Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.user > #255594

call for beta testing emu10k1 driver improvements

From Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Newsgroups linux.debian.user
Subject call for beta testing emu10k1 driver improvements
Date 2023-03-04 18:50 +0100
Message-ID <G5Eb7-a48d-9@gated-at.bofh.it> (permalink)
Organization linux.* mail to news gateway

Show all headers | View raw


moin,

i did some hacking on the emu10k1 driver, and now need it tested on a 
variety of hardware. the cards still have a market value, so *someone* 
must be still using them, or at least have some in the basement ...

the driver targets SoundBlaster Live! / 512PCI / Audigy and the E-MU 
0404/1010/1212/1616m/1820 PCI/PCIe/CardBus series. the exact labeling of 
the relevant cards can be found in the table at 
https://github.com/torvalds/linux/blob/master/sound/pci/emu10k1/emu10k1_main.c#L1273

the primary objective of the patch series was adding support for 96/192 
khz sample rates on the e-mu cards, but a variety of other (mostly 
minor) improvements and bugfixes is included as well. i have only an 
e-mu 0404 card, so my own testing is obviously limited.

to fully unleash the driver, emu_das=1 needs to be added to the module 
options. this switches the driver to a mode which is incompatible with 
existing configuration files and saved mixer states (the cost/benefit 
ratio of extending the legacy mode in a compatible way is unreasonable).  
only the multi-channel device is available in this mode (it's still 
usable as a regular sound card, but you won't have much fun with it 
without a sound server - but most people have that nowadays anyway).

things that need testing in particular:
- the source & destination routing in the mixer; the internals were 
   pretty much completely rewritten. test at different settings of the 
   clock multiplier, as that massively affects the routing.
- capture on 2nd generation e-mu 1010 and 1616m cardbus cards: 8 
   channels at 192 khz and 16 channels at 96 khz (whether the upper half 
   of the channels actually records something, and whether these channels 
   aren't delayed by a single sample or otherwise distorted)

then there are a few things that don't need testing as such, but rather 
help me with understanding the hardware in the first place:
- inspect the contents of /proc/sound/card?/spdif-in with various 
   digital sources - adat and s/pdif, both on the card and the dock, and 
   bnc sync. what are the lock status and the displayed rates with no 
   signal, good signal async to internal clock, and synced signal? test 
   s/pdif at least with 48 & 96 khz (192 works on 1010b & 1616m cards on 
   the RCA port). use the mixer to switch around the clock source between 
   internal and the connected external one.
- if you have a second e-mu card:
   1. if it's in a different machine, make sure it's also running my 
   patched kernel
   2. connect its s/pdif output to the 1st card's s/pdif input
   3. make sure the clock multiplier is the same on both cards
   4. select s/pdif as the 1st card's clock source
   5. switch the 2nd card's word clock between 48 and 44.1 khz
   6. see if the kernel journal (dmesg) for the 1st card reports 
   "emu1010: External clock rate changed"
   7. if it reports "emu1010: External clock lost" instead, go back to 
   point 4. retry a few times.
   8. if it (occasionally) doesn't report anything, that's valuable info, 
   too. but switch either card's clock multiplier to make sure that a 
   clock loss is actually reported at all.

for soundblasters barely anything changed, so it's only necessary to 
test whether i didn't break anything - playback and the mixer in 
particular. /proc/sound/card?/emu10k1 should be somewhat more useful 
than before. on cards with an additional CA0151 chip (various audigy 2/4 
cards), playback over hw:?,4 should be tested.

the code is available from 
https://github.com/ossilator/linux/tree/ossis-alsa and is supposed to 
work out of the box (like any rc kernel ...).

if someone wants to provide me with hardware (temporarily or
permanently), go ahead (i'm in berlin, germany).

please send responses directly to me. you can cc the list if you have 
questions relevant to a wider audience.

regards

Back to linux.debian.user | Previous | Next | Find similar | Unroll thread


Thread

call for beta testing emu10k1 driver improvements Oswald Buddenhagen <oswald.buddenhagen@gmx.de> - 2023-03-04 18:50 +0100

csiph-web