Talk:Codec

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Apple Lossless Encoding[edit]

Shouldn't [Apple Lossless Encoding] be under lossless, or am I missing something? 1189 13:56, Jun 24, 2004 (UTC)

Coder/Decoder[edit]

I added this wording, since it's also commonly used. It came back with 44,000 Google hits [1], one of which was a definition in the CNET glossary [2]. --Blackcats 15:59, 20 May 2005 (UTC)

I just looked through the article history and found out that a few months ago it used the coder meaning and then it was replaced with compressor. If anything, I think the coder meaning should actually come first, as it has more than twice as many Google hits [3]. It would be nice to find a source for the origin of the term, but just going by the sound of it, "codec" sounds a lot more like "coding" than "compressing." I would think that if the original meaning was compressor/decompressor, then it would be called "comdec" or something like that. --Blackcats 16:20, 20 May 2005 (UTC)
(en)coder/decoder is simply a more general concept. A codec doesn't necessarily compress data, it just converts data from one representation to another. For example, a YUV codec would convert data from RGB to YUV and/or vice-versa. That has nothing to with compression, it might just be a more appropriate representation respectively one that a software or hardware can handle. --82.141.48.65 02:44, 7 Jun 2005 (UTC)
Codec is an acronym for compressor/decompressor, *not* coder/decoder. The YUV "codec" is not a real codec, i.e. it should be named a "filter", because that's what it is. Unfortunately the wrong "coder/decoder" interpretation can be found on many web sites, which doesn't make it more correct, though. The copy-paste mentality makes errors spread world wide nowadays...
This being Wikipedia, we'll need an authoritative citation for that strongly-worded-yet-unlikely claim. The more encompassing term is "coder/decoder", of which compression is one of many possible purposes. As a counterexample, a base-64 codec actually makes the content *larger*. And no, a tool to transform between RGB and YUV would not be a "filter", as "filter" implies at least the potential to deliberately intercept and discard ("filter out") some portion of the content. 206.205.52.162 (talk) 23:45, 3 February 2012 (UTC)

Misleading Information[edit]

On the first Paragraph, "Compressor-Decompressor" or "Coder-Decoder" is basically not really the same thing, which need to be explained separately.

On second paragraph, It tells about Video/Audio Codec, which will be more appropiate to be on "Video Codec" or "Audio Codec" Article

The Third paragraph, I believe CoDec is not only about Encoding or Decoding.

The 4th is similar as the second.


^^ agreed --Elin

Conciseness at the cost of clarity[edit]

The explanation "This function is carried out by a video file format (or container), such as *.mpg, *.avi, *.mov, *.mp4, *.rm, *.ogg or *.tta. Some of these formats are limited to containing streams conforming to a small fixed set of codecs, while others are intended to be more general purpose." has just been replaced by a simple link to container formats. I don't think that was a good idea.

I don't really understand all this but reducing everything down to the bare minimum makes it far harder for people like me to understand. To understand what a codec is you need to know what a container format is and to understand what a container format is you need to know what a codec is. Trying to use wiki to get my head round this is a bit like reading an account of the battle of Gettysburg which has been split into Confederate actions at the Battle of Gettysburg and Union actions at the Battle of Gettysburg. Repeating stuff that appears on another page is not the worst thing in the world.Dejvid 19:46, 28 September 2005 (UTC)

IMHO this kind of information should go into Audio file format. --Hhielscher 20:15, 28 September 2005 (UTC)

Needlessly complicated?[edit]

Is it just me or does this article seem needlessly complex?

I don't think the first paragraph very clearly explains what a codec is. It should separate the two and aim to explain the basic function of a codec, maybe something along these lines:

In Telecommunications, a codec is a portmanteau of COder-DECoder, and is a device or program which deals with the transformation of data signals from analogue to digital, and from digital to analogue.

It also stands for COmpressor/DECompressor, and in this case it is a device or program which compresses data for storage. The same codec then takes data in a compressed state and decompresses it so that it can be run.

Or maybe a device or program which deals with the transformation of data files from a compressed to decompressed state and vice versa. I dunno.

