home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.atari.st.tech
- Path: sparky!uunet!mcsun!fuug!funic!nntp.hut.fi!nntp!eye
- From: eye@lesti.hut.fi (Petri M Kuittinen)
- Subject: Re: Sound Chip: What are the volume levels?
- In-Reply-To: warwick@cs.uq.oz.au's message of 20 Jul 92 23:41:29 GMT
- Message-ID: <EYE.92Jul23122240@lesti.hut.fi>
- Sender: usenet@nntp.hut.fi (Usenet pseudouser id)
- Nntp-Posting-Host: lesti.hut.fi
- Organization: Helsinki University of Technology, Finland
- References: <9466@uqcspe.cs.uq.oz.au>
- Date: 23 Jul 92 12:22:40
- Lines: 79
-
- According to the Abacus Internals book, the volume levels on the Yamaha sound
- chip in the ST uses "logarithmically raised" volume levels.
-
- I'd like to know exactly what the volume levels are, in order to calculate a
- better "magic table" for my sampled sound player. I've looked at the tables
- used in some programs, and they don't seem too accurate according to some of
- the calculations I've done - but I don't know how they came up with them.
-
- I have also noticed the same problem, I looked through many sample tables I
- got. After several weeks, I finally came up with a solution, which
- seems right (please don't flame me if I am wrong!). The accurate volume levels
- seem to be:
-
- Sound chip Volume Real Volume (down from max.)
-
- 15 -0 dB
- 14 -3 dB
- 13 -6 dB
- 12 -9 dB
- 11 -12 dB
- 10 -15 dB
- 9 -18 dB
- 8 -21 dB
- 7 -24 dB
- 6 -27 dB
- 5 -30 dB
- 4 -33 dB
- 3 -36 dB
- 2 -39 dB
- 1 -42 dB
- 0 -45 dB
-
- So a change of 1 in sound chip volume value will change the real volume by
- 3 desibels. And desibels of course are calculated like this:
-
- dB=10*lg(I/10^-12 W/m^2)
-
- where lg is 10-based logarithm and I is intensity of sound.
-
- With one sound channel you can achieve maximum of 45dB dynamics, but sound
- quality will be poor. With combined 3 channels you will get max. 49.8 dB
- dynamics (10*lg(3*10^4.5=49.7712...). Most sample players use all the sound
- channels to get more accurate simulation of the 256 8bit-sample values. I
- don't know how the freq. values (in sound chip register 0-5) affect the volume,
- but I guess with the use of them one could achieve even better dynamics.
- (Normally sample players on ST, set the frequency-registers to zero and set
- all sound channel off -> mixer register).
-
- If someone can find out what the "logarithmically raised volume levels" are,
- I can post the be-all-and-end-all set of "magic tables", so we can all
- benefit.
-
- Yes, post the table! It would be nice if you would also calculate a table for
- 1 sound channel samples (eg. many ST-musicians use only one channel to produce
- digi-drum sounds).
-
- I tried a guess of sqrt(2)^V, and it seems pretty good - maybe even better
- than the tables I was using before.
-
- With the program I've written, I can produce "magic tables" of varying
- characteristics - eg. one to play softer-recorded sounds at increased
- volume without first doing an information-losing scale-up. The Sun Sparc
- "ulaw" format could be trivially supported by this program too - it should
- give quite EXCELLENT reproduction of those sampled sounds - probably even
- better than is possible with the STe/TT DMA sound hardware.
-
- All I need is the coefficient of the "logarithmically raised" volume levels.
-
- When you calculate the "magic" table, try to get the the most silent sample
- values correct, human ear becomes more inaccurate with louder values and it's
- not so important to get them correct.
-
-
- The <(O)> Eye
- --
- | The Eye of Brainwash Company | LEA POINT(PC),A3 ;Don't try |
- | E-mail:eye@niksula.hut.fi | ADD.L #$465707B7,(A3) ;this M68000|
- | S-mail:Timpurinkuja 1 C 39 | LEA $8000,A4 ;assembler |
- | 02600 Espoo, Finland |POINT: ORI.W #$5945,D5 ;program!! |
-