VideoHelp Forum

Our website is made possible by displaying online advertisements to our visitors. Consider supporting us by disable your adblocker or Try ConvertXtoDVD and convert all your movies to DVD. Free trial ! :)
+ Reply to Thread
Page 66 of 66
FirstFirst ... 16 56 64 65 66
Results 1,951 to 1,969 of 1969
Thread
  1. Youtube generally doesn't accept x264 lossless.
    Quote Quote  
  2. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    I'm also a bit unsure how to get better quality with my videos. They've always been pretty good but I notice that going to 480p does not yield an improvement any more, in some cases videos with a lot of motion look more stuttery at the higher setting. The video is interlaced and I don't like how youtube tries to make it progressive scan.
    this was added the past August. the raw video file was only 683 mb.
    http://www.youtube.com/watch?v=on3-ArMfpnI

    In Sony Vegas 4.0, I rendered it as a Windows Media Video V9 (*.vmv). Audio Settings were:
    Mode: Quality VBR
    Format: Windows Media Audio 10 Professional
    Attributes: VBR Quality 98, 96 kHz, 2 channel 24 bit VBR

    For video-
    Mode: Quality VBR
    Format: Windows Media Video 9
    Image Size: DVD quality (640x480)
    Frame Rate: 29.970 NTSC Seconds per keyframe: 1
    Quality: 93%
    The rendered file on my computer looked pretty good but it did seem shrunk a little like there was a thin black line around it all. Also, trying to open one of these rendered files in vegas itself will cause the program to close. Any idea why that is?

    for one really short video, I just rendered it (also on Sony Vegas 4.0) as an avi file with the same NTSC frame rate, lower field first and NTSC DV for the video format. This type of rendered file will open fine in Vegas.
    The file size is 793 mb and video size is 720x480 but even that didn't look as good as I hoped, once again the high quality settings didn't seem to improve anything. Even the still photos at the end appear blocky and there's nothing moving at that point.
    http://www.youtube.com/watch?v=xVPL8s1fNKo

    I know the file size limit is 2 gb but my videos don't come close to that so I really can't understand why they don't look any better with the higher viewing setting anymore. So yeah, these look decent but I suspect there's a lot of room for improvement.

    But the main reason I really want super high quality is because of the contest with cosponsored by ESPN to submit your best original sports clip on youtube. Most of the entries look pretty clear and vibrant and it's not like they appear to be made by any professional companies, this is your usual average dad holding the camcorder deal.
    http://www.youtube.com/ESPN
    So if they can just copy the video and upload in a high quality, why can't I?
    Thanks.
    Quote Quote  
  3. BuskerAlley.com zoobie's Avatar
    Join Date
    Feb 2005
    Location
    Colorado Rocky Mountains
    Search Comp PM
    I suppose a feature length movie on youBoob is out of the question...
    Quote Quote  
  4. Originally Posted by Knightmessenger View Post
    The video is interlaced and I don't like how youtube tries to make it progressive scan.
    It's your job to deinterlace it before uploading. YouTube doesn't deinterlace at all.
    The file size is 793 mb and video size is 720x480 but even that didn't look as good as I hoped,
    Don't you notice the aspect ratio is wrong? People are way too tall and thin. You uploaded a 720x480 video and it plays in that same incorrect ratio. It can be fixed by adding a tag as it looks to me like it's supposed to be 16:9. Or you can resize it properly before uploading it. Also, your whites are completely blown out and the blacks horribly crushed. Besides looking horrible, it makes it harder to compress properly. And the deinterlacer used was poor.
    So if they can just copy the video and upload in a high quality, why can't I?
    Maybe they know what they're doing? And some of the ones featured at the ESPN channel don't look so hot. Some have a lot of aliasing remaining after being deinterlaced poorly.
    Last edited by manono; 2nd Nov 2010 at 19:36.
    Quote Quote  
  5. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    Originally Posted by manono View Post
    Don't you notice the aspect ratio is wrong? People are way too tall and thin. You uploaded a 720x480 video and it plays in that same incorrect ratio. It can be fixed by adding a tag as it looks to me like it's supposed to be 16:9. Or you can resize it properly before uploading it. Also, your whites are completely blown out and the blacks horribly crushed. Besides looking horrible, it makes it harder to compress properly. And the deinterlacer used was poor.
    The spirit week video is 4:3 like all of my videos. The 720x480 does look a little wider than the 640x480 in my other videos but what's the difference. 720x480 is still 4:3, right? There were options to render it as widescreen in Vegas but I didn't select those.

