If it's purely an engineering challenge why bother designing yet another DAC?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I fully believe your experience here. The problem is that with an other room and other speakers, the finding may be reversed. Agreed?

So how does one choose a DAC?

//

If one keeps all other variables the same and only change the clock, and if you also keep a reference dac to calibrate your subjective listening impressions as required, then differences in clocks are quite apparent.

However, the differences are not dramatic, and no claim is made to that effect as Wurcer disingenuously suggests.

What does happen is that the quality of the stereo illusion improves (or worsens) and or there is some change in subjective tonal balance while measured frequency response stays flat. That is when 'good' clocks are substituted.

Using clocks of phase noise performance less than 'good' results in increased audible artifacts including increased distortion of a type that may not show up well on THD FFTs.

I did some checking of RF coming out of the I/V stages for the AK4499 eval board and for D90. For CD audio, most of the RF was at twice the bitclock frequency (roughly 2MHz). It was also slightly modulated by MCLK frequency.

After the I/V output, the RF is attenuated by 2 passive filtering stages leaving it harder to see with a scope (at the I/V output is ~200mV peak to peak). The passive filtering is nearly the same for D90 and for the eval board. However, the eval board doesn't have any more opamps after the passive filters, whereas D90 does, but not with any additional filtering (only resistors around the opamps, I sketched up a schematic).

Using the passively filtered output of the eval board as a source I was able to measure some remaining ~2MHz RF at the headphone output jack of Neurochrome HP-1 when set to maximum gain. That suggests to me that enough RF could come out of the dac to possibly cause some low level audible distortion in equipment downstream of a dac that might not show up very well on an FFT. In fact, it could be that some (likely not all of) subjective 'brightness' heard could turn out to at least partially arise from such a cause.

While taking a look at the above with a pretty good scope (the cheap scope couldn't sync on the RF noise), it occurred to me there is probably a way to cancel out most of the RF if pairs of DAC channels are summed in a particular way (maybe more about that some other time).

All I have to say about it for now.
 
Last edited:
Member
Joined 2014
Paid Member
Oh I agree its good news. I was merely noting that all Jam's assertions about speakers 'dissapearing' is not hard. Also I love vinyl as a fun passtime, but with at best 25dB seperation it's not going to give image specificity like digital. Just not possible.



Back in the 90s with a marantz CD80 some orchestral CDs would make one 'believe' the kettle drums were 6ft behind the speakers and 4ft to the left. All an illusion!
 
I was merely noting that all Jam's assertions about speakers 'dissapearing' is not hard.

It all depends how rigorously 'disappeared' is defined. Jam's description wasn't particularly a hard one I agree. I've only ever heard the trick pulled off completely (in the sense they can't be found even when walking around, not merely from the listening position) by a pair of MBLs.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.