da_burl
100+ Head-Fier
- Joined
- Feb 13, 2002
- Posts
- 164
- Likes
- 10
One more thought on what is an inspiring thread. Even if you have a non-upsampling sound card, there is a "bug" in MS Windows (more specifically DirectX and Direct Show ) that re-samples everything, even 44.1khz audio. Yes, it will internally resample 44.1 khz to 44.1 khz! I didn't believe it either, until I read some posts.
”If I understand the K-mixer correctly, It translates 44.1 to 48 KHz, then retranslates it back to 44.1. Am I correct or not?”
You are not correct. The purpose of the SRC is to resample wave audio that has a sample rate different than the default or currently used (by client software) sample rate to enable multiclient compatibility. When Windows boots the KMixer/SRC defaults to the Redbook CD audio sample rate of 44.1kHz and any wave audio input, at that sample rate, to the KMixer should bypass SRC and be passed to the soundcard unmolested at 44.1kHz and if another wave file is played at the same time with a different sample rate (ie. Windows sounds at 22.050kHz) the SRC would resample them to the default sample rate of 44.1kHz. The bug in Windows is that the SRC is applied to any wave audio regardless of the sample rate, even when it is at the default of 44.1kHz. The SRC uses a resampling algorithm with a quality factor that changes the bits including when going from 44.1 to 44.1kHz.
”Anyway, whatever it does it degrades the signal and should be abolished. Every software program should have ASIO drivers built in.”
Agreed! If the bug were corrected ASIO would only be needed to lower latency."
This is why the fuss over asio and kernel streaming to bypass the Windows Kmixer (affectionately referred to as the KMangler at avsforum).
I don't know a lot of specifics, now I am wondering if it applies to CD Audio, or just wav files, or for that matter any other lossless or lossy compression.
Anyway, I safely use foobar2K and kernel streaming to avoid any complications!!
”If I understand the K-mixer correctly, It translates 44.1 to 48 KHz, then retranslates it back to 44.1. Am I correct or not?”
You are not correct. The purpose of the SRC is to resample wave audio that has a sample rate different than the default or currently used (by client software) sample rate to enable multiclient compatibility. When Windows boots the KMixer/SRC defaults to the Redbook CD audio sample rate of 44.1kHz and any wave audio input, at that sample rate, to the KMixer should bypass SRC and be passed to the soundcard unmolested at 44.1kHz and if another wave file is played at the same time with a different sample rate (ie. Windows sounds at 22.050kHz) the SRC would resample them to the default sample rate of 44.1kHz. The bug in Windows is that the SRC is applied to any wave audio regardless of the sample rate, even when it is at the default of 44.1kHz. The SRC uses a resampling algorithm with a quality factor that changes the bits including when going from 44.1 to 44.1kHz.
”Anyway, whatever it does it degrades the signal and should be abolished. Every software program should have ASIO drivers built in.”
Agreed! If the bug were corrected ASIO would only be needed to lower latency."
This is why the fuss over asio and kernel streaming to bypass the Windows Kmixer (affectionately referred to as the KMangler at avsforum).
I don't know a lot of specifics, now I am wondering if it applies to CD Audio, or just wav files, or for that matter any other lossless or lossy compression.
Anyway, I safely use foobar2K and kernel streaming to avoid any complications!!