NNenov

Ello, I was wondering, is it possible to work with videos in the same way that we can with audio buffers? e.g. store a video in ram in a buffer, and have multiple players be able to access it and output different segments simultaneously. 

I'm trying to make a polyphonic video player where the source video is shared in memory. I'm not sure if its possible if the video is long. I saw one thread suggesting to put jit.movie in a poly object, but jit.movie, with viddll engine seems to store the video per instance? so you couldnt just spawn poly instances of it each doing loadram. Im hoping theres an obvious approach I just missed. Any advice appreciated!

Ello, I was wondering, is it possible to work with videos in the same way that we can with audio buffers? e.g. store a video in ram in a buffer, and have multiple players be able to access it and output different segments simultaneously. 
I'm trying to make a polyphonic video player where the source video is shared in memory. I'm not sure if its possible if the video is long. I saw one thread suggesting to put jit.movie in a poly object, but jit.movie, with viddll engine seems to store the video per instance? so you couldnt just spawn poly instances of it each doing loadram. Im hoping theres an obvious approach I just missed. Any advice appreciated!

Ello, I was wondering, is it possible to work with videos in the same way that we can with audio buffers? e.g. 

reading-from-same-video-memory-with-multiple-"play-heads"

inside a poly~ everything loads per voice.

buffer outside - playheads inside.


if using viddll engine with loadram you can create this kind of video buffer structure allowing for multiple playheads. if using with output_texture enabled, an extra texture copy is needed on the playback end. here is a basic example of the logic needed. the playback module could easily be stuffed into a poly~ abstraction.

inside a poly~ everything loads per voice. buffer outside - playheads inside."

This is the workflow I'd like to do, but afaik the buffer is stored inside the jit.movie object, which afaik is the "playhead". although I see from what Rob Said below my idea of playhead is abstract, instead I could send read frame messages from the poly object and execute them per frame (if I understood correctly).

This is interesting, thank you for setting up the example, I think I follow:

Every bang, the frame from each desired segment is sent to each destination texture, and you don't use the jit.movie playback features (start stop etc). Is this the best technique, performance wise if using a bunch of objects for routing each frame per frame? It makes sense though, I cant imagine another approach now.

inside a poly~ everything loads per voice.  buffer outside - playheads inside."

This is the workflow I'd like to do, but afaik the buffer is stored inside the jit.movie object, which afaik is the "playhead".  although I see from what Rob Said below my idea of playhead is abstract, instead I could send read frame messages from the poly object and execute them per frame (if I understood correctly).


"ROB RAMIREZ

if using viddll engine with loadram you can create this kind of video buffer structure allowing for multiple playheads. if using with output_texture enabled, an extra texture copy is needed on the playback end. here is a basic example of the logic needed. the playback module could easily be stuffed into a poly~ abstraction."

This is interesting, thank you for setting up the example, I think I follow:
Every bang, the frame from each desired segment is sent to each destination texture, and you don't use the jit.movie playback features (start stop etc). Is this the best technique, performance wise if using a bunch of objects for routing each frame per frame? It makes sense though, I cant imagine another approach now.

reading from same video memory with multiple "play heads"