So I was having a blast (esp. in open firmware ;) ) with my new G5. Compiled Cinelerra-gg on Slackware ~current ppc32 (gcc 15.2.0). But sadly sound was broken? I suspected ffmpeg but even native way sounded wrong, but rendered correctly?! Turned out both ALSA and Pulseaudio hardcoded LE variant, and at least on Slackware alsa offered BE variant - picked by speaker-test utility. I added code to use this or old variant depending on BC_Resources::little_endian and it seems to sound fine now on both outputs? Can you please test on le ? (usual intel/arm)
The patches install without any issues, and CinGG launches without errors. But I haven't tested it (I only have Ryzen).
ср, 3 июн. 2026 г., 21:06 Andrea paz <gamberucci.andrea@gmail.com>:
The patches install without any issues, and CinGG launches without errors. But I haven't tested it (I only have Ryzen).
Thanks. I surely mean testing on normal x86(-64) - my machine for now donated her monitor to this G5 ;) You can play sound normally, and also record sound via both pulseaudio and alsa?
Andrew, I see Andrea has successfully compiled and tested. I have to figure out which patch is causing me an error as I have not been able to do a complete build on this older laptop so I will add one patch at a time (can't find the error in the log yet). On Wed, Jun 3, 2026 at 6:35 AM Andrew Randrianasulu <randrianasulu@gmail.com> wrote:
So I was having a blast (esp. in open firmware ;) ) with my new G5.
Compiled Cinelerra-gg on Slackware ~current ppc32 (gcc 15.2.0). But sadly sound was broken? I suspected ffmpeg but even native way sounded wrong, but rendered correctly?!
Turned out both ALSA and Pulseaudio hardcoded LE variant, and at least on Slackware alsa offered BE variant - picked by speaker-test utility.
I added code to use this or old variant depending on BC_Resources::little_endian and it seems to sound fine now on both outputs?
Can you please test on le ? (usual intel/arm)
Build with patch 0002-Fix-mjpegtools-build-on-ppc-g5-and-gcc-15.2.0.patch fails and I have not yet found the actual problem in the log. I will send it to just Andrew. On Wed, Jun 3, 2026 at 6:35 AM Andrew Randrianasulu <randrianasulu@gmail.com> wrote:
So I was having a blast (esp. in open firmware ;) ) with my new G5.
Compiled Cinelerra-gg on Slackware ~current ppc32 (gcc 15.2.0). But sadly sound was broken? I suspected ffmpeg but even native way sounded wrong, but rendered correctly?!
Turned out both ALSA and Pulseaudio hardcoded LE variant, and at least on Slackware alsa offered BE variant - picked by speaker-test utility.
I added code to use this or old variant depending on BC_Resources::little_endian and it seems to sound fine now on both outputs?
Can you please test on le ? (usual intel/arm)
Thanks to Andrew, checked in these 4 patches to GIT - minus changes on the line " -Wl,--end-group -lm -lstdc++ -pthread -latomic \" due to latomic not functional on older systems. On Wed, Jun 3, 2026 at 6:35 AM Andrew Randrianasulu <randrianasulu@gmail.com> wrote:
So I was having a blast (esp. in open firmware ;) ) with my new G5.
Compiled Cinelerra-gg on Slackware ~current ppc32 (gcc 15.2.0). But sadly sound was broken? I suspected ffmpeg but even native way sounded wrong, but rendered correctly?!
Turned out both ALSA and Pulseaudio hardcoded LE variant, and at least on Slackware alsa offered BE variant - picked by speaker-test utility.
I added code to use this or old variant depending on BC_Resources::little_endian and it seems to sound fine now on both outputs?
Can you please test on le ? (usual intel/arm)
participants (3)
-
Andrea paz -
Andrew Randrianasulu -
Phyllis Smith