At least that's how I understand it. The article does not make it clear that the two kinds of codec are used for different things, and this is perhaps one of the most complicated definitions of codec I've seen, and I'm not saying that I'm great at explaining things- but I'm sure someone could do this better.

--Elin

I always use codec to refer to something which compresses and decompresses digital data - for transmission or storage. Stephen B Streater 18:54, 29 June 2006 (UTC)

YES I absolutely agree that the discussion is needlessly complicated[edit]

Exactly how is the codec used to compress and decompress data. It is clearly not a stand alone computer program. It might be a device? Is is a software library with an interface? Is the interface standard? This topic concerning codec is indeed needlessly complicated and obtuse. —Preceding unsigned comment added by 130.76.96.153 (talk) 16:28, 24 May 2011 (UTC)


NPOV[edit]

See fourth paragraph. Someoneinmyheadbutit'snotme 02:49, 10 July 2006 (UTC)

I'm removing the POV tag, since there is no real controversy visible, either here on the talk page, or the article page. --Frescard 04:18, 29 August 2006 (UTC)

Non-sequitur[edit]

The Metal Gear Solid reference might belong on a disambiguation page, but not here.

Codec Hell[edit]

I think the article should mention the problem that many users face, of having to download hundreds of codecs to just try to watch a video.. And sometimes, if the codec is obscure or hard to get, it's not only impossible to watch the file, but you also end up with many installed codec packs that might conflict, sometimes messing up your whole setup so that even playback of previously known files is altered. (And the problem may be hard to track down.. Not all codecs or codec packs have a straightforward uninstall process) --164.77.84.202 16:20, 21 October 2006 (UTC)

There needs to be something about the existing tools for identifying what format a particular media file is in, (or rather what format its contained data is in), and therefore which codec or type of codec needed to decode it. --AC 05:57, 20 October 2007 (UTC)

Fake codecs[edit]

The article should also mention fake codecs, which are used to install spyware/adware.

Definions and consistency[edit]

The article today reads "A codec is a device or program capable of performing encoding and decoding on a digital data stream or signal."

Questions:

  1. This uses the word "and", implying a codec must do both. Yet surely most of the 'codecs' that one downloads, or finds in a device, will only decode, there being no need for encoding. Is the definition wrong, or is it incorrect to call these codecs?
  2. Having established that a codec is a device or software, it then says "There are lossless codecs". Surely, there are lossless encoding/decoding methods, but that is not a codec according to the definition in the article. There will also be lossless/lossy encoding implementations, potentially slightly different.
  3. Similarly, "While many people explain that AVI is a codec" (given that it says this is incorrect but not from a definition point of view) seems to imply that a codec can also be a file format rather than a piece of software.
  4. I observe the quick summary says "This article is on the compression/decompression algorithms...". Is it about algorithms, encoding specifications, file formats, or implementations? All of these are quite distinct it seems to me.

Can anyone clarify? Notinasnaid 09:14, 12 April 2007 (UTC)

