robertl@bucsb.UUCP (Robert La Ferla) (08/17/89)
/* Go ahead, eat my line. */ There has been discussion about how the audio can be improved on the Amiga. One person suggested that CBM provide us with 8 channels of audio. Two channels of stereo are fine. What needs to be improved is the frequency response. It would be nice to recreate any sound from 20hz-20khz. I believe that to do this we must increase the sampling rate. Is this possible? Which chip handles audio? Is it the Paula? Does anyone know the audio specs of the Amiga. For instance, how much stereo seperation is there? Is it possible for the Amiga to generate Dolby encoded Surround Sound? Rec.audio fans speak! __ / \ / __/_ /___/ __ /_ __ __ / / \ '_' /_/ |_- / ' /
swarren@eugene.uucp (Steve Warren) (08/18/89)
In article <3114@bucsb.UUCP> robertl@bucsb.UUCP (Robert LaFerla) writes: >/* Go ahead, eat my line. */ > >There has been discussion about how the audio can be improved on the Amiga. >One person suggested that CBM provide us with 8 channels of audio. Two >channels of stereo are fine. What needs to be improved is the frequency >response. It would be nice to recreate any sound from 20hz-20khz. I >believe that to do this we must increase the sampling rate. Is this Two areas could be improved, the sample rate and the resolution. I think the sample rate is pretty simple to deal with. 20 khz requires a min. of 40 khz sample rate, but in the real world somewhere above 40. That's one sample every < 25 usec. Not exactly taxing on Amiga's bandwidth (I know it could _get_ taxing, but it's not inherently). But 8 bits of resolution means audible quantization noise, which could only be improved by increasing the number of bits. I wish we could get 16 bit sound on the Amiga, it would really sound great. But that would require significant redesign of the sound chip, and I don't know how it could be made compatible with old software. Maybe a dual-mode design that only uses the 8 MSbits of the sound channel for old stuff. Just dreaming :-). --Steve ------------------------------------------------------------------------- {uunet,sun}!convex!swarren; swarren@convex.COM
trantow@csd4.csd.uwm.edu (Jerry J Trantow) (08/19/89)
In article <1542@convex.UUCP> swarren@eugene.UUCP (Steve Warren) writes: >In article <3114@bucsb.UUCP> robertl@bucsb.UUCP (Robert LaFerla) writes: >>There has been discussion about how the audio can be improved on the Amiga. >>One person suggested that CBM provide us with 8 channels of audio. Two >Two areas could be improved, the sample rate and resolution >But 8 bits of resolution means audible quantization noise, which could >only be improved by increasing the number of bits. I wish we could get Don't forget about adding channels (not adding waveforms) and volume modulation. When you use all 8 bits the sample doesn't sound that bad. The dissatisfaction occurs during the quiet parts of the sample when the quantization is relatively large. If you use volume modulation you can adjust for this and use 8 bits keeping the relative quantization error the same thru the sample. Now for my suggestion, The waveform and volume D/As do not change simultaneously which gives a short glitch in the sound when doing volume modulation. How about having a S/H or synchronous action? _____________________________________________________________________________ Jerry J. Trantow | The concern for man and his destiny must always 1560 A. East Irving Place | be the chief interest of all technical effort. Milwaukee, Wi 53202-1460 | Never forget it among your diagrams and equations. (414) 289-0503 | Albert Einstein _____________________________________________________________________________
ltf@attctc.Dallas.TX.US (Lance Franklin) (08/20/89)
In article <1542@convex.UUCP> swarren@eugene.UUCP (Steve Warren) writes: >But 8 bits of resolution means audible quantization noise, which could >only be improved by increasing the number of bits. I wish we could get >16 bit sound on the Amiga, it would really sound great. But that would >require significant redesign of the sound chip, and I don't know how it >could be made compatible with old software. Maybe a dual-mode design >that only uses the 8 MSbits of the sound channel for old stuff. Well, how about adding a "Ham-Mode" for the audio channel... Let's say, perhaps, that we have a 12 (or more) bit D-A on each channel. Each byte, when the channel is in this mode, is interpreted thusly: Bit 7 : Mode bit...when 0, bits 6-0 will be loaded directly into the high order 7 bits, with all low-order bits set to zero. When set to 1, bits 6-0 are interpreted as a signed 7 bit number which is added to the current value of the D-A I beleive the scheme is commonly called Delta-Modulation, although I may be a little off on that...been a while since I've seen the scheme used. Of course, this setup requires that the data be massaged a bit before it gets sent to the channel, but that should not be too horrible, since most samples just get pumped directly from a sample file...just mean adding an extra step when creating the file. Lance -- +-------------------------+ +------------------------------------------+ | Lance T Franklin | | "And all who heard should see them there, | ltf@attctc.DALLAS.TX.US | | And all should cry, Beware! Beware! +-------------------------+ + His flashing eyes, his floating hair!"
