Don't get me wrong. I could hear subtle improvements over the ODAC (which I think is pretty sterile), but I don't think it justifies the upgrade cost. I was hoping for a significant improvement... perhaps from entry to mid-tier category.
If I get a chance for a demo unit again, I'm gonna try burning in and see if it changes my opinion.
So the "burn in" thing was just driving me crazy so I did a little research. The best I found from Mike Moffat
here which states:
"It is not the DSP which needs to thermally stabilize. It is the DAC chip which needs to settle into its INL spec."
@Baldr
Now the problem is he has never directly answered the question about thermal stabilization of the DAC chip, which has been asked on at least a dozen occasions that I have found; and these were on threads he was posting on.
I think this may be a marketing issue where he is being intentionally vague. The issue is INL is a specification of the DAC, in fact there is a specification for it over voltage and temperature range. So it does not need time to "settle" into this specification. However the DAC being used has a LIN COMP register in which they can apply one or more dynamic nonlinearity compensation techniques. There are many techniques that you can find using the digital domain, analog domain, or both. Some of these use LIN COMP, while others adjust the gain and offset error registers. What it does appear, based on several thesis papers (you can find on the subject), that you can do this accurately or fast; and the larger the bit depth the longer it takes.
This would seem to be far more accurate as to why the Modi Multibit (16 bit) seems to only need a few hours whereas the Yggdrasil (20 bit) needs a day or more to "settle."
I hope
@Baldr responds because this is driving me crazy. The Ideal operating temp for these DAC's are +25 degree C (ambient), and should come "up to temp" in a matter of minutes.
This may be a proprietary dynamic nonlinearity compensation technique that they do not want to discuss, or maybe this part of the code is a black box to him. I would really like to see a real response to people believing the chip is still "coming up to temp" over the course of 24 hours (some even claim upwards of 150 hours).
In essence, something is stabilizing; and I i have not seen a single paper talking about thermal stabilization (at room temperature ambient) over the course of hours, let alone days.
I just noticed that the Holo Audio Spring is claiming to be the first and only DAC with linearity compensation.... Interesting, though I don't think this is true based on other articles I read.
-Ray
