PCM1792A vs. PCM1792 + influence of oversampling on THD etc.

[capslock]

Those As are confusing, sometimes it is a higher grade version, sometimes a lower grade. Then with DACs sometimes it seems money was saved on the digital filter section, at other times (AD) it seems there is a difference in the analog section.

Anyway, there is now a PCM1792A, a lower cost version of the 1792. According to the parametric selection table, it has 127 dB SNR vs. 132 for the 1792. Looking into the data sheet, the 1792 reaches 132 dB only in mono mode, and the 1792A is specified at excactly the same value. Scanning through the rest of the specs, I have not been able to detect any diffference at all, not even in typical vs. maximum values. Has anybody found a difference?


Another interesting observation: the oversampling ratio in the 1792(A) is user selectable. However, neither the spec numbers nor the typical performance graphs give any indication as to which os ratio was used nor how for a given fs, the os ratio influences noise and distortion. As fs grows, both distortion and noise get worse, as is seen in most other DACs. I have often wondered whether this effect is actually related to the modulator switching frequency. In this case, one should select as low an oversampling ratio as possible.

[ojg]

I disagree with your conclusion here. The distortion grows with increasing Fs because that results in less oversampling of the delta-sigma modulator. You get the best performance when using 44.1 with a 128x OS. Since 192kHz only works with OS <=64 you will get worse performance.

[capslock]

Hmm, as I said, they don't even state the os ratio used for those various fs. 128x can be used for both 48 and 96 kHz, so it would make sense to do that in a comparison. Or they just stuck to the default (64x)?