    The camera used was a Sony video 8mm TRV-32 which had really good color (often almost as good as a 3ccd) but did not have a manual exposure option. So yeah, it was slightly overexposed but I didn't really know what to do about that, especially back then in 2003 when I knew a lot less about video. I don't know why you think the blacks are crushed, I've played this clip on tapes, dvd and many different screens and never felt the shadows were too dark. When I uploaded it last year, I barely made any adjustments.

    I did not deinterlace the video either. I rendered lower frame first because when I've tried to convert it to progressive scan in Vegas, it has come out weird with the frame rate looking jittery, especially with transitions or wipes. Now, the 480p option makes most of my videos look more pixelated and shows more jaggies but could this be the result of the lower resolution blurring that stuff away? Why doesn't the 480p button on youtube not improve the playback like clicking the HQ button used to? Are they not the same thing?

    Are there better settings I could use in Vegas? Or is there something I could download that would give good deinterlacing or better compression for youtube?
    Quote Quote  
  6. Originally Posted by Knightmessenger View Post
    720x480 is still 4:3, right?
    No, it's 3:2. 720/480=1.5.

    But I was wrong about it being 16:9 and people being too tall and thin. As you say, it's 4:3 but you're not uploading a 1.33:1 ratio video, and it's not being played as a 1.33:1 video, but as a 1.5:1 video and people are too fat (not too thin as I said before). And as I said, it can be fixed in one of two ways. One is to add the tag:

    yt:stretch=4:3

    http://www.google.com/support/youtube/bin/answer.py?hl=en&answer=146402

    The other is to do a proper crop and resize before encoding. Here are before (top) and after (bottom):
    Image Attached Thumbnails Click image for larger version

Name:	Dance11.png
Views:	335
Size:	363.7 KB
ID:	4154  

    Click image for larger version

Name:	Dance2.png
Views:	788
Size:	339.4 KB
ID:	4155  

    Quote Quote  
  7. Originally Posted by Knightmessenger View Post
    The camera used was a Sony video 8mm TRV-32 which had really good color (often almost as good as a 3ccd) but did not have a manual exposure option. So yeah, it was slightly overexposed
    Not just slightly. It's horrible. I use AviSynth and not Vegas. The brightness and contrast need to be corrected. Also, the deinterlacers in these NLEs are usually pretty bad. Here's a pic showing the whites (green) and blacks (red) outside of 16-235. All shadow detail is lost as what should be dark grey is solid black. Afterwards is a pic showing some improvement, but it can't be entirely fixed at this point. Does Vega have a histogram with which you can check the white and black levels? It's very obvious when you put one on. Heck, I don't even think the colors are any good as skin tones look kind of 'rosey' to me.
    Image Attached Thumbnails Click image for larger version

Name:	Dance3.png
Views:	812
Size:	364.1 KB
ID:	4156  

    Click image for larger version

Name:	Dance4.png
Views:	759
Size:	345.8 KB
ID:	4157  

    Quote Quote  
  8. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    I don't think my version of Vegas has a histogram. But what probably happened is some areas were in the stage light so they were overexposed while other areas outside of the light were underexposed, hence the crushed blacks. And that black space between the projector board and gray curtain should be black, it's the back of the stage without a lot of light in that area. To expose that properly, the camera would have had to wash out everything else in the picture.
    To be honest, I'm surprised you are spending as much time on this video. Sure it might not look that great but it's still better than your typical auto focus/exposure youtube video. I suppose I could have lowered the gain further and boosted the shadows (in Vegas) before uploading it like you did with the image but I didn't want to risk adjusting it too much that contrast was lost. Also, you adjusted one screen cap off youtube, I adjusted the whole video so I had to be aware of what it would do for every frame.
    The yt: stretch 4:3 will probably make it too narrow also.