murphy@pur-phy (William J. Murphy) (08/21/89)
In article <9051@attctc.Dallas.TX.US> ltf@attctc.Dallas.TX.US (Lance Franklin) writes: >Well, how about adding a "Ham-Mode" for the audio channel... > >Let's say, perhaps, that we have a 12 (or more) bit D-A on each channel. >Each byte, when the channel is in this mode, is interpreted thusly: > > Bit 7 : Mode bit...when 0, bits 6-0 will be loaded directly into the > high order 7 bits, with all low-order bits set to zero. > When set to 1, bits 6-0 are interpreted as a signed 7 bit > number which is added to the current value of the D-A > > >| Lance T Franklin | | "And all who heard should see them there, >| ltf@attctc.DALLAS.TX.US | | And all should cry, Beware! Beware! I don't see the purpose in doing what you suggest. If you have a 12-bit DAC, then with the 16-bit DMA we already have, why do you need to do the audio-HAM mode? I would think that it would just be easier to do 12-bit audio directly. Besides, the Audio-HAM mode would not reduce the storage requirements of a raw audio sample. Rather, what you suggest increases the required space by a bit or two per sample. I do see where if you had a "low-volume" passage of sound (where only the first six bits are used for a period of time), this might save in storage of the sample. The hardware could assume that the bit 7 being set implies that if the immediately following byte were bit 7 off, it would be loaded in the lower 6 bits. byte 1 bit 7 byte 2 bit7 byte 1 loaded as byte 2 loaded as on on high 6bits next sample high 6bits on off high 6bits this sample low 6bits off on low 6bits next sample high 6bits off off low 6bits next sample low 6bits Is this what you were thinking of Lance? William J. Murphy murphy@newton.physics.purdue.edu
swarren@eugene.uucp (Steve Warren) (08/21/89)
In article <9051@attctc.Dallas.TX.US> ltf@attctc.Dallas.TX.US (Lance Franklin) writes: >I beleive the scheme is commonly called Delta-Modulation, although I may be >a little off on that...been a while since I've seen the scheme used. > Yes, but to buy anything with delta modulation you have to increase the sample rate. If the sample rate is the same then you have not increased the highest frequency component in the output wave. But you know that if you are familiar with delta modulation. --Steve ------------------------------------------------------------------------- {uunet,sun}!convex!swarren; swarren@convex.COM
jms@tardis.Tymnet.COM (Joe Smith) (08/25/89)
In article <2450@pur-phy> murphy@newton.physics.purdue.edu.UUCP (William J. Murphy) writes: >I don't see the purpose in doing what you suggest. If you have a 12-bit >DAC, then with the 16-bit DMA we already have, why do you need to do >the audio-HAM mode? I would think that it would just be easier to do >12-bit audio directly. Besides, the Audio-HAM mode would not reduce the >storage requirements of a raw audio sample. Right now we have an 8-bit DAC being driven by an 8-bit data stream. Fibbonacci delta encoding allows driving an 8-bit DAC from a 4-bit stream. With a 12-bit DAC, we could use a 16-bit data stream (wasting 4 bits), a 12-bit stream (optimal), or an 8-bit stream. If most of the changes in the audio signal are less than +/- 63 units per time, using 8 bits per sample wins over using 1.5 bytes per sample. In article <9051@attctc.Dallas.TX.US> ltf@attctc.Dallas.TX.US (Lance Franklin) writes: >Well, how about adding a "Ham-Mode" for the audio channel... >Let's say, perhaps, that we have a 12 (or more) bit D-A on each channel. >Each byte, when the channel is in this mode, is interpreted thusly: > Bit 7 : Mode bit...when 0, bits 6-0 will be loaded directly into the > high order 7 bits, with all low-order bits set to zero. > When set to 1, bits 6-0 are interpreted as a signed 7 bit > number which is added to the current value of the D-A byte 1 bit 7 off = put bits 6-0 into DAC bits 11-5 and zero DAC bits 4-0 byte 1 bit 7 on = bits 6-0 are a number from -64 to +63. Add this value to the 12-bit number currently in the DAC. byte 2 exactly the same as byte 1. End result: 33% fewer bits required to store an audio sample (8 vs 12). Zero distortion if the original audio signal is reasonably well behaved. -- Joe Smith (408)922-6220 | SMTP: JMS@F74.TYMNET.COM or jms@tymix.tymnet.com McDonnell Douglas FSCO | UUCP: ...!{ames,pyramid}!oliveb!tymix!tardis!jms PO Box 49019, MS-D21 | PDP-10 support: My car's license plate is "POPJ P," San Jose, CA 95161-9019 | narrator.device: "I didn't say that, my Amiga did!"