[Summa]

Originally Posted by LivePsy

Actually, I don't see any reason why you can't have DC from digital or analog outputs if there's no DC decoupling in the analog output stage. I assume this is a discreet stage output of the DAC chip but that could be my ignorance.

Since at least when it comes to analog outs this could kill your speakers, the inductors will overheat when exposed to a constant current. When it comes to the mix, the DC-Offset is shifting the waveforms symmetry above or below the X-Axis and thus creating level without loudness.

I've sampled the S/PDIF into several sampling workstations, direct optical. Its always the same result AND not just the TI. The waveform of low bass notes is obviously low cut visually. You really should be able to have a perfect 20Hz saw wave coming out of the TI if you really want one.

If the S/P-DIF signals shows the same problems it's created by the synths engine/software and no DAC problem.
As mentioned in my last posting, mixing and integration can be easier and needs less EQing when the synth already outputs a processed sound with amplified mid-frequencies by dampening low and high frequencies or even (as some synths seems to do) cutting bandwith to conserve some processing power. Most ppl. seem to prefere that type of sound even so the synth tends to lose quite a bit of its flexibility that's why I don't have exactly a crush for that type of synths and a flexible synth engine allows to cut the unneeded frequencies at will.
Anyway when it comes to a mathematical perfect saw, if this is what you're looking for, even none of my analog synths create one and when I played arround with different saw samples the perfect ones don't sound very well.
Still, when it comes to the Virus and Bass sounds with lots of low end, it takes quite a bit effort to create them.

Perhaps all my sampling workstations are all at fault here?

Don't think so, but I could be wrong, since I mainly use my Computer for sampling/recording sounds.

I'm intrigued by this idea but I just don't understand it. The only slight differences I can imagine would be similar to eq. And you might be able to compensate for that with eq in the DAW. Can you give an example (non TI example is fine)

Well it is similar to an EQ, it's an effect widely known as Fletcher Munson Curves, that explains how different levels influence the frequency perception.

Here a link to the revised version of the curve.

http://en.wikipedia.org/wiki/Equal-loudness_contour

So when comparing DACs one have to make sure the Level is identical and testing double blind, since ppl. often tend to hear what they expect to hear.

...Summa