123 dB (A) | Mytek Stereo192 ADC
114 dB (A) | SD 722
110 dB (A) | V3
108 dB (A) | M-Audio Firewire 410
105 dB | Mini-Me
100 dB | Fostex FR-2
98 dB (A) | M-Audio MicroTrack
94 dB | Marantz PMD-671
? | Tascam HD-P2
unpublished | Edirol UA-5
unpublished | Edirol R-1
unpublished | Edirol R-4
unpublished | Tascam DV-RA1000
Added the 722 to the above for completeness and fluffing..
Talk to Doug (and many others) and they will probably tell you that great specs don't necessarily reflect good audio quality. In many cases, excessive negative feedback is used to reduce noise at the expense of good sound quality.
Case in point, there are quite a number of 722 owners who insist on fronting the 722 with a V2 or V3. However, even the V3 users seem to prefer using the 722's A/D. I expect that the "specs" on that combination aren't as good.
As far as measuring these devices ourselves, I imagine that doing these tests with no load on the input is incorrect. There should probably be a loading resistor to simulate a mic, etc (and probably another one to simulate 48v load).
While we assail the MT, it would be interesting to see the same data on the R-1 in 24 and 16 bit mode, etc. For that matter, how about a 148>modsbm? 148>AD1000? Etc.
I'll still take noisy 24 bit over noisy 16 bit because there is more detail in the portion of interest.