Poking around at -60. 16 vs 24 bit

[mixsit]

Reading Vurt's thread about converters, thought you might be interested in an experiment I did.
Tryng to get a better handle on the nature of converter guality, I did some comparison recording between ADAT and Tascam DA-30 at 16 bit, and the RME ADI-8 at 24 bit, at -40 and -60 (miced solo acoustic guitar).
First off, the difference at -60 was a real eye opener. The 16 bit versions are swamped with gross distortions, clock noises, ect, with the worst of it modulating with the signal(!), while the 24 bit's are rather quiet, mostly some analog sounding hiss in the background. (This is with the guitar fading into -80-ish.)
What was more surprising, especially in 16 bit mode, was that the signal was coherent and 'normal' sounding. I expected it to sound distorted or coarsely grained I guess. What I didn't see was the expected loss of detail at low levels often mentioned, just the signal surrounded by noise. (Not that I could realy hear much 'detail' under these conditions.) This is not what I expected.

At this point I have to agree with Dobro and JuSum, the bigest improvement was is in the noise floor, but up in the top (normal)ranges, smaller gains are from converter quality, regardless of bit depth. I think we have it really good now, with great sounding gear being so affordable!
A few side notes.
I tried 24 bit ad-da recording via light pipe to/from the RME and ADAT (chopped to 16 bit, -60ish). The adat's noise level went down quite a bit. An up-grade for 16 bit tape?!
The ADI-8 has 'aux' light pipe in/outs (for bit spliting 24 bit recording to 16 bit media). Turns out they're also duplicate in/outs, usefull for routing options. (I can use the adats's meters in digi-in mode. Nice, since the rme has squat for meters.)
The ADI's "over" led hits at about -3 or -4, which I suppose is ok if you have no meters.
I still have to settle down and do some listening work at normal levels, but so far I got a big grin on my face!
peace
mixsit