title should read multibit*
Hello,
I'm trying to find multibit d/a converters (not CDP) from the 90s to modify to accept 96k direct.
So far from what i've found most consumer D/A converters at this time were limited to 48k. 96k input 'wasn't necessary' until 1998 by which time newer 192k delta sigma hybrids had already emerged and 96k lights started appearing on pro applications.
For example, Apogee mini-DAC and MSB Link dac I feature 96k input, but are not multibit designs.
I find this is odd because it would suggest that recording studios with DAW in the 90s did not record/monitor beyond 48k? I can't find devices accepting anything beyond.
At first my search was refined to devices with PCM63 but realiseing how scarce this is I then widened scope to other multibit chips.
As long as it's multibit right?
FWIR earlier Yamaha products were popular early 90s because of the 63, but I struggle to find any with digital input (nevermind 96k).
Scary pricetags exempt 'multibit audiophile' designs made after 2001 from my search also.
Device shortlist so far (open to suggestions or alternatives).
- Audio Alchemy Dac in a Box (converter: AD1860) (PCB pic)
- Denon DAC-500 (converter: PCM1702)
Audio Alchemy is a good contender - plenty on ebay. portable too, but would modding one to accept 96 be possible/viable?
There's also diy kits, noteably
Armonia PCM63 DAC 2.0
S-Audio DIY (not sure if any of these are multibit or 96).
I'm aware every multibit made can accept upto 384khz directly, but was rarely used because it was always 44.1, which needed upsampling.
Therefore a multibit recieving 96k isn't a problem if relevant filter is bypassed.
I note SPDIF (optical) specification is 192k 20bit maximum.
Rendering 24bit multibit un-optional.
I'd like to ask aswell about the lack of 48k+ products in the early 90s? (even pro field?) unless I'm missing something. - Ive only conducted research using 'complete' d/a converter list' on dutchaudioclassics.nl and this seems rather odd... unproffesional.
Hello,
I'm trying to find multibit d/a converters (not CDP) from the 90s to modify to accept 96k direct.
So far from what i've found most consumer D/A converters at this time were limited to 48k. 96k input 'wasn't necessary' until 1998 by which time newer 192k delta sigma hybrids had already emerged and 96k lights started appearing on pro applications.
For example, Apogee mini-DAC and MSB Link dac I feature 96k input, but are not multibit designs.
I find this is odd because it would suggest that recording studios with DAW in the 90s did not record/monitor beyond 48k? I can't find devices accepting anything beyond.
At first my search was refined to devices with PCM63 but realiseing how scarce this is I then widened scope to other multibit chips.
As long as it's multibit right?
FWIR earlier Yamaha products were popular early 90s because of the 63, but I struggle to find any with digital input (nevermind 96k).
Scary pricetags exempt 'multibit audiophile' designs made after 2001 from my search also.
Device shortlist so far (open to suggestions or alternatives).
- Audio Alchemy Dac in a Box (converter: AD1860) (PCB pic)
- Denon DAC-500 (converter: PCM1702)
Audio Alchemy is a good contender - plenty on ebay. portable too, but would modding one to accept 96 be possible/viable?
There's also diy kits, noteably
Armonia PCM63 DAC 2.0
S-Audio DIY (not sure if any of these are multibit or 96).
I'm aware every multibit made can accept upto 384khz directly, but was rarely used because it was always 44.1, which needed upsampling.
Therefore a multibit recieving 96k isn't a problem if relevant filter is bypassed.
I note SPDIF (optical) specification is 192k 20bit maximum.
Rendering 24bit multibit un-optional.
I'd like to ask aswell about the lack of 48k+ products in the early 90s? (even pro field?) unless I'm missing something. - Ive only conducted research using 'complete' d/a converter list' on dutchaudioclassics.nl and this seems rather odd... unproffesional.
Last edited:
Could it be that the professionals realized they didn't need more than 48K? In the 90's the difference between 48k and 96k carried a real cost factor in terms of processing power and storage/memory, unlike these days when CPU power, memory and disk space is so cheap that you might as well do 96k anyway...
The AA DITB does look to be a good starting point. But its using the Crystal CS8412 receiver which the DS says only goes up to 55kHz. So that'll need replacing with a more modern part. As you say you'll also have to bypass the oversampling filter chip - from the picture I can't work out which one is fitted, perhaps its a Yamaha part.
The Monarchy 24, aka Lite DAC50, is a PCM63 dac that does 24/96 input, outputting 20/96, and can be made into a fantastic unit. The Pioneer DV-S9(japan version DV-09) is a 24/96 player that can be used as a dac unit, using PCM1702's(easily piggybacked for better sound).
Interestingly, the Denon DCD-3000/DCD-S10, players that also can be used as dacs, while not spec'd to handle beyond 16/48, actually accept & process properly up to 24/96, and they use a pair of differential-connected PCM1702 per channel. The DA500 may also have this capability, but I have not verified it.
There are also several current production Lite dacs that are 24/96 w/PCM63/1702/1704 ladder dacs, e.g., DAC60, DAC72, DAC38 and DAC83.
Interestingly, the Denon DCD-3000/DCD-S10, players that also can be used as dacs, while not spec'd to handle beyond 16/48, actually accept & process properly up to 24/96, and they use a pair of differential-connected PCM1702 per channel. The DA500 may also have this capability, but I have not verified it.
