Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Compression Techniques compression yields artifacts

  • compression yields artifacts

    Posted by Dorit Grunberger on February 8, 2011 at 8:25 pm

    Hi all

    The setting:
    I have a 1hr, 20 minute SD project ending on DVD. MacPro, 8 core, 12GB RAM, OSX 10.5.8 and FCP Studio 3. The project is divided into 6 chapters and all the footage is shot in a studio. The footage is of a Tai C’hi practitioner and shot over the course of several month under different lighting conditions.

    The problem:

    There are several places that result in visual artifacts upon compression (best 90 minutes for DVD setting in Compressor). I have tried all kinds of things to try to eliminate the artifacts, but at this point I’m pretty much stumped. The problem seems to appear when the subject is slowly moving in front of 8 identically sized dark picture frames with off white mattes and black calligraphy hanging on an off-white wall…Hope you can wrap your brains around that one :-).

    The Questions:

    1. Are there situations where no matter what, you can’t resolve such an issue?
    2. any other brilliant ideas?
    3. And while I’m at it, Compressor has decided to fail all my compressions (not related to the above problem at all), even ones that worked just fine a day ago. Error -120. Any help with that would be swell as well.

    Many thanks as usual,

    Dorit

    Thanks in advance,
    Dorit

    Jeff Greenberg replied 15 years, 2 months ago 4 Members · 13 Replies
  • 13 Replies
  • Chris Blair

    February 9, 2011 at 12:43 am

    If you’re using those settings you really shouldn’t be seeing artifacts. It would also help to know more specifically what you’re seeing when you say artifacts. Is it blockiness, pixelation, stair-stepping?

    The only thing I can think of that could cause issues when compressing using modern MPG2 codecs is incorrect field order, or resizing from 720×486 to 720×480 instead of just cropping off 3 pixels top and bottom to go from 486 to 480.

    Keyframing can cause issues in compression sometimes but many MPG2 codecs don’t even concern themselves with keyframing, especially if you use a preset, which is often optimized for DVD use.

    Chris Blair
    Magnetic Image, Inc.
    Evansville, IN
    http://www.videomi.com
    Read our blog http://www.videomi.com/blog

  • Dorit Grunberger

    February 9, 2011 at 2:28 am

    Hi Chris. Perhaps I can send you a little snippet to view. It’s pretty hard to describe, hence the less than satisfying “artifact” description. The footage where this occurs was shot as progressive, but the entire project is a mixed bag of 16:9 and 4:3, 23.98 and 29.97.
    I’m pretty sure the frame controls in Compressor were off so the output fields would be same as source, right?

    Let me know if I can send you something and if so where shall I send it to.

    Many thanks,
    dorit

    Thanks in advance,
    Dorit

  • Mark Spano

    February 9, 2011 at 9:29 pm

    [dorit grunberger] “Compressor has decided to fail all my compressions (not related to the above problem at all), even ones that worked just fine a day ago. Error -120.”

    Don’t know what that error means, but any time Compressor errors out on my compressions, I usually just go into the app, click on the Compressor menu, and choose Reset Background Processing. From there, I choose Reset and cancel jobs. Then I’m good to go.

  • Chris Blair

    February 10, 2011 at 1:41 am

    You could upload a snippet right here on the cow and actually embed it into your post I believe. That would be easiest as others on the list could view it and help with feedback.

    However, if your project is a mix of different aspect ratios and frame rates, that’s likely your problem. When we are forced to use mixed footage, we typically decide which format is used the most, then convert all the other clips to that format, aspect ratio, interlacing and frame rate. Yes…it’s a pain, but it’s the only way to eliminate artifacting when you export for final delivery. It’s no different than color correcting or grading a project clip by clip, which we often do on projects with decent budgets.

    So if you have clips in your timeline shot at 30p, and clips shot at 24P with pulldown and clips shot at native 24P (no pulldown), and clips shot interlaced, you end up relying on Final Cut for conforming all the clips. It may or may not do a good job, depending on your original footage, your timeline settings and how you interpret the footage.

    If there are only a few offending clips, you could always convert just those clips to match (frame rate, interlacing, aspect) your other footage.

    Chris Blair
    Magnetic Image, Inc.
    Evansville, IN
    http://www.videomi.com
    Read our blog http://www.videomi.com/blog

  • Dorit Grunberger

    February 10, 2011 at 1:48 am

    Thanks Mark. I decided to restart the computer before doing any resting in Compressor and for some unknowable reason, it decided to work again…

    Thanks for your input

    Thanks in advance,
    Dorit

  • Jeff Greenberg

    February 10, 2011 at 12:27 pm

    [dorit grunberger] “1. Are there situations where no matter what, you can’t resolve such an issue?
    2. any other brilliant ideas?”

    1- yup. But if your stuff is as slow moving as you say…it should be good.
    2- You mentioned your were mixing all sorts of formats and sizes…what’s the sequence type in FCP – this could very well be part of your issue. Additionally, what method are you using to get from FCP to Compressor?

