[sdiy] uLaw DACs in the modern era?

Tom Wiltshire tom at electricdruid.net
Wed May 23 11:54:28 CEST 2018

I don’t see the point of such exhaustive testing. We know well enough what the DAC does, since the datasheet tells us the exact levels in detail. As others have said, you can use a look-up table to get the same outputs from a modern higher-resolution DAC (14 bit would do, 16 is ample). As far as the u-Law encoding goes, the output from the AM6070 and the modern DAC will be identical. While there might be small differences due to different DAC glitches, I don’t think anyone has ever claimed that the *DAC glitches* are an important part of the sound, and such an effect would be utterly swamped by effect of the 8-bit encoding anyway.
There may be other stuff that’s important after that (distortion in following stages, CEM filters on some sounds not others, etc etc) but again, these are not specifically anything to do with the DAC or the encoding.

I find the encoding the most interesting part because it allows a much wider dynamic range from only 8-bit sample data, which gives the units more “punch”, while the low bit depth and extremely sketchy reproduction of high level signals gives a measurable (although not offensively perceptible) “crunch”. This is a good match for drums.

       Electric Druid
Synth & Stompbox DIY

> On 23 May 2018, at 05:28, cheater00 cheater00 <cheater00 at gmail.com> wrote:
> That's what people did - anyone using those as their main drum machine would have them modified to take external eeproms, with a zif socket, usually in a side car connected by flat tape. People did that and samples played back through those still had that extra quality, but obviously that is subjective. The best way to figure this out would be to get some high resolution samples of common waveforms coming out of those - bass drum, pink noise, white noise, sine waves, triangle waves, square waves - at a few different frequencies (35 Hz, 65 Hz, 255 Hz, 1005 Hz (all frequencies offset by 5 Hz to prevent capturing mains)), and amplitudes (0 dBFS, -0.5 dBFS, -3 dBFS, -6 dBFS, -24 dBFS, +0.5dBFS (or whatever reconstruction overshoot would allow). Record with at least 20 bit headroom (ie bit rate minus noise floor minus recording level below 0 dBFS). Then, look at it in a wave editor and sonograph, searching for glitches.
> What do you mean by layouts, Paula? I guess you don't mean top posting since you top post too.
> On Tue, 22 May 2018 14:08 Tom Wiltshire, <tom at electricdruid.net> wrote:
> Yeah, I’m not sure swapping the EEPROMs counts as “able to take external samples”, although it’s literally true - you *could* put whatever you wanted in there if you’re willing to do the work of preparing the samples, converting to 8-bit u-Law and burning the EEPROMs.
> > On 22 May 2018, at 11:46, paula at synth.net wrote:
> > 
> > Not sure what you mean by "take external samples", the LInndrum, Drumulator, and many others are fixed (unless you change the EPROMs).
> > 
> > Paula
> > p.s. I'm struggling to follow emails with layouts like this.
> > 
> > 
> > 
> > On 2018-05-22 11:02, cheater00 cheater00 wrote:
> >> Those drum machines take external samples, too.
> >>   On Tue, 22 May 2018 11:00 , <[1]rburnett at richieburnett.co.uk> wrote:
> >>     I suspect that a lot of the perceived "punch" in the sounds of early
> >>     digital drum machines is down to the same sorts of processing
> >>     applied to
> >>     the acoustic drum recordings before sampling that we would apply to
> >>     drums these days to obtain a more punchy and "produced" sound.�
> >>     Namely,
> >>     things like dynamic range compression with artistically chosen
> >>     attack/release times, gating, dynamic EQ, etc.�  Oversights in the
> >>     recording process like allowing a fair degree of clipping also make
> >>     things sound a lot harder too.�  Heavy clipping is not uncommon in
> >>     the
> >>     samples of these early drum machines!
> >>     -Richie,
> >>     _______________________________________________
> >>     Synth-diy mailing list
> >>     [2]Synth-diy at synth-diy.org
> >>     [3]http://synth-diy.org/mailman/listinfo/synth-diy
> >> References
> >>   1. mailto:rburnett at richieburnett.co.uk
> >>   2. mailto:Synth-diy at synth-diy.org
> >>   3. http://synth-diy.org/mailman/listinfo/synth-diy
> >> _______________________________________________
> >> Synth-diy mailing list
> >> Synth-diy at synth-diy.org
> >> http://synth-diy.org/mailman/listinfo/synth-diy
> > _______________________________________________
> > Synth-diy mailing list
> > Synth-diy at synth-diy.org
> > http://synth-diy.org/mailman/listinfo/synth-diy
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy

More information about the Synth-diy mailing list