There are also several current production Lite dacs that are 24/96 w/PCM63/1702/1704 ladder dacs, e.g., DAC60, DAC72, DAC38 and DAC83.
Early Audio Note NOS DACs with CS8412 and AD1865 can be updated to 96K with the substitution of a CS8414. Or, substituting a CS8415A with a little glue logic can take it to 192K. Although the CS8415A is spec'd to 96K, the ones I have tried do go to 192K.
Regarding your last question: There was no need for >48K because 44.1K/16 was declared to be "Perfect Sound, Forever!"
Regarding your last question: There was no need for >48K because 44.1K/16 was declared to be "Perfect Sound, Forever!"
It has to be remembered that that marketing slogan referred to the fact that the sound quality remained the same, despite how many hundred or thousand times the CD was played (unlike vinyl). It was not a claim that 44.1/16 was "perfect" from a sound quality point of view (although extensive testing and research has shown that 44.1/16, when correctly executed, is "good enough" so that any audible improvements are hard to achieve).
Thank you for the helpful replies.
Stephensank, are you familiar with the receiver used in those denons?
this info would be helpful to cross reference the vasiltech list for similar designs (which appear to be early DVD-Audio types, before SACD/HDCD took over 1996 ish).
I now realise older receiver chips can also be upgraded with 'drop in' mods.
see audiotuning.de
In turn with bypassing digital filter, should considerations be made regarding mismatching word lengths with this approach?
For example, replacing YM3623B (16out) with CS8414 (24out), prior to a 16bit dac I can imagine would result in problems, so ultimately it would be up to the user to pad/condition the signal to the correct depth, prior to SPDIF transmission, am I right?
Before
16 v input
16 v YM3623
16 v SAA7220
16 v TDA1541
After mod
16 v input padded to 24
24 v CS8414
16 v SAA7220 (bypassed)
16 v TDA1541
Obviously 24 into 16 can't work, so user would need to pad the data prior, is my point...
Stephensank, are you familiar with the receiver used in those denons?
this info would be helpful to cross reference the vasiltech list for similar designs (which appear to be early DVD-Audio types, before SACD/HDCD took over 1996 ish).
I now realise older receiver chips can also be upgraded with 'drop in' mods.
see audiotuning.de
In turn with bypassing digital filter, should considerations be made regarding mismatching word lengths with this approach?
For example, replacing YM3623B (16out) with CS8414 (24out), prior to a 16bit dac I can imagine would result in problems, so ultimately it would be up to the user to pad/condition the signal to the correct depth, prior to SPDIF transmission, am I right?
Before
16 v input
16 v YM3623
16 v SAA7220
16 v TDA1541
After mod
16 v input padded to 24
24 v CS8414
16 v SAA7220 (bypassed)
16 v TDA1541
Obviously 24 into 16 can't work, so user would need to pad the data prior, is my point...
The Denons I mentioned use the YM3623, which is not supposed to be capable of beyond 48/16, but it seems that it is governed only by what format coding bits it sees, so can actually process 24/96, at least properly enough for the SM5845(custom-for-Denon version of more or less an SM5847) digital filter to make sense of it. It is certainly possible to upgrade the input receiver to DIR9001/CS8414/WM8805 with an adapter/daughter board, but there isn't much point, since the ladder dacs used are far less sensitive to jitter than bitstream dacs.
The word length sent to the dacs is controlled by the digital filter chip, with the SM5845 in the Denons being configured to send 20bit out to the PCM1702 chips.
The word length sent to the dacs is controlled by the digital filter chip, with the SM5845 in the Denons being configured to send 20bit out to the PCM1702 chips.
Are you suggesting Y3623B can already accept 96khz?
Also,
However if the digital filter is bypassed, larger words would need to be padded out at source, right? Something the DF chip would otherwise do on it's own...
Also,
Aye this is what I thought.
However if the digital filter is bypassed, larger words would need to be padded out at source, right? Something the DF chip would otherwise do on it's own...
The YM3623 was not designed to accept 24/96, but it somehow does it.
I am absolutely not an NOS believer, so have never looked into what it might take to get 24bit words truncated to 20bit for 20bit dacs, or 20bit into 16bit dacs, etc. I don't think that is a simple thing, by any stretch.
I did find my Denon DA500 dac service manual last night, and it does use the same YM3623 dir & SM5845 df as the DCD3000/S10, so should also do 24/96, but just has one PCM1702 per channel, rather than push-pull 1702's per channel.
I am absolutely not an NOS believer, so have never looked into what it might take to get 24bit words truncated to 20bit for 20bit dacs, or 20bit into 16bit dacs, etc. I don't think that is a simple thing, by any stretch.
I did find my Denon DA500 dac service manual last night, and it does use the same YM3623 dir & SM5845 df as the DCD3000/S10, so should also do 24/96, but just has one PCM1702 per channel, rather than push-pull 1702's per channel.
I think it just a case of padding the word length using software, some playback programs can do this I think.
I actually have a dac/amp with ym2363 already so i'll experiment and post results when I get it to work.
Thanks for the help
I actually have a dac/amp with ym2363 already so i'll experiment and post results when I get it to work.
Thanks for the help
If you're talking strictly computer source, yes, you absolutely can tell various audio softwares to "only output 16/44.1" or whatever format you need. It's a common feature.
- Status
- Not open for further replies.