1) I think you are taking the words on this article too literally. A codec you download may just be the decoder but mpeg4, divx, etc. are codecs. The both encode (compress) and decode (decompress) data. mplayer has all of these and can do both for, as far as I know, all of them (except maybe real). Cburnett 13:07, 12 April 2007 (UTC)
Well, yes, I am taking the words on the article literally. If they are not literally correct, what are they doing there? Maybe we can fix them if they are wrong, but they certainly shouldn't be left alone if they are wrong. It seems to me... (I should clarify: I am not asking these questions out of idle curiosity but because the article seems to contradict itself, something that I think needs fixing). Notinasnaid 13:17, 12 April 2007 (UTC)
I agree. How can you take the article "too-literally"? Either it is correct or it is incorrect. As it is now, a codec that only encodes isn't a codec, and this is incorrect- for one of many examples, check out the one offered here: http://mp3decoders.mp3-tech.org/decoders_acm.html As the articles definition is incorrect, and inconsistant with the usage of the word, it should be changed. That the article encompases most codecs isn't relevant. If the defintion of a shirt claimed it must have a collar and sleeves, would we leave it despite the existance of sleeveless-shirts? When Cburnett cites the overwhelming majority of shirt have sleeves would we be convinced? No, cuz a shirt doesn't need sleeves to be a shirt, and a codec doesn't need to be able to encode to be a codec.65.185.93.86 (talk) 12:41, 19 March 2008 (UTC)
2) Why does the lossless attribute of a codec make it not a codec? I don't think you understand either lossless or codec very well (no offense intended yet you seem to insist you do know what it means). Lossless means that there is no information lost during the encoding. FLAC is lossless. JPEG can be lossless with a particular quantization matrix. Both FLAC and JPEG are definitely codecs. Cburnett 13:07, 12 April 2007 (UTC)
I may have phrased my question badly. I am quite happy that there are lossless and lossy methods and software. If JPEG is a codec, then presumably the definition in the article is wrong, since JPEG is an encoding technique, not a piece of software or device. Notinasnaid 13:17, 12 April 2007 (UTC)
If JPEG is an encoding technique only then how do you expect to view the JPEG? That's write: you decode it. You decode it by doing things in the opposite order of encoding yielding an image approximately/exactly the same as the original depending on if you use a lossy/lossless quantization matrix. Cburnett 13:40, 12 April 2007 (UTC)
Sorry, I'm still not explaining myself. Yes, JPEG has a specification which defines or implies encoding and decoding methods. But clearly JPEG does not fit the definition "A codec is a device or program capable of performing encoding and decoding on a digital data stream or signal." How about if the sentence was A codec is a device or program capable of performing encoding and/or decoding on a digital data stream or signal, or the specification or algorithm that describes how that that encoding or decoding is to be done? Notinasnaid 13:48, 12 April 2007 (UTC)
3) A file format is not a codec. Cburnett 13:07, 12 April 2007 (UTC)
Good. So should the paragraph starting "While many people explain that AVI is a codec, they are incorrect" be removed, since it does seem to suggest that some file formats (which are not containers) - yes? Notinasnaid 13:17, 12 April 2007 (UTC)
Not removed, changed. Many people assume that AVI is a codec (I can hear it now: "it's a file suffix like .jpg so isn't a codec?") and there's nothing wrong with exaplining that AVI, RIFF, ogg, etc. are containers not codecs. Though, if you really really wanted to, you can argue that a container format is a codec because it does encode data into a specific form an decode it back into its original form. More of a trivial codec in comparison with "real" codecs. This is more of an academic argument though. Cburnett 13:40, 12 April 2007 (UTC)
4) They are all different perspectives on the same thing. A specification outlines how a codec should perform. An algorithm is a specific way to perform a codec. An implementation is a specific written way of doing an algorithm. You have to store the encoded data somehow and thus a file format is needed. Cburnett 13:07, 12 April 2007 (UTC)
So, how should the definition be rewritten so that it is correct, since it suggests it only applies to implementations? Notinasnaid 13:17, 12 April 2007 (UTC)
Personally, I'm fine with it. However, I think it could be partially rewritten and broken up into sections. (Most articles could use this anyway.) Cburnett 13:40, 12 April 2007 (UTC)
Sorry, I think pedantry has its place and an encylopedia is one of those places. The definition should be accurate, so I can indeed take the words in the article literally. I don't think we should aim any lower than that. I have no view (yet) on whether the article should be broken up, though a proper lead section which summarised/repeated what followed would be good. Notinasnaid 13:48, 12 April 2007 (UTC)
At the end of the day, JPEG is a codec and JFIF is a file format commonly used to contain JPEG encoded data. Cburnett 13:07, 12 April 2007 (UTC)

Lower bit rate codecs[edit]

I agree that Lower bit rate codecs should be merged with Codec. Biscuittin 17:15, 13 October 2007 (UTC)

Capitalization[edit]

Since when did the the abbreviation CODEC become an all lowercase word? —Preceding unsigned comment added by Andreas Toth (talkcontribs) 02:04, 22 February 2008 (UTC)

I've seen it exclusively that way since I first saw it, over a decade ago. Iain Dalton (talk) 06:50, 18 December 2009 (UTC)

How is a Codec chosen?[edit]

In Windows Media Player, how is a codec chosen for playing a selected video file? David (talk) 16:27, 16 March 2008 (UTC)

Using the DirectShow merit and media types system. In short, DirectShow tries every possible combination of splitters, decoders, renderers and converters on the file, in order of merit. Obviously there's a method to this madness. The configuration that works is chosen, behind the scenes. If no working configuration exists, WMP will ask you to download a codec from the web. C xong (talk) 07:57, 26 May 2010 (UTC)

merge endec into codec[edit]

I suggest that the "endec" article should be merged with the "codec" article, since the current version of the endec leaves me confused about the distinction (if any) between an "endec" and a "codec". Even if there is some subtle distinction, perhaps they are similar enough that a single article describing both of them (and also the subtle distinction between them) would be better than 2 separate articles -- similar to the way one article covers both Myanmar and Burma. --68.0.124.33 (talk) 14:50, 29 May 2008 (UTC)

More clarity[edit]

The Related Concepts section (a better title might be "Introduction") bypasses the main point: WHY codecs exist. Presumably compression is to allow more data (higher-quality sound or video) to flow over limited bandwidth and some encoding is for copyright protection. Perhaps someone with greater technical knowledge than I can make the missing connection.

The fourth paragraph of this section is one long, ungrammatical, muddled sentence: "The raw encoded form of audio and video data is often called essence, to distinguish it from the metadata information that together make up the information content of the stream and any "wrapper" data that is then added to aid access to or improve the robustness of the stream." I've tried to clarify it as follows, but hope others can continue to improve it: "The raw encoded form of audio and video data is often called essence, to distinguish it from metadata information. The essence and metadata together make up the information content of the stream. "Wrapper" data may be added to aid access to or improve the robustness of the stream."

"Metadata" in this paragraph also requires an explanation or link to an article that explains it (the link provided is not useful).

Billfalls (talk) 15:02, 7 May 2009 (UTC)

The 4th para on "essence" is definitely an WP:OR (no books talk about it), I will remove it, and will add the WHY part when time permits... Thanks for raising these. --Nvineeth (talk) 16:52, 7 May 2009 (UTC)

Digital voice coders[edit]

I would like to add an entry in Wikipedia to discuss Voice Codecs (or if you prefer Voice Coders). These are the devices that convert your voice to a digital representation (at fairly low data rates) in all Cellular telephone systems, and most VOIP implementations. SDG —Preceding unsigned comment added by Sdgardner (talkcontribs) 13:12, 6 June 2009

Sounds like it's worth a mention. It's a way to get low datarates, and if phonemes are used, the codewords can be searched. Stephen B Streater (talk) 11:04, 21 March 2010 (UTC)

Codecs and Formats[edit]

There is a lot of confusion within the general public (and even in a number of wikipedia articles) about the difference between codecs and formats, and it would be nice to mention this in this article. C xong (talk) 07:58, 26 May 2010 (UTC)

Do you mean container formats like Ogg? It is probably worth a sentence with a wikilink to an article about these. Stephen B Streater (talk) 08:34, 26 May 2010 (UTC)
Not only container formats, but formats in general. Audio and video compression formats. The vast majority of people believe that a codec is a format. But as you can see from the definition of the word "codec", it is "a device or computer program" that encodes/decodes data. Formats do not encode and decode data. Formats are not devices or computer programs. Formats are specifications, describing how to make devices or programs that encode/decode data. How to make codecs. They are coding standards. Now, people are extremely confused about the difference between a specification and implementation—between a format and a program (and not only in the multimedia area). Most people call audio/video formats "codecs" (for example, "audio encoded in the MP3 codec", "DVD video uses the MPEG-2 codec", etc.), and most people call audio/video codecs "formats" (for example, "I downloaded a Xvid video", "I encoded the file into the x264 format" etc.) They also believe that multimedia containers can "contain audio and video codecs" (how can you put a codec into a container, or why would you want to do it?), that's why they "compare DivX to H.264", apples to oranges, that's why all those "codec lists" freely mix codecs with formats, that's why the HTML5 video discussion is full of controversy about "the H.264 codec" (is there only one?) and so on and so on. The word "codec" is really a misnomer in the vast majority of cases it is used today, thanks to the proliferation in texts written by laymen.—J. M. (talk) 20:44, 26 May 2010 (UTC)
That's enough people to write a section on it then. A couple of external references to codec and format definitions (and ideally a reference to the confusion itself) would be ideal. Stephen B Streater (talk) 21:49, 26 May 2010 (UTC)

