Did a recent upgrade to Debian Etch from Debian Sarge.

Everything is working fine except I'm no longer able to get .aiff files to play back. Houdini wouldn't even load audio files under linux i686_gcc3.4 h8.1.866, (as noted in a previous post)… however, it does load audio files under h8.2.97… but no audio.

No audio from aplay either.

Other linux media programs play the .aiff files without any problem.

Any ideas?

O.K. So, I guess no one else is having a problem playing audio files in Houdini on linux.

Since the previous post I've been trying .aiff and .wav files saved with different sampling rates, and various settings within Houdini… without success. I found that if I reboot the machine, then aplay WILL play back the .aiff and .wav files without any problem. But once I start up houdini aplay no longer works.

There are no error messages from houdini or aplay.

Again, other linux multimedia programs play the audio files just fine.

Debian Etch/KDE 3.5.5
Houdini 8.1.866 & 8.2.97

Any help or suggestions appreciated… can't work without the audio track.

Did you try changing the audio format with sox or something similar? Doesn't Houdini accept pcm or wav now? Can't remember. I did just check trying to read an aiff right now with 8.2.91 and it indeed failed - but I seem to recall that recently certain sample rates with aiff would cause problems in Houdini.

Cheers,

J.C.

Hello,



just curious, but how can you load an audio file in Houdini…?



thanks




bern

Bernard
Hello,

just curious, but how can you load an audio file in Houdini…?

bern

Go into Chops, lay down a file chop and read in a .aiff or .wav (or other supproted formats). Set the audio flag. Bring up the audio panel to enable audio, set volume etc.

You can also specify an Audio file directly in the Audio Panel, if you're just using it as animation reference (see the bottom of the panel).

twod
You can also specify an Audio file directly in the Audio Panel, if you're just using it as animation reference (see the bottom of the panel).

Tried that too.

I've managed to get audio files playing back under houdini on linux. I'm not sure about the technical details, but it has something to do with the linux “artsd” process. Whenever this is running, I can't hear audio from aplay or houdini.

I've shortened the “Auto-suspend if idle after” time under the KDE Sound System “General” controls… this shuts down artsd sooner… allowing houdini audio output.

However, once you save or do anything that causes audio feedback from within houdini, artsd starts up again and you will no longer be able to hear the audio track. So, I went back to the KDE Sound System “Hardware” controls and set “Select the audio device:” to “No Audio Input/Output”. I know this isn't the best solution, as it probably disables other audio functions, but at least it keeps artsd from starting up and I can animate to an audio track.

If there's a more elegant and/or appropriate solution to the artsd “problem” please let me know.

My apologies for assuming this was a Houdini bug.

haven't tried on linux yet, windows works fine (I'm starting to develop some audio tutorials with houdini etc).

I've noticed that the audio import data caps at 44100, stereo. Does anyone think it would be an easy or hard etc for future versions to support higher sampling rates … like up to 384k for samleing, and 32 bit float for dynamic range?

Some synthesiser functions could actually use the resolution and now that computers have more than 256meg of ram it's practical.

Houdini chops seem to work at 32bit float natively anyway.

One thing I can verify, btw, is that the latest 8.2 on Linux does indeed seem to kack on trying to read many aiff files, and I suspect the sampling rate is the problem. You post indicates that Houdini is capable of resampling on the fly, so perhaps the aiff reader is a little busted?

Personally I only ever need a certain ceiling of sampling for all the day to day work I use Houdini for(frankly, just getting in the audio for animation), so I don't have a need for the higher overhead. But yes, assuming you're trying to do any serious audio work I can see it.

Oh, and also - on Linux, if you ever get stuck with aiff files as Floyd did, I've also verified that sox works great - many distros don't install sox by default but it's terribly useful and it's likely there somewhere in the list of available apps. Converting to a .wav file worked for any problems I had(including the aiff files that Houdini won't read).

Cheers,

J.C.

Yeah I'm completely understand that there isn't really a need for higher sampling rates right now (except for maybe 48000 from DV cam). If it's typing in 192000 instead of 44100 than by all means at least one person would appreciate it but I suspect it wouldn't be quite that simple.

But I'm thinking in terms of eventually using Houdini for Tonal pattern recognition and isolation, as well as realistic spatial audio (not just fx) … also to sell Houdini to mastering Houses that need very specific functions of pipeline tools and aren't afraid of "tech … these would definately be in the area of noise reduction and tone isolation, which would be somewhat in the higher and lower frequencies which 44100 doesn't really provide for.

Houdini loads all audio into RAM - so even with a bunch of ram and clever use of CHOP unloading, it's not really adept at work with larger audio track numbers and lengths. I tried what you're doing, and ended up using Houdini for generating control rate data (100-1000hz) and then funneling that control data into an outside synthesis package called Csound to do all the high sample rate processing. The Csound community has probably covered a lot of the problems you're looking to cover - and it's free! Here's a description (in badly dated marketing guise) of what I did:

http://www.users.bigpond.com/tstex/pan.htm [users.bigpond.com]

The whole “Pan” system is available for download on odforce codex “others” when it comes back. The interface was pretty clunky - what with there being no CHOP Hda's or python back in version 4. With python coming in v9 you'll probably be able to generate the interface more easily than I was able to with the hscript-perl-csound cludgeflow.

There's these as well:
http://www.users.bigpond.com/tstex/research.htm [users.bigpond.com]

Cheers

Wow very interesting, I'll have to look those over! Yeah I know about the Ram etc, but I think pretty soon this will be a non-issue (or at least by the time I have all my systems worked out).

It will get cheaper and cheaper and systems will soon easily be able to accomodate more than enough for serious audio processing in Houdini.