[sdiy] CD sound quality evolution

Mike Bryant mbryant at futurehorizons.com
Wed Feb 17 02:07:51 CET 2021

Hi Brian

> I cannot find reference to any Burr Brown DAC chip that's less than 16-bit.

They said 16 bit on the datasheet, but the LSB was almost random, and the next bit not placed too accurately.  

> My direct experience with BB DAC chips started in 1988, so that would be the time frame. I cannot find the exact part number that was used in my NeXTdimension and NeXTstation computers, but the sound quality was phenomenal.

My experience of them is 1982/3.  It was the PCM53 and was an impressive step forward.  They even admitted they trimmed it for best audio distortion rather than linearity, which was fine for consumer audio but weren't up to what we were trying to do with them and so we had to cancel the project.  But as usual at HP nothing goes to waste and several ended up in homers, including my first additive synth.

> I'm talking about the PCM58 (1988), PCM63 (1990), PCM64, PCM65, PCM1701. Some of those were a decade later, though. Burr-Brown pioneered techniques like active laser trimming, which I witnessed at their competitor, Analog Devices. 
> This was all before delta-sigma became common, although Southworth had a 24-bit delta sigma audio card for the Macintosh II computer perhaps as early as 1987 before they went out of business.

We were breadboarding with sigma-delta at HP in 1984 and eventually got to near 20 bits.  But it's a long way from a breadboard to a viable chip, and we didn't have the funds for a custom one on that project, especially as we'd need a ADC as well.
The PCM63 was good when it came out many years later, and I think some digital mixers used it, but the rest didn't really have the resolution for pro audio.  BB then seemed to lose the plot a bit when confronted with AKM and Cirrus, before being bought by TI.  That said I've had to convert to using the PCM5101A in a current project.  It's not as good as the AKM devices but after their fab fire, AKM supplies are uncertain at best.

> I frequently get the impression that Sony did not try to make the *best* digital audio system, but merely an affordable one that was good enough

I think you're mixing up Sony consumer and Sony Pro Audio divisions.  Pro Audio was part of their television and film equipment division back then and not separated out until later.  But the chips they designed were available thru Sony Semiconductor and were well received.  And it was more Philips that tried to hold things back - they wanted to just use 14 bits which really would have been a disaster.   But even Sony's consumer division produced the D-99 in 1988 which was like something from another planet, it was so far ahead of anything else at that time.  I bought one of the first ones when in Japan that year and still use it now and then.

> Meanwhile, PCM releases are being made that are not over-compressed. See HDtracks - although it's basically impossible to glean the provenance of their media without buying it first. Warp Records has artists like Squarepusher who do not over-compress their releases, which are available in 24-bit via Bleep.com

I've seen some places like this, though not HDtracks.  Looks quite promising, though it depends on which artistes allow their old material to be remastered.  I may order one and see just how good it is.  Never heard of anyone on Warp Records.  

>>  The Phil Collins problem was infamous at the time, 
> That sounds like the EFM (eight-to-fourteen modulation) or some other very low level aspect well ahead of the audio samples was in effect. Are you talking about the lossless CIRC error correction, or the lossy sample interpolation that takes place only after error correction fails?

I've no idea of the exact cause - I never liked Phil Collins :-)   You should realise that the Red Book was taken as a rough guideline at best in the early days and so plenty of shortcuts were taken, and this was one too far.  Of course this happens even today with the various copy-protection schemes that keep on appearing.


More information about the Synth-diy mailing list