Look into the thread for test file and instructions.
https://www.audiosciencereview.com/forum/index.php?threads/can-you-trust-your-ears-by-tom-nousaine.1669/page-23#post-54233
It seems that I don't have a good enough analyzer so I attached the audio files as well. I used my X-Fi Titanium HD to record the output of Realtek ALC892 at 96k. "native" is direct WASAPI playback, "sox" is resample the signal to 96k using foo_dsp_resampler_mod with default settings (linear, 95% passband, no alias, normal quality).
Of course there are some differences in the screenshots, but it seems that nothing is "broken". IMO the Realtek should be a good candidate of "typical" filters due to its popularity, can someone look into the files or do some tests on your devices?
Look into the thread for test file and instructions.
https://www.audiosciencereview.com/forum/index.php?threads/can-you-trust-your-ears-by-tom-nousaine.1669/page-23#post-54233
It seems that I don't have a good enough analyzer so I attached the audio files as well. I used my X-Fi Titanium HD to record the output of Realtek ALC892 at 96k. "native" is direct WASAPI playback, "sox" is resample the signal to 96k using foo_dsp_resampler_mod with default settings (linear, 95% passband, no alias, normal quality).
Of course there are some differences in the screenshots, but it seems that nothing is "broken". IMO the Realtek should be a good candidate of "typical" filters due to its popularity, can someone look into the files or do some tests on your devices?
No problems here, either.
However, the SA9023A + ES9018K2M USB DAC attached to this machine misbehaves very badly with something far simpler: 19+20 kHz that peaks at - 1 dB FS. (attached).
However, the SA9023A + ES9018K2M USB DAC attached to this machine misbehaves very badly with something far simpler: 19+20 kHz that peaks at - 1 dB FS. (attached).
Old versions of RMAA used such frequencies and made a lot of AC97 soundcards got the "poor" rating.
[EDIT] Tested your files.
While no one pointed out my mistake, the 19+20kHz test was done by resampling the Realtek to 48kHz, that explains the aliasing on 28 and 29kHz in the previous post.
For the sake of correctness here is the correct version.