Merge Proposal[edit]

Uh, how exactly is there a tag on there without a relevant discussion section?

Instead of being merged, this aritcle should be expanded to talk about different types of audio codecs and their advantages/disadvantges (rather then the incprehensible tree that is sued in the articles that are incorrectly entitled "comparison"), and discuss what audio codecs do specifically that other codecs do not, and/or what is nessecary for an audio codec to have.                     ~Rayvn  13:40, 6 July 2010 (UTC) —Preceding unsigned comment added by RayvnEQ (talkcontribs)

Needs reference to MXF container[edit]

MXF is missing from the list of container formats. —Preceding unsigned comment added by 71.231.50.137 (talk) 14:23, 27 September 2010 (UTC)

Rename proposal[edit]

Codec is not an unambiguous term (see Codec (disambiguation)) and the use of the term described here is not obviously dominant in my experience. I propose that this article be renamed Media codec. This would allow the disambiguation page be renamed Codec and would give readers immediate context as to what they're reading about. --Kvng (talk) 14:06, 24 March 2011 (UTC)

I disagree. Codec is a fairly unambiguous term. Yes, there can be audio codecs or video codecs. This article talks about codecs in general. Audio codec or video codec is a subset of the term codec, and they have their own separate articles. And which use is dominant in your experience? I think the use explained in this article is exactly the most dominant (and correct) use of the term. And how exactly would renaming the article to "Media codec" help? This article is not only about media codecs. This article explains what a codec is. That's why it is named "Codec". Hence, I can't see a more suitable name for the article. "Media codecs" is only a subsection of this article.—J. M. (talk) 14:22, 24 March 2011 (UTC)
Audio codec is, in fact, an fully ambiguous term. I don't believe there is a clearly dominant use of the term, that's why I have proposed the rename; It brings the reader straight to the disambiguation page when they search for "Codec". --Kvng (talk) 19:57, 24 March 2011 (UTC)
This seems stale, but in any case, I disagree with any rename. There doesn't seem to be much confusion... as evidenced by the lack of confusing entries and possible targets on the disambiguation page. Shadowjams (talk) 03:22, 10 February 2012 (UTC)

Readability: Approach by reader type[edit]

I think most visitors to this article can be classified as users (viewers and posters) or software developers.
(As a viewer, I presume that viewer users predominate.) With that in mind, I recommend that a separate, titled section or two aimed at these users, describe how to find an appropriate codec or format (most viewer users won't know or care about anything but the filename extension or the program complaint that the 'codec could not be found' or the 'format is not supported'.
Posting users will be interested in producing their file so their recipients will have least difficulty viewing the produced file(s).
Certainly the most common (popular) programs should also be addressed: Browsers, plug-ins, and viewer programs.
Perhaps a comparison table would be a convenient way to start, with XROCs as appropriate to more descriptive articles.
Separating the information of this article this way will make it more useful to both/all three types of visitor.
Note: This problem of reader types is fundamental to the nature of Wiki, and pervasive, but typically ignored by savants. I have not seen it addressed in the Wiki-contributor guidance.
--Wikidity (talk) 01:11, 27 November 2012 (UTC)

Ubuntu Link[edit]

The Ubuntu link in the references is broken. — Preceding unsigned comment added by 195.70.5.235 (talk) 09:03, 31 October 2013 (UTC)