It is not clear to me how I should write the movie without using
AddMediaSample2.
Is it possible to just add the sample's data to a new empty file, and
remember each samples offset to later create the movie, tracks, media
and at last add the sample references using the offsets and insert
the movie to the file.
Does it mean that I should then write the UInt8 *dataIn that I am
passing to AddMediaSample2 directly to the file?
Shall I simply interleave audio and video data this way?
Regarding the 2GB limitation of AddMediaSample2: the QuickTime
Sequence Grabber DOES support capturing files larger than 2GB over AFP.
Might it be that writing the file as Jan suggests would solve this
problem too?
QuickTime engineers: could you please comment if this is how the
Sequence Grabber writes movies - is this the solution to the problem?
Best regards,
Mark
All of the above is working, but AddMediaSample2 fails as soon as the
file reaches 2 GB if capturing over an AFP share - when using the
sequence grabber to write the file there is not such a limitation.
Can't help you here. I remember reading somewhere that Tiger has AFP
3.0 which has no 2GB limit but in my experience this is simply not
true.
Also, when using the sequence grabber to write uncompressed (YUV422)
samples, activity monitor reports as much as 40MB/s throughput. When
using my custom writer, throughput never gets above 20MB/s - what is
actually limiting my frame rate.
It is faster to write the data directly to a file asynchronously and
create the movie at the end using AddMediaSampleReferences.
I used to do this on Mac OS 9 using completion procs. These days in
Carbon you create an MP thread and wait on a queue, then call
FSWriteFork.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
QuickTime-API mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/quicktime-api/email@hidden
