] [ifnenv] DESCRIPTION granule is a Csound unit generator which employs a wavetable as input to produce granularly synthesized audio output. Wavetable data may be generated by any of the gen subroutines such as gen01 which reads an audio data file into a wavetable. This enable a sampled sound to be used as the source for the grains. Up to 128 voices are implemented internally.The maximum number of voices can be increased by redefining the variable MAXVOICE in the grain4.h file. granule has a build-in random number generator to handle all the random offset parameters. Thresholding is also implemented to scan the source function table at initialization stage. This facilitates features such as skipping silence passage between sentences. The characteristics of the synthesis are controlled by 22 parameters. xamp is the amplitude of the output and it can be either audio rate or control rate variable. All parameters with prefix i must be valid at Init time, parameters with prefix k can be either control or Init values. SUMMARY OF PARAMETERS xamp - amplitude. ivoice - number of voices. iratio - ratio of the speed of the gskip pointer relative to output audio sample rate. eg. 0.5 will be half speed. imode - +1 grain pointer move forward (same direction of the gskip pointer), -1 backward (oppose direction to the gskip pointer) or 0 for random. ithd - threshold, if the sampled signal in the wavetable is smaller then ithd, it will be skipped. ifn - function table number of sound source. ipshift - pitch shift control. If ipshift is 0, pitch will be set randomly up and down an octave. If ipshift is 1, 2, 3 or 4, up to four different pitches can be set amount the number of voices definded in ivoice. The optional parameters ipitch1, ipitch2, ipitch3 and ipitch4 are used to quantify the pitch shifts. igskip - initial skip from the beginning of the function table in sec. igskip_os - gskip pointer random offset in sec, 0 will be no offset. ilength - length of the table to be used starting from igskip in sec. kgap - gap between grains in sec. igap_os - gap random offset in % of the gap size, 0 gives no offset. kgsize - grain size in sec. igsize_os - grain size random offset in % of grain size, 0 gives no offset. iatt - attack of the grain envelope in % of grain size. idec - decade of the grain envelope in % of grain size. [iseed] - optional, seed for the random number generator, default is 0.5. [ipitch1], [ipitch2], [ipitch3], [ipitch4] - optional, pitch shift parameter, used when ipshift is set to 1, 2, 3 or 4. Time scaling technique is used in pitch shift with linear interpolation between data points. Default value is 1, the original pitch. EXAMPLE Listing of orchestra file. sr = 44100 kr = 4410 ksmps = 10 nchnls = 2 instr 1 ; k1 linseg 0,0.5,1,(p3-p2-1),1,0.5,0 a1 granule p4*k1,p5,p6,p7,p8,p9,p10,p11,p12,p13,p14,p15,p16,p17,p18,p19, p20,p21,p22,p23,p24 a2 granule p4*k1,p5,p6,p7,p8,p9,p10,p11,p12,p13,p14,p15,p16,p17,p18,p19, p20+0.17,p21,p22,p23,p24 outs a1,a2 endin Listing of score file. ; f statement read sound file sine.aiff in the SFDIR directory into f-table 1 f 1 0 524288 1 "sine.aiff" 1 0 i1 0 10 2000 64 0.5 0 0 1 4 0 0.005 10 0.01 50 0.02 50 30 30 0.39 1 1.42 0.29 2 e The above example reads a sound file called sine.aiff into wavetable number 1 with 524,288 samples. It generates 10 seconds of stereo audio output using the wavetable. In the orchestra file, all parameters required to control the synthesis are passed from the score file. A linseg function generator is used to generate an envelope with 0.5 second of linear attack and decade. Stereo effect is generated by using different seeds for the two granule function calls. In the example, 0.17 is added to p20 before passing into the second granule call to ensure that all of the random offset events are different from the first one. In the score file, the parameters are interpreted as: p5 (ivoice) the number of voices is set to 64 p6 (iratio) is set to 0.5, it scan the wavetable at half of the speed of the audio output rate p7 (imode) is set to 0, the grain pointer only move forward p8 (ithd) is set to 0, skipping the thresholding process p9 (ifn) is set to 1, function table number 1 is used p10 (ipshift) is set to 4, four different pitches are going to be generated p11 (igskip) is set to 0 and p12 (igskip_os) is set to 0.005, no skipping into the wavetable and a 5 mSec random offset is used p13 (ilength) is set to 10, 10 seconds of the wavetable is to be used p14 (kgap) is set to 0.01 and p15 (igap_os) is set to 50, 10 mSec gap with 50% random offset is to be used p16 (kgsize) is set to 0.02 and p17 (igsize_os) is set to 50, 20 mSec grain with 50% random offset is used p18 (iatt) and p19 (idec) are set to 30, 30% of linear attack and decade is applied to the grain p20 (iseed) seed for the random number generator is set to 0.39 p21 - p 24 are pitches set to 1 which is the original pitch, 1.42 which is a 5th up, 0.29 which is a 7th down and finally 2 which is an octave up. Csound is developed by Barry L. Vercoe at the Experimental Music Studio, Media Laboratory, M.I.T., Cambridge, Massachusetts. ------------------------------------------------------------------------ ATARI Csound ============ The files are csound.ttp extract.ttp hetro.ttp lpanal.ttp pvanal.ttp scale.ttp scsort.ttp sndinfo.ttp I do not have any of the standard compression programs for the Atari at present, so these are raw binary files. This is an initial port of 3.14, with no support for graphics or MIDI. The code is not yet optimised, and has only been subjected to limited testing. If there is sufficient interest I will optimise and extend, but my Atari is currently failing to boot. The code does not assume the existence of a floating point co- processor, and is built for the 68000 (lowest common denominator). Please report any bugs or comments to me. I am using LatticeC and the sources and scripts are available from me, assuming I can read the Atari disk... There is a better-supported commercially available Atari Csound from CDP. They also have a large suite of music synthesis programs on Atari and PC. Contact Tom Endrich: tendrich@cix.compulink.co.uk Tel: +44-1904-613299 Composers' Desktop Project 11 Kilburn Road York YO1 4DF England