Log in

View Full Version : 32bit floating point wav


datascab
9th February 2004, 16:49
Hi,

When i try to use wavsource it says ACM failed to find decompressor or something.

When i try directshowsource it says the filter graph wont talk.

Can i not use 32bit PCM streams within avisynth?

I added the stream manually in vdubmod and it shows ok in there.

Thanks

Data

sh0dan
9th February 2004, 17:11
Could you post a second or two of it (preferably with some sound in it).


DSS should be able to open it, if it can be played correctly by media player, zoomplayer, etc - not entirely sure about WavSource, though.

(32 bit might not be a bit tricky, as there might not be any way to tell if it is 32bit int samples or 32bit float samples.

datascab
9th February 2004, 17:34
Hi,

Thanks for the quick reply.

These waves were created with foobar v8.7b with the floating point 32bit option.

I just checked, the 1st wav out of the 12 does play in WMP9, MPC but pops in winamp.

The others dont play at all.

Most of the wave editor programs i have wont open any of the files at all.

I might have a look at how the files were written 1st and then get back to you.

Thanks

Data

OK, i just downgraded to the version 7 of foobar so i could write a 32bit fixed point wav.

This is ok, no popping in winamp, opens using wav source etc.

So the problem seesm to be when the wav is floating point.

Now i'll have to find out what the difference is, LOL.

My conversion is ac3 - wav - mp4(aac) and I wanted the intermediate wav to be higher than 16 bits since I read it will give a better output.

Thanks

Data

sh0dan
9th February 2004, 17:51
You might be better off with 24bit integer samples (unfortunately). You _should_ be able to open these with DirectShowSource.

It looks like I should have a look at vdub's WAV-support, as maximum seems to be 16bit/2 channels. 24bit / unlimited channels _should_ be possible with WAV-files.

Edit: Oh - you put in some more data. As far as i can tell there isn't any rock-solid way of determining wether a WAV-file is int or float. But I'll have a look at it, if time permits.