View Single Post
Old 20th November 2014 | Show parent
Lives for gear

Originally Posted by funka View Post
living sounds, did you have the time to do the Aurora DAC / TI PCM4222 EVM test?
Here you go. These are the raw recordings with only manual DC removal applied, they have not been processed by Diffmaker.

Aurora is clocked to EVM via HDSP9632 and DA24 MK2 as described above. Each time, Aurora is DAC, EVM is ADC.

Aurora running at 44,1khz synced externally to EVM, synchro lock off

Same as before, but synchro lock turned on.

Aurora at 44,1khz but clocked internally.

Aurora at 88,2khz clocked internally. Signal is provided at 44,1khz and EVM records at 44,1khz.

The versions with the Aurora clocked on its own require sample drift compensation in Diffmaker to be turned on to get useable results. Since the software crashes with the 2 minute long files, the only way to get proper Diffmaker files is to run it on half a file (1 minute).

The Diffmaker files don't tell the whole story, BTW. My general observations comparing the converter settings listening to music and comparing to different other converters are the following:

The Aurora sounds better with synchro lock turned off. It mushes/muddies the sound, even with the Aurora clocked internally.

88,2 khz has more high end, probably due to the anti aliasing filter having less impact in the audio range.

However, the Aurora sounds best clocked to the EVM at 44,1khz (no synchro lock). Best transient response and clarity, definition, which is more important than flat frequency response IMO.

I'd love to test the Aurora clocked to the EVM at 88,2khz, but the route I have to take (through the Swissonic) doesn't allow going beyond 48khz. There is no proper clock output on the EVM itself.