That something is possible doesn’t necessarily mean you’ll find someone willing to donate their time and effort to make it happen if it isn’t something they would need/use themselves. As mentioned, hanatos does have an interest in video, so that project may be more aligned with your needs.
Other than that I believe you can trigger a lua script after each image export so it might be possible to use ffmpeg to add it to a growing container and delete before starting the next export, meaning you wouldn’t have all images on disk at the same time. @wpferguson ?
right, vkdt has an ffmpeg output format you can select in the gui that will write h264 streams. it also reads .mov containers or magic lantern .mlv raw video for real-time display/processing in the gui, but it sounds like the image sequence mode is more what would be required by the op here? the raw loading module supports reading IMG_%04d.DNG kinda filenames and will load a bunch of images as animation. you can grade/keyframe this in the gui and then export to h264 via ffmpeg.
would be curious to hear how that works for you if you try. i suppose the ffmpeg output format might require some extra parameters exposed in the gui to write exactly the required format or so.
as i recall, i did tried that lua script in the past, but was not beneficial because it does something that i want to avoid…
I dont want to render images and after create the clip, what i want is to straight send from DT into mov container.
In this moment the workflow is: i am exporting as exr that can hold 16bit files where quality is lossless(but always needs more space and time), if i would choose jpegs quality would be limited to 8 bit which would be a fail.