tool to add metadata fields to MPEG2

[jagabo]

I've only seen it work for MP3 files. It wraps the MP3 file in an ID3 container.

http://en.wikipedia.org/wiki/ID3

It doesn't work for MPG files.

Why doesn't VLC do the same job on MPEG2's? The Wikipedia-link you sent indicates it should be possible:

"ID3 tags were designed with MP3 in mind, so they would work without problems with MP3 and MP3Pro files. However, the tagsets are an independent part of the MP3 file and should be usable elsewhere. In practice, the only other format which widely uses ID3v2 tags is AIFF, where the tag is stored inside an IFF chunk named "ID3". The same could be accomplished in WAV, but isn't. "

If you wrap the MPG file in an ID3 container players won't know what to do with it (they're not expecting to see an MPG file inside). If you write metadata info into the MPG file's USERDATA field you have to define your own metadata standard and convince all the other tool makers to follow it. I suspect most of them would just tell you to wrap your MPEG data into an MKV container and use MKV's metadata tags.

[HitTheRoad]

If you wrap the MPG file in an ID3 container players won't know what to do with it (they're not expecting to see an MPG file inside)

Are you sure of this? Do players really expect to find a MP3 inside a ID3 container? MP3's aren't the only files that have a ID3 container wrapped around them. AIFF's also have a ID3 container. I have never heard of a player that can't play AIFF's, because the player doesn't expect to see an AIFF inside.

If you write metadata info into the MPG file's USERDATA field you have to define your own metadata standard and convince all the other tool makers to follow it.

Can you please explain this to me? I find it hard to understand what you're trying to tell me here. Suppose it's possible to wrap a ID3 container around MPEG2's and suppose players know how to play them, then why should you want to write metadata into the MPEG file's USERDATA field? In that case you wouldn't need the USERDATA field. Right? In that case you would write the metadata into the ID container. Or not?

[jagabo]

It was an either/or situation. If you can't wrap your MPG file in an ID3 container your other option is to use the USERDATA field in the MPG file.

Regarding whether wrapping MPG in ID3 will work in existing players, I have no evidence either way. But as a software engineer I can tell you this: nothing in handled in software unless it has been specifically programmed for. Since nobody has yet (as far as I know) wrapped an MPG in ID3 it's doubtful any player will handle it correctly. You're the one who's interested in this, you will have to do your own research.

[Cornucopia]

If you wrap the MPG file in an ID3 container players won't know what to do with it (they're not expecting to see an MPG file inside)

Are you sure of this? Do players really expect to find a MP3 inside a ID3 container? MP3's aren't the only files that have a ID3 container wrapped around them. AIFF's also have a ID3 container. I have never heard of a player that can't play AIFF's, because the player doesn't expect to see an AIFF inside.

If you write metadata info into the MPG file's USERDATA field you have to define your own metadata standard and convince all the other tool makers to follow it.

Can you please explain this to me? I find it hard to understand what you're trying to tell me here. Suppose it's possible to wrap a ID3 container around MPEG2's and suppose players know how to play them, then why should you want to write metadata into the MPEG file's USERDATA field? In that case you wouldn't need the USERDATA field. Right? In that case you would write the metadata into the ID container. Or not?

You're completely misunderstanding these file formats. There is no such thing as an "ID3 Container" file format. There is an "ID3 container segment" embedded WITHIN an MP3 file, and an ID3 container chunk appended to a certain chunk branch in an AIFF. There "could" be one for WAV, but there is no need, since there already is a WELL-USED and WELL-DEFINED professional standard for tagging of WAV files - it's known as the BWF (Broadcast Wave Format) format.
MP3 lends itself well to adding an ID3 tag segment container because the structure of the MP3 file has "dummy" areas or "gaps" if you will, where it doesn't hurt the file (or the understanding/playing of the file) to have it added in. And ID3 was commonly appended to MP3's early on in their popularity, so it became much more easily associated with MP3s (though certainly not necessary).
MPEG video files are another matter.

They don't use an elementary stream format like MP3, they use a multiplexed, System/Program/Transport stream format. This both makes it easier to include ID3 (into a separate muxed stream), and complicates things (because the positioning could be a number of possible places and a player would have to be smart enough to check around to find the placing).

Also,
Your whole argument about metadata within the file is only good from a certain ponit of view. The industry as a whole has already decided to go the other route. Not counting external databases like you've already mentioned, MXF and MPEG7 are growing ways of implementing standardized, abstracted, wholistic metadata in a format that is combination super-container and database.

Scott

[usually_quiet]

MP3 only supports audio. MP4 also allows embedding of ID3 data. (iTunes uses it.) However, using MP4 may require you to convert audio at times, because MP4 doesn't support all the same kinds of audio that DVD allows. Plus you will need to find a different player that can use ID3 tags in a MP4 container and MPEG-2 video, or convert your video to .MOV to use iTunes.

