Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Adobe After Effects Suitable compression format for alpha channel

  • Suitable compression format for alpha channel

    Posted by Joakim Dalfors on July 20, 2005 at 5:36 pm

    Hi!
    What would be the best and most space efficient way to encode an alpha channel to use as a matte for DV clips in e.g. Premiere Pro?

    At the moment I use the Microsoft DV codec with default settings for the actual clip and add another rendermodule with the same codec and settings except for channels where I choose alpha instead of RGB (and no audio). This works fine but since the alpha channel is just greyscale i thought that theoretically it could be saved with almost only a third of the space compared to an rgb clip with maintained quality.
    Is this possible in reality as well and how would I do that?
    (For archieving purposes, space is an issue for me)

    best regards
    Joakim

    Joakim Dalfors replied 20 years, 9 months ago 2 Members · 4 Replies
  • 4 Replies
  • Steve Roberts

    July 20, 2005 at 5:50 pm

    As far as I know, there are no grayscale codecs, only RGB. There are some 256-colour codecs (“graphics” codec?), but they’d have to render within your Premiere timeline and may look like crap.

    However … an alpha-only clip will compress more efficiently in an RGB codec just because it’s usually a lot of solid white, solid black, and some gray. There’s just less information to worry about. So I’d just compress it to the same codec as your timeline, that is, Microsoft DV in your case.

    BTW, I wouldn’t archive motion graphics to DV, since it degrades too much. Archive to Photo-JPEG for better quality, and only use DV when you have to.

    Steve

  • Joakim Dalfors

    July 20, 2005 at 9:14 pm

    Thanks for your input Steve.

    As you might have guessed I am a newbie in regards of compression and in my “world” as a hobbyist I have only “encountered” DV, mpeg, mpeg2, divx and lossless encoding formats.

    Using a DV codec will give the same filesize regardless of information in the picture since it is kept at the same bitrate (3.4 MB/s i think) but I tried the Photo-JPEG you suggested and then the alpha channel was about1/3 or even less in size with the same quality setting.

    The funny thing is that as you indicated, the RGB clip looked better than the corresponding DV version although the bitrate (and thus total file size) was lower, about 2MB/s.
    Whats the drawback with this compression?
    I have been tought that with the DV codec, although not initially lossless, the files can be packed and unpacked time after time with no loss in quality. Is this the case with Photo-JPEG? (I would guess not since normal jpeg pictures loses qulity in repeated pack-unpack)

    Would it be a good strategy to use the DV codec when editing in PPro when there is not to much altering of the clips (especially since most of my raw material is DV footage) and use the Photo-JPEG codec to save all renders from AE?

    Maybe to many questions at once but any input is welcome and your input helped me a lot Steve, so thanks again.

    best regards
    Joakim

  • Steve Roberts

    July 20, 2005 at 9:39 pm

    Ah — constant bitrate with DV. Yep, you’re right. Sorry for the misleading bit o’ advice. Shows how much I use DV. 🙂

    You can view all video files without changing them, because you’re actually just uncompressing them. You don’t recompress them when you stop viewing them. 🙂 I’m assuming that’s what you mean by pack/unpack … otherwise you’re talking about rendering and recompressing, which will degrade any video, some less than others, but DV degrades a fair bit.

    Regarding recompression, whenever you take a video source and render a new video file from that source (either in AE or in an NLE by adding effects or transitions) you are degrading the file by recompressing. AE always renders, so it always recompresses.

    On the other hand, in your NLE, when making a cuts-only video, you are just taking parts of the DV file and placing them into a new file in a different order. However, when you render by adding effects or transitions, you are changing the content of each frame, and therefore recompressing.

    As a general rule, if your footage is DV, leave it as DV. No codec can make it better. If you are making fresh graphics in AE and don’t expect to do anything to them (more effects, transitions) you can render them from AE to the DV codec and drop them into your NLE show with the rest of the DV footage.

    If however, you’re achiving your graphics and want them to stay relatively clean, render the originals to the Animation codec (or the Photo-JPEG codec if you’re short on space). Those two codecs show little degradation after a couple of recompression passes. (I haven’t done exhaustive tests.)

    Hope that helps,
    Steve

  • Joakim Dalfors

    July 20, 2005 at 10:09 pm

    Hi again Steve!
    I am obviously not good at using the right terminolgy because I meant rendering/recompress when I said pack/unpack. And what my former knowledge (or lack thereof) said was what I think you are trying to tell me (i.e. parts with no effects or changes are just “copied” with no loss when editing with the DV codec)

    So thanks a lot for all advices and info. I am definetly going to start using the Photo-JPEG codec for my AE projects.

    Since its is late night here in Sweden I will go to bed now.

    best regards
    Joakim

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy