YouTube distords my videos.

Why is it that sometimes, not always, youtube distords pixels for 1-2 second time at one point during the video. If i watch on my pc, the file is fine, if i upload anywhere else, like vimeo, its fine, but for some reason with some videos of mine youtube distords the image for a second or two, and it always does that only one time per video.

Also, if i watch in other resolution, like in 720 instead of 1080 it wont sometimes distord it, or if i use my friends pc the image is fine. Its so completely random and i have no idea wtf is causing it.

Anyone had any similar experiences, or has an idea what it could be?

I've been wondering this as well. I remember when cable tv went to digital early on I'd occasionally see a similar artifacting.

Hopefully this will get you started. Video encoding is kind of way too complicated to try and explain details in a couple paragraphs to address a vague description of some sort of video problem, so I'm kind of all over the place.

It's a subject far too lengthy to go into here, but it sounds like what you're talking about has to do with how video files are encoded. Raw lossless video files are, by definition, a series of full and complete images for every frame. When you encode a video, you're using various sets of algorithms to reduce the size through stripping of data that isn't necessary to make the video watchable to a human being.

The frequency of each type of frame is dependent on what settings you've configured your encoder to use. Generally, the more key frames you pack into a video, the larger the file is going to be. The further you space them apart, the smaller it'll be. The frames in-between these key frames perform various types of interpolation relating only to the data in the picture which is actually changing from one image to another (another keyword to look into here is "group of pictures").

Depending on the encoding (and Youtube suggests very specific optimal encoding workflow for the content you're going to upload), the space between these primary points can be up to hundreds of frames. This is part of the explanation for why you often can't just seek to absolutely any precise point in a video. If you were to take an encoded 30fps video that has 300 frame group of pictures setting and chop it just after one key frame, you could have just under ten seconds of useless data before the next key frame starts off a full usable image. Without seeing your video, this is what I'm guessing you're seeing.

Anyway, there's a ton of information to read into and one thing leads to another, if you really care about the topic of video and encoding (you'll end up learning about color spaces, the difference between RGB types and YUV, b-frames, and a ton of other crap). I'd suggest starting out looking at the manual for your codec of choice and then googling for information on all the options and phrases you encounter that you're not sure about. It also wouldn't hurt to go read up at sites like doom9.org .

http://en.wikipedia.org/wiki/Video_compression_picture_types

https://support.google.com/youtube/bin/answer.py?hl=en&answer=1722171&topic=1728573&ctx=topic

http://en.wikipedia.org/wiki/Macroblock (not exactly related to what you're talking about, but D_W was talking about that with regard to what he saw with streaming cable).