    Besides, I wasn't really asking about the video itself, I was wondering about the blockiness and artifacts introduced by the youtube compression. I want to avoid blockiness with my next upload. So if you knew of any ways to render the video that youtube will not junk up as much, it would be most helpful.
    For example, this is another video uploaded by somebody else and shot in the same auditorium. I'm pretty sure she didn't use any special encoding or compression techniques to max out the youtube quality yet this video does not look as blocky as mine.
    http://www.youtube.com/watch?v=-0hxOuaippo
    Any possible explanation for that?

    And I'm still not sure if the 480p mode is the same as what the HQ setting used to be on youtube.
    Last edited by Knightmessenger; 10th Nov 2010 at 15:14. Reason: mentioned other video
    Quote Quote  
  9. You have to optimize it for the low bitrate youtube compression

    1) Denoise it . There is horrible chroma and other noise - Noise is the detrimental to compression and requires a lot of bitrate to encode

    2) Stabilize it . Motion is detrimental to compression. Hand held jerky stuff is the worst and "eats up" a lot of bitrate

    3) Add intros/outros - Still image titles at the front and end can improve bitrate distribution

    4) Upscale to 720p using a good algorithm (e.g. avisynth nnedi3) - This is one case where upscaling is beneficial - It allows you to access higher bitrates on youtube (even the 480p version of the 720p encode looks better than the normal 480p version)

    5) Better quality deinterlacer (e.g. avisynth qtgmc) - Vegas leaves residual jaggies from dropping a field and resizing the other, with no interpolation or antialiasing. These jaggies "eat up" bitrate as well


    PS. everything manono said is spot on . And those girls won't like it that you made them "fat"
    Quote Quote  
  10. Originally Posted by Knightmessenger View Post
    Sure it might not look that great but it's still better than your typical auto focus/exposure youtube video.
    And how is that an excuse to upload crap? Because everyone else does it? You came here asking for help so apparently you do care.
    Also, you adjusted one screen cap off youtube, I adjusted the whole video so I had to be aware of what it would do for every frame.
    What I said applies to everything, and applying the fix I did for that frame would have improved the entire thing. I could have pointed out much worse frames, the ones with more white. The black levels can be fixed; the white levels are too far gone to fix entirely. Sure, the stage lighting plays some role, but most of the fault is yours and the camera's. Besides, I thought the rappers just awful and I was intrigued by the blond amazon in the girls' part.
    The yt: stretch 4:3 will probably make it too narrow also.
    Why make a claim like that until you've tried it? Did you bother to read the link? So many people do what you did (upload unresized 720x480 videos) that they came out with a fix.

    pdf has answered the questions about the blockiness. When uploading handheld camera stuff you're just asking for poor quality YouTube results. Use a tripod if possible or a camera with a built-in stabilizer or, as he suggests, 'deshake' it. The more movement in the individual frames (not even a stable background with yours), the more encoding artifacts you'll get. Use a better deinterlacer. Denoise it. All will help to improve the compressibility and the resulting quality, given the low bitrate YouTube uses. And I didn't really find the other video to which you linked any better.
    Last edited by manono; 13th Nov 2010 at 19:42.
    Quote Quote  
  11. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    Originally Posted by manono View Post
    And I didn't really find the other video to which you linked any better.
    You didn't think it looked less compressed?

    If there isn't a denoise function, deinterlace or stabilization filter in Vegas or something similarly easy to use, I would not know how to do it. And don't post production stabilization just smear the picture out and lose detail? Also, I believe I had the steady shot function on the camera turned on.
    I do not know how to use script like avisynth.

