Hope it's okay I post this.
ASR has posted measurements of the Oppo-105. The reason I'm mentioning it here is that there appears to be an anomaly in the bit linearity test in the 105 causing triangle waves to happen rather than a straight line. By the ASR definition of 0.1 db variation used to declare a dac a certain number of bits, this would make the 105 a 0 bit dac. This is the same test intepretation that has resulted in the Bifrost being labelled a 10 bit unit and the Yggy being a 16 bit unit when it varied from 0.1 db linearity. I pointed out that a 0 bit DAC in a Oppo 105 doesn't seem to make sense and maybe the wording of the test interpretation needs to be refined.
He reported different results than previous measurements on the Soekris DAC too. Amir does not know how to take proper measurements, document a test setup, or understand his own data. I would not take anything seriously from that hack.
I'm with you
@Rtg97229 In today's world (thanks to the internet) it's not too hard to apply a modicum of rigour fuelled by knowledge that is readily obtainable. To summarise the diatribe below, I can't see alignment between the arbitrary ENOB "measurement" method referred to in
@garbulky 's quote from ASR and recognised standard. So, I think using the term "hack" is not unreasonable. I'm happy to be corrected on that point (refer the detail below if/when anyone does).
I'm not a measurements guru but from a quick finger walk on the interweb I found various references describing the calculation of ENOB. From a deeper search, I actually found a copy of IEEE Std 1658-2011 Standard for Terminology and Test Methods of Digital-to-Analog Converter Devices (I won't post a link but if you're resourceful, well, ... enough said) and the calculation for ENOB according to the standard which notably differs from the generic equation that you'll come across in places like Wikipedia:
I also found this paper,
Some Critical Notes on DAC Frequency Domain Specifications, Balestrieri et. al.,
http://grouper.ieee.org/groups/1658/NextMeetingArchive/061030DAPONTE/DAPONTE_P.pdf that appears to have provided input to the IEEE Std 1658 and from reading it, the work appears to have been inspired by the lack of a common test frame work and general confusion.
Balestieri, et. al. comments on the above ENOB equation as follows:
Defining ENOB in terms of another figure of merit, as a measure of the signal-to-noise and distortion ratio makes this parameter useless as it is redundant.
Taking into account the previous considerations, the proposed definition for ENOB is:
“For an input sinewave of specified frequency and amplitude, after correction for gain and offset, effective number of bits is the difference between the DAC digital resolution and the number of bits affected by DAC nonideality. This number has to be obtained as the binary
logarithm of the ratio between the measured rms noise and the ideal rms quantization noise”.
In this way ENOB definition is not directly related to the SINAD.
The above narrative helps explain why IEEE 1658-2011 defines ENOB as per the excerpt below:
I'm happy to be corrected about this but I can't see any reference to a 0.1 dB criteria being used to determine ENOB in the detail above.
I find it difficult to take data and analysis from anyone seriously, if it's not supported by reference back to recognised standard and some demonstration that the standard has been competently understood and applied.
The concluding statement from this paper
https://www.mccdaq.com/PDFs/anpdf/DT-Application-Notes/ENOB-Overall-Accuracy.pdf puts the issue better than I can articulate at short notice:
ENOB is not just a specification, but a set of conditions under which the specification was measured. SNR and error measurements between products can only be compared if they were taken under the same circumstances: at maximum throughput, on alternating full-scale and zero-scale measurements, measured from the module’s input connectors using data output to the computer’s bus. If measurement conditions are not specified, you should not trust the specification.
The issue of procedual rigour is at the heart of the advice above and that's why I lack confidence in the musings coming from the source in question.
