Lucas - great to hear that the endpoint patch fixed your crackles! I'm curious if you get any different results enabling feedback on the capture endpoint. I'm also curious about the output from "lsusb -v" for your device and whether the implicit feedback setup looks like the GT-1. I'm not exactly sure what it even means to have the capture endpoint set up with implicit feedback as most of the implicit feedback code seems to deal only with playback. The bit of code in endpoint.c that I patched out, for example, only runs on playback endpoints. I *think* the existing code may effectively just be completely turning off implicit feedback for these devices. This results in audio playback/capture that "works", but is susceptible to pops/crackles due to small timing mismatches. Mike On Wed, Apr 7, 2021 at 10:36 PM Lucas <jaffa225man@xxxxxxxxx> wrote: > Well, I just got back from testing Mike's suggestion of setting the Roland > Boutique D-05 as a playback quirk, disabling its capture quirk, and setting > that endpoint test section to always be skipped. Sure, enough, the D-05's > playback is now perfectly crystal clear without the crackles associated > with an LP record. > > This is just a guess until I recompile again, but my capture from the D-05 > reports a read error, seemingly just as the vanilla mainline kernel does, > and I think it's because I disabled its capture quirk line. So, it seems > to me that the D-05 needs both quirk table entries, whether or not the GT-1 > does. >