    If I add a yt:function stretch, does that mean my video will be permanently set that way, or there is no way to undo the change (without further quality loss) if I think it looks worse? If the yt: stretch 4:3 is designed for making a 720p video a little more narrow, what is the function that takes a 4:3 video stretched to fill the 16x9 frame and restores the pillar bars?

    Also, why do I keep getting emails when somebody replies to a post I make? I thought I turned that setting off.
    Quote Quote  
  12. Don't know anything about the emails.

    That particular tag is for properly resizing 4:3 720x480 material that's been uploaded to YouTube. You put it in the 'Tag' part, like the others such as
    No reencoding is done. The YouTube player just resizes it at playback. You'll still have the black bars on the sides, just a bit more of them. If you don't like what it does, just remove the tag. If the video is 16:9 and 720x480 (which it isn't), then you'd use a different tag to fill the screen. If you want to zoom it to fill the player, losing the top and bottom of the video, then use the yt:crop=16:9 tag. I'm not entirely sure that'll work to improve the aspect ratio. It's meant mainly for widesreeen videos uploaded with black bars above and below. I have a few from before YouTube began using a widescreen window and did that to them to get them to fill the screen. And I don't know if you can combine the two tags, one to fix the aspect ratio, the other to zoom the video. You could always try. Like I said, it's easy to undo. 720p stuff (1280x720 or 960x720 ) shouldn't need a tag as it's already at the proper aspect ratio.
    Quote Quote  
  13. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    Yeah, I understand the tags, I just didn't realize how they worked. But as long as nothing permanent is done to the video, I think I might try it.

    But I understand that my video might not be optimized for youtube compression, is there more I can do with encoding? Should I try to make the file size smaller or is that not as big of a factor? Does a good Mpeg 2 file turn out better than an avi? Was the old HQ setting the same as the 480p now or are they different?

    Basically, how can I optimize for youtube in ways that don't involve advanced video filtering or adjustment, just rendering and encoding?

    Does every video on youtube get allocated the same bitrate per frame? Or does it try to give more to videos it detects have a higher resolution or are more detailed?

    Is there such a thing as a youtube friendly file? One that will be natively accepted?
    Last edited by Knightmessenger; 13th Nov 2010 at 16:55.
    Quote Quote  
  14. Originally Posted by Knightmessenger View Post
    But I understand that my video might not be optimized for youtube compression, is there more I can do with encoding? Should I try to make the file size smaller or is that not as big of a factor? Does a good Mpeg 2 file turn out better than an avi? Was the old HQ setting the same as the 480p now or are they different?
    They lowered the bitrates on both audio & video recently for the SD versions , hence it's even more important now than ever to optimize your videos

    Basically, how can I optimize for youtube in ways that don't involve advanced video filtering or adjustment, just rendering and encoding?
    Not much you can do to improve unless you are willing to get more involved. The new version , Vegas 10 , does have a stabilizer, but it's not very good.

    You can get a fairly easy to use plugin called neat video for vegas and it works well, but its not free

    All the other optimization suggestions earlier were free methods, and I guarentee that you will get better results using them

    You can upload uncompressed and it doesn't change the fact that there is too much noise and movement that overloads the youtube compression.

    Does every video on youtube get allocated the same bitrate per frame? Or does it try to give more to videos it detects have a higher resolution or are more detailed?
    No. As mentioned earlier, the 720p version gets a higher average bitrate allocation. Even the 480p version encoded from the 720p version looks better than the 480p version uploaded as 480p


    Is there such a thing as a youtube friendly file? One that will be natively accepted?
    They all get re-encoded, hence the importance of optimizing your videos

    The main issue here is the low bitrate re-encoding practices of youtube. You have to optimize your videos to get slightly better results.

