Looking for a primer on digital video.
November 14, 2010 3:44 PM Subscribe
Primer on digital video codecs, containers, etc. While working with video on my Linux PCs, I am often confused because I just don't know much about all the different video formats (MPEG 4, h.264) "containers" (OGG, AVI) the different kinds of compression, how the audio and video stay in sync, what various terms like "muxing" mean, and so forth. I'm looking for a good document that explains all these things to someone with little knowledge of it. A web document, book, Wikipedia article, etc. would be good. I just don't know where to start to see how these things fit together. Thanks.
One site which is often useful for those who work with video files is http://www.doom9.org/
posted by dhens at 4:22 PM on November 14, 2010
posted by dhens at 4:22 PM on November 14, 2010
Forgive the self-link, but I wrote a short overview of this topic for my students that you might find useful. It leads into a brief article on video production workflow that may also help answer your questions.
posted by Bora Horza Gobuchul at 5:11 PM on November 14, 2010
posted by Bora Horza Gobuchul at 5:11 PM on November 14, 2010
Wikipedia has a lot of good information, for example here's a comparison of container formats and what types of video/audio each can hold.
Muxing is short for multiplexing, which means essentially, "N things in, 1 thing out." The opposite is demultiplexing (or demuxing for short) which means, "1 thing in, N things out." In the context of multimedia files, muxing means combining individual elemental streams into a single file, for example an audio stream and a video stream are combined into a single "video clip". It makes a little more sense when you consider more complicated cases, such as a file with multiple audio streams like a DVD with commentary track, which might also have a subtitle track as well -- all of those streams are multiplexed into one VOB file. Demuxing refers to the process of splitting such a file back into its parts.
posted by Rhomboid at 5:12 PM on November 14, 2010
Muxing is short for multiplexing, which means essentially, "N things in, 1 thing out." The opposite is demultiplexing (or demuxing for short) which means, "1 thing in, N things out." In the context of multimedia files, muxing means combining individual elemental streams into a single file, for example an audio stream and a video stream are combined into a single "video clip". It makes a little more sense when you consider more complicated cases, such as a file with multiple audio streams like a DVD with commentary track, which might also have a subtitle track as well -- all of those streams are multiplexed into one VOB file. Demuxing refers to the process of splitting such a file back into its parts.
posted by Rhomboid at 5:12 PM on November 14, 2010
Also: Know the difference between the specification and an implementation. There is a huge difference, because a poor implementation of an advanced standard can be beaten in quality by a good implementation of a less-sophisticated standard. Additionally, standards names can be confusing: MPEG-4 is a standard that has 27 parts, each describing completely different and unrelated things.
For example, MPEG4 Part 2 describes a video format with two profiles: Simple Profile and Advanced Simple Profile (ASP). Common implementations of the ASP are XviD and DivX. So "MPEG4 Part 2 ASP" and "XviD" are referring to the same thing, but one is the standard and one is a particular implementation of that standard. Often people refer to the format of a video clip by the name of the implementation used to encode it, so you might see someone talk about an "XviD format file", even though any MPEG-4 Part 2 capable decoder can decode it.
Similarly, MPEG-4 Part 10 describes a video format which is also known as H.264. A common (and some would say the best) implementation of an encoder for this format is x264, and sometimes you see people talking about clips in x264 format. This is done because identifying the particular encoder tells you more about the quality than if you referred to it by the standards-name, because again, the quality of the implementation of the encoder is just as important if not more important than the standard.
posted by Rhomboid at 5:35 PM on November 14, 2010
For example, MPEG4 Part 2 describes a video format with two profiles: Simple Profile and Advanced Simple Profile (ASP). Common implementations of the ASP are XviD and DivX. So "MPEG4 Part 2 ASP" and "XviD" are referring to the same thing, but one is the standard and one is a particular implementation of that standard. Often people refer to the format of a video clip by the name of the implementation used to encode it, so you might see someone talk about an "XviD format file", even though any MPEG-4 Part 2 capable decoder can decode it.
Similarly, MPEG-4 Part 10 describes a video format which is also known as H.264. A common (and some would say the best) implementation of an encoder for this format is x264, and sometimes you see people talking about clips in x264 format. This is done because identifying the particular encoder tells you more about the quality than if you referred to it by the standards-name, because again, the quality of the implementation of the encoder is just as important if not more important than the standard.
posted by Rhomboid at 5:35 PM on November 14, 2010
If you're using Linux, do head over to the cinelerra group on vimeo and ask there...lots of friendly people making good-quality videos and giving good advice on the things you're asking about.
posted by circular at 7:44 PM on November 14, 2010
posted by circular at 7:44 PM on November 14, 2010
A Digital Media Primer for Geeks is a great introductory video presented by one of the main ogg vorbis authors. Only one episode so far, so it doesn't go very deep but it still makes for an excellent base to work from.
posted by pharm at 12:29 AM on November 15, 2010
posted by pharm at 12:29 AM on November 15, 2010
This thread is closed to new comments.
posted by mmascolino at 3:58 PM on November 14, 2010