Re: [Mjpeg-users] Mplex lpcm bug+fix (1.8.0)

2005-10-12 Thread Steven M. Schultz
On Tue, 11 Oct 2005, bn wrote: > Below is a patch to fix this (excuse the word-wrap, > courtesy of yahoo mail; I'll attach it separately as > well), by adding a 'whole_unit' variable to class The attached patch worked fine. Not necessary to include both a yahoo version and an at

Re: [Mjpeg-users] Mplex lpcm bug+fix (1.8.0)

2005-10-12 Thread Steven M. Schultz
On Thu, 13 Oct 2005, Dave Chapman wrote: > My understanding from examining 20-bit DVD-Audio AOB files (authored > using a commercial DVD-A authoring package) is that they are packed in a > similar way to 24-bit. It would appear reasonable enough then to extend the patch to treat

Re: [Mjpeg-users] Mplex lpcm bug+fix (1.8.0)

2005-10-12 Thread Dave Chapman
Steven M. Schultz wrote: > Does anyone know how 20bit LPCM audio is packed? According to > http://www.dvddemystified.com/dvdfaq.html#3.6.2 LPCM of 16, 20 or 24 > bits (at either 48k or 96k) is allowed. My understanding from examining 20-bit DVD-Audio AOB files (authored using a

Re: [Mjpeg-users] Mplex lpcm bug+fix (1.8.0)

2005-10-12 Thread Steven M. Schultz
On Tue, 11 Oct 2005, bn wrote: > I've just been experimenting with 1.8.0's vastly > improved lpcm capabilities, and seem to have found one more problem: > > Mplex's 24-bit lpcm PES packet payloads aren't aligned > to the nearest fully-resolved sample boundary, e.g. a... 24 bit lpcm cuts

Re: [Mjpeg-users] Wrks with some Jpegs. Does not with others.

2005-10-12 Thread Bernhard Praschinger
Hallo > Has anyone face this issue before? > > - > I got the jpeg2yuv.exe directly from somewhere. > > If I use a jpg created from PaintBrush of Microsoft, the folowing line works > fine. > C:\Program Files\XYZ>jpeg2yuv -f 25 -I p -j food.jpg | mpeg2enc -o fo

[Mjpeg-users] Mplex lpcm bug+fix (1.8.0)

2005-10-12 Thread bn
I've just been experimenting with 1.8.0's vastly improved lpcm capabilities, and seem to have found one more problem: Mplex's 24-bit lpcm PES packet payloads aren't aligned to the nearest fully-resolved sample boundary, e.g. a typical packet, (subtracting headers) works out to, in the case of 24-b