    All the optimization suggestions mentioned earlier add up - it might not be a huge improvement - but a few % improvement here, a few % improvement there and your video will look noticeably better
    Quote Quote  
  15. Originally Posted by Knightmessenger View Post
    Does every video on youtube get allocated the same bitrate per frame?
    No, each at the same resolution height (240p, 360p, 480p, 720p, 1080p) gets the same average bitrate. The bitrate per frame, as you put it, varies from very low in dark and static scenes to high in brightly lit scenes with more movement. You described CBR encoding. YouTube uses VBR encoding.
    Or does it try to give more to videos it detects have a higher resolution or are more detailed?
    It uses a fixed average bitrate depending on the resolution height at which it's been encoded. Each video gets reencoded several times by YouTube, up to the resolution of the source. Your 720x480 video has 240p, 360p, and 480p versions available. If you had uploaded a 1080p video, in addition to those 3 you'd also get 720p and 1080p versions made. It doesn't make any difference how 'detailed' or complex the video (and yours is very complex - the whole screen is in motion constantly, a real bitrate hog), each video only gets the same fixed average bitrate they use for that particular version. Which is why you want to make it as compressible as possible (within reason). So it'll turn out looking as good as possible.
    Quote Quote  
  16. Knightmessenger, I don't know if you already found a solution, but I agree with poisondeathray. While upscaling a 480i video to 720p won't improve on the source, you will get the advantage of getting higher bitrates on Youtube, so there's that.

    When I work with DV footage in Vegas I have made it a habit to correct the levels by using the preset, "Studio RGB to Computer RGB," but drag the "Input end" slider to 1.00. This way I boost the black levels, while preventing overexposure on the whites. For deinterlacing, I frameserve with Debugmode to Avisynth using QTGMC (if I haven't deinterlaced already) and then feed the script into Virtualdub. Save as AVI using FFV1 with PCM audio and then upload to Youtube. This can result in pretty big files and long upload times, however. The results are great though - at least they have been for me. The idea is to upload a lossless video and let Youtube do it's own encoding.

    QTGMC is one of the best deinterlacers out there, but it's slow. Deinterlace, and then upscale to 960x720p if your source is 4:3.

    This is the latest video I've uploaded using the method I talked about above. It's a mashup of two different sources: a music video and a commercial. One source was 480i and the other was 1080i, so I upscaled and downscaled them respectively to 720p.

    http://www.youtube.com/watch?v=490dZoEpCCw

    You'll see quality improvements on the 720p option over the 480p.

    The only problem I had was ghosting when upscaling the 480i source on some shots, but I doubt it's not noticeable, especially on Youtube.
    Quote Quote  
  17. Member Knightmessenger's Avatar
    Join Date
    Aug 2007
    Location
    United States
    Search Comp PM
    Okay so at the campus media lab, I did find someone who rendered a video file at 640x480 with a program called Compressor using the Youtube optimization setting. So I used them for this video that I entered into the ESPN Your highlight contest. The result is better than what I would have gotten using the settings I've done with most of my other videos. However, it still looks soft compared to some other videos.
    http://www.youtube.com/watch?v=HjQPp-IQEJs

    But like even with it optimized, it still doesn't look as sharp as this:
    http://www.youtube.com/watch?v=07c-DF3UArQ
    I know my Hi8 camera has lower resolution but it shouldn't look that more blurry, should it? The highest resolution setting is 480 (not HD) so it should have gotten the same bitrate from youtube as me. The fact that the sound is so terrible makes me think that the camera isn't even that high end. Do these tapeless cameras record in a way that it is easier to optimize them for a low bitrate or something?

