About WASAPI/WaveRT device switch issue

Started by jmkz, March 26, 2021, 18:08:50

Previous topic - Next topic

jmkz

Recently discovered that switching from WASAPI Default Device to WaveRT Default Device (PortAudio) sounds the same (not bypassing driver APO filter). Changing from WASA.PI default to others like WaveRT direct device seems to not change or lock to it. Looks like it's a bug, but i'm unsure. Also noticed in status box, there's no 'OK.' message nor 'Exclusive stream'
Looks like faulty ASIO driver, because corrupts audio device after some device switch.

1. https://youtu.be/FXVRnoFydbs
2. Compressed dump: https://drive.google.com/file/d/1wOwt1l8PZxLh04bL5ADtZyjGSjxsgD05/view

Edit: Added pictures, ASIO assert dump, video

manx

Quote from: jmkz on March 26, 2021, 18:08:50
Recently discovered that switching from WASAPI Default Device to WaveRT Default Device (PortAudio) sounds the same (not bypassing driver APO filter). Changing from WASA.PI default to others like WaveRT direct device seems to not change or lock to it. Looks like it's a bug, but i'm unsure. Also noticed in status box, there's no 'OK.' message nor 'Exclusive stream'

"Exclusive stream" is only shown for WASAPI mode exclusive streams. "OK" is also not shown for WaveRT devices. In general, the information displayed in the status box is highly backend-dependant.

As to why WaveRT would not bypass any driver APO or other sound-manipulating filters, I cannot tell. This would very likely be a driver problem (like, whatever your driver does here might not be implemented as an APO but instead later in the chain, and thus also affect WaveRT streams).

Quote from: jmkz on March 26, 2021, 18:08:50
1. https://youtu.be/FXVRnoFydbs

Switching to WaveRT from a running WASAPI stream (the first switch you did) might not immediately work because WASAPI might not have closed a device yet at the point when WaveRT tries to open it. Also, could you try a different sample format for WaveRT (i.e. 24bit instead of 32bit)?


Regarding the ASIO issue, please also provide the actual assert message.

jmkz

Quote from: manx on March 27, 2021, 11:15:18

Regarding the ASIO issue, please also provide the actual assert message.


Added window capture, and additional info from VS output box in asioassert_vslog.txt attachment.