    Best,

    Jeff G

    Apple Master Trainer
    Avid Cert. Instructor DS/MC
    Avid & Color Videos Vasst.com
    Compressor Essentials Lynda.com

  • Dorit Grunberger

    February 10, 2011 at 8:51 pm

    Hi Jeff
    Here’s a blow by blow (1hr, 20min Tai C’hi video):

    1. I received QT .movs from a camera shot at 16:9, DV/DVCPRO-NTSC, lower field dominance.
    2. For newer material, I used our Panasonic DVX100A, shot in 24p advanced (progressive).
    3. For newest material shot with 2 cameras (and because I didn’t pay enough attention, rental place gave me a DVX100 that couldn’t do 24 at all), was shot as 4:3, DV/DVCPRO-NTSC, lower field dominance.

    So those are the 3 different format/sizes (all SD).

    4. All sequence settings in FCP were : QT Video settings (compressor): Apple ProRes 422 (HQ) – overkill…, Video processing: Render 10 bit, Motion: Best

    To get around some of the issues, I compressed the project from 7 partial sequences. All the 23.98 were compressed, (best 90 minutes for DVD) separately, and was not problematic. All the 29.97 were compressed in small groups, because I was concerned that the length of the project was getting a little too close to 90 minutes, and thought it might be introducing artifacts that way.

    In all cases, I exported uncompressed, QT movies and submitted those to Compressor.

    Thanks in advance,
    Dorit

  • Chris Blair

    February 11, 2011 at 1:56 am

    dorit grunberger wote:
    1. I received QT .movs from a camera shot at 16:9, DV/DVCPRO-NTSC, lower field dominance.
    2. For newer material, I used our Panasonic DVX100A, shot in 24p advanced (progressive).
    3. For newest material shot with 2 cameras (and because I didn’t pay enough attention, rental place gave me a DVX100 that couldn’t do 24 at all), was shot as 4:3, DV/DVCPRO-NTSC, lower field dominance.

    I think the above mish-mash of formats, aspect ratios and frame-rates is likely what’s causing your issue. As I mentioned before, the only way we’ve ever had success fixing final output problems is by grading the whole project and making all the shots match in terms of pixel aspect, frame-rate and interlacing. You’ve got 60i (interlaced) material at 16:9, 60i (interlaced) at 4:3 and 24P (with no pulldown) at whatever aspect you shot it at. So when you go to export, if you pick progressive, you’re interlaced material is probably going to look “stroby.” If you pick interlaced, your progressive material is going to look bad. If you pick 29.97fps, your 24P stuff is likely going to look odd etc.

    Sure..Final Cut can conform your material for you in the timeline, but it’s an editing app, not a high-end format conversion program. The best method to convert footage at high quality is with compositing and color grading programs. So my advice is still figure out which format, size, frame rate is used the most, and shot by shot…conform everything else to that if at all possible.

    Chris Blair
    Magnetic Image, Inc.
    Evansville, IN
    http://www.videomi.com
    Read our blog http://www.videomi.com/blog

  • Dorit Grunberger

    February 11, 2011 at 2:06 am

    Hey Chris

    I don’t see how the mix of different formats is the issue as (like I’ve already explained), the different formats were each compressed separately, i.e. not on the same timeline at all. I will try to embed a snippet and that might be an easier way to address this. I am very interested in what “compositing and color grading programs” you might recommend. I’m not clear what you mean by “conform everything else to that if at all possible.”

    How do I assess if it is or is not possible, and would I do it in Compressor?

    Thanks

    Thanks in advance,
    Dorit

  • Jeff Greenberg

    February 11, 2011 at 1:20 pm

    Wow, I have no nice way of saying this; you have a mess on your hands. I don’t know if it’s fixable ‘as is.’ I’m going to put the areas that are of concern to me; I hope it’s helpful. I think that the problem is you were mixing frame rates; and to fix it might not be worth your effort.

    * 1. I received QT .movs from a camera shot at 16:9, DV/DVCPRO-NTSC, lower field dominance.
    This makes it sound like you didn’t capture it yourself. This could have been handled incorrectly. Did you do it? Or did someone else?

    *2. For newer material, I used our Panasonic DVX100A, shot in 24p advanced (progressive).
    You’re now talking about mixing progressive material with interlaced material. In a 30i timeline this does different things than a 24 timeline. Generally, FCP is so-so at best at mixing frame rates in the same timeline.

    *3 3. For newest material shot with 2 cameras (and because I didn’t pay enough attention, rental place gave me a DVX100 that couldn’t do 24 at all), was shot as 4:3, DV/DVCPRO-NTSC, lower field dominance.
    You’re now mixing more interlaced material – and if you don’t want pillar boxing (black bars left and right) you blew up the footage.

    So, after thinking about this for a bit, I’d suggest as a test, you send the most problematic footage to compressor and see how it looks. Please note – I didn’t say put it on a timeline, I said, send it directly to compressor. You can use in/out points to make the compressed job smaller.

    Basically, what I’m suggesting is that you test the compression by taking the workarounds you (and FCP) have done, out of the picture. If you send the problematic areas to compressor and use the same settings as you did prior and it works, we’ve learned something ; and if it fails, you’ve learned even more. No exporting uncompressed movies – just Send To->Compressor.

    This should take all of about 10-20 min (far less than I spent thinking on how to identify the problem.)

    I suspect that your timeline settings (the frame rate), the mixing of different formats and your workflow (exporting and sending to compressor) are all parts of the problem; I want to boil it down to the simplest possible test and work our way backwards from there.

    Best,

    Jeff G

    Apple Master Trainer
    Avid Cert. Instructor DS/MC
    Avid & Color Videos Vasst.com
    Compressor Essentials Lynda.com

Page 1 of 2

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