    Even this from which says it was from an Android phone, appears to have better encoding.
    http://www.youtube.com/watch?v=ktnU9IbUf_E

    I know HD video has a much higher resolution than the Hi8 format is capable of but there's a lot of really tiny low quality cameras made these days. You know how a higher megapixel camera doesn't always mean a better picture. The Hi8 I have is a good camera, it has a manual focus and exposure setting that are both very easy to use. The lens and camera parts are much bigger than today's consumer models because the cameras are bigger. I did use the auto focus for my clip (which I now realize is vastly inferior to the manual setting) but it still had a pretty decent auto focus. Most new cameras don't even have those manual options. And they use hard drives instead of a physical tape medium. (most professionals prefer tape because that means there is a physical master that can't crash)
    But is even a SD video shot on a newer camera going to be that much better than my Hi8? I actually have a MiniDV (Canon ZR70) but it's terrible with no manual focus, no manual exposure (both auto modes are also worse than the Hi8) harder to hold steady, choppy digitally crunched sound and does much worse in lower light with grain unless it's broad daylight.
    Perhaps I should mention I've used a DV capture box (a crappy Pyro brand that drops frames like crazy) to convert the Hi8 video to digital. This means that the resulting video is compressed in the DV format. I have experimented with uncompressed captures but haven't gotten to work properly. However, I can tell they would be less grainy and more vivid with color.

    I guess what I'm getting at is that I might be interested in getting a higher resolution camera but it has to be significantly better in every aspect, just as easy to use and not cost over $700. I do not under any circumstances want another ZR70. So sorry to get off topic but yeah that's been something that's annoying me.

    This was another video entered into the contest that I couldn't understand how it looked so compression free. After all, whoever was holding the camera has quite a bit of room for improvement so I find it unlikely that person would be an encoding expert.
    http://www.youtube.com/watch?v=dCNApb6acos
    I would think he/she just used default settings and those turned out quite good. But then again, since it is an HD video youtube gives a higher bitrate than it would for one of my 480i videos. Correct?
    Quote Quote  
  18. One big difference between those other videos and yours, is lighting conditions. Low light is the horrible for video. They could have been improved significantly and optimized as well for youtube with those suggestions above. They weren't that impressive.

    However, it still looks soft compared to some other videos.
    A big reason is you're using vegas. If you use the default settings, you are blend deinterlacing. This means consecutive fields are blended together into a blurry mess. When you select "interpolate" it's not much better. It leave jaggies and you get a shimmering mess that impairs compression


    But then again, since it is an HD video youtube gives a higher bitrate than it would for one of my 480i videos. Correct?
    yes , but upscaling 480i in a poor fashion could actually make things worse. Of course starting with HD is better


    All you can do - if you keep your old equipment - is to optimize shooting conditions as best you can (eg. lighting, stabilization, slow pans) , and then optimize as much as you can in post production.



    I did some testing a while back, and that's where that list of optimizations above came from. Unfortunately those tests aren't on YT anymore, but trust me there was a drastic difference in quality. I only saved a representative screenshot, they were from the exact same 4:3 DV-AVI video but processed in 2 different ways. 1 was processed only in vegas ( like you did) , and the other used the optimizations listed earlier. The vegas upload was even 2x the filesize , but half as long in duration (ie. the 720p bitrate was way lower), but the vegas upload still looked like crap compared to the optimized version uploaded to youtube . Look at the details difference in the vest. Now youtube killed it too - all the grain is gone and everything looks "plasticky" and oversharpened but at least it's not a blurry, blocky mess like vegas' export.


    The moral of the story is quality begins at the source - shooting conditions, lighting, decent equipment, stabilization. This is the most important. But at least some improvement can be done in post, and these small optimizations add up and can improve the final quality on Youtube. This example below demonstrates gains between optimizing vs. not optimizing, because the source video was the exact same.


    Processed with Vegas only




    Using Optimizations

    Last edited by poisondeathray; 3rd Dec 2010 at 00:04.
    Quote Quote  
  19. With 16:9 1440x1080p footage, would it be advisable to apply a resize to 1920x1080 before uploading? From what I have seen Youtube resizes this type of footage to that size. I'm sure Avisynth scales better so would there be any noticeable quality increases if I resized prior to uploading?
    Last edited by seven_deuce offsuit; 14th Jan 2011 at 01:40.
    Quote Quote  



Similar Threads