ID3 is a formal standard. It only allows kinds certain kinds of information to be stored, and each kind of information is associated with a unique identifier. From a computer programming perspective, that makes it much easier to work with than MKV metadata.

MKV isn't used for commercial products like MP3 or MP4. MKV metadata doesn't have the same kind formal standard controlling the type of information that can be stored and provides no pre-defined identifiers for the information. My best guess is that MKV metadata is mainly used as a convenient way for people who trade video files to provide information about the video they are uploading. Figuring out what is there requires a human to read the output from MediaInfo or similar.

[HitTheRoad]

Sorry if I misunderstand you guys. Compared to you I'm Alice in Wonderland. But I'm trying my best learning.
I have a better understanding now of the difficulties and possibilities of ID3 and MPEG2's. I understand now it's possible to build an editor that includes ID3 in MPEG2's, but that it might be impossible to build a player (or modify existing players) that can read ID3 in MPEG's:

MPEG video files are another matter. They don't use an elementary stream format like MP3, they use a multiplexed, System/Program/Transport stream format. This both makes it easier to include ID3 (into a separate muxed stream), and complicates things (because the positioning could be a number of possible places and a player would have to be smart enough to check around to find the placing).

So, will there ever be a media player that can both edit and read ID3 tagged MPEG2's? It's unlikely, I now understand, but still, if nobody ever tries, we're sure it'll never happen.

@ Cornucopia:

What about the option Jagabo mentioned, about using the USERDATA field in MPEG2 files? Suppose I would try to build my own media player and metadata editor. What would be harder? Building one that tags and reads meta data using the USERDATA field? Or building one that tags and reads metadata using something like ID3 container segments?

The industry as a whole has already decided to go the other route. Not counting external databases like you've already mentioned, MXF and MPEG7 are growing ways of implementing standardized, abstracted, wholistic metadata in a format that is combination super-container and database.

What do you prefer? Metadata exclusively written into the files? Or metadata in a format that is a combination of super-container and database? And why is 'industry as a whole going the other route'? And do we really have to care? There's GStreamer. I'm sure you guys are familiar with that. I just stumbled upon it. This is what the GStreamer handbook says about tagging and ID3:
http://gstreamer.freedesktop.org/data/doc/gstreamer/head/pwg/html/chapter-advanced-tagging.html

[Cornucopia]

"Will there ever be a media player...?" Probably not, unless you or a friend of yours, or someone you pay/commission to write your own. Most players go with standards, for longevity & compatibility-sake. When there is a standard way of doing things (including tags & metadata), the ROI on a programmer's time investment is much better. Plus, the likelihood of it being used is greater.
ID3 segments in USERDATA fields of MPEG2 aren't a BAD idea (prior to jagabo, I myself mentioned it in one of my earlier posts on this thread), it's just not one that's really going to get off the ground. MOST PEOPLE don't keep track of "Artist", "Album" and "Genre" when describing/catalogging Video files, so your interest/need is not standard, in this respect. Plus, a lot of other things are currently using the USERDATA area (closed-captioning, broadcasting info, etc)
For portability, having metadata encapsulated within the file is fine, but its self-contained-ness is also a weakness when you talk about libraries. Makes it more difficult to support with more and more formats (as the metadata is all in different places with different file formats), and slows down read-writes.
Plus, if the industry has already created something like MPEG7 (which is expansive and extremely capable of ANY metadata you might think to include), and it supports just about ALL the various MPEG formats (MPEG1, 2, 4, h.264, h.265, etc), why reinvent the wheel? Just find or make an app that takes advantage of that.

It's already been suggested to you that you either:
1. Build your own
2. Use a different tagging system, like a database (this would also include MXF or MPG7)
3. Remux into containers that already support the tagging system you like

Take your pick. I already know which one I would choose (#2).

Scott

[usually_quiet]

iTunes only needs to support the types of media that Apple sells, and Apple has a lot of influence over the standards applied to the containers used for its media products. That's why everything can be very simple and straightforward in the iTunes ecosystem.

Third-party media centers need to use external databases. There are too many different types of media files that people want to use with them. It is the only sane way to store and organize data for so many types of files and types of media, most of which don't have a way to store metadata themselves.

I thought you might be interested in reading this http://forum.xbmc.org/showthread.php?tid=12721 Based on what it says, it appears XBMC can make use of the ID3 data in music files. (XBMC provides a way to add ID3 information to its internal database) Although it won't help you, there are apparently a few XBMC users who want to use the ID3 tags present in their iTunes music videos too, and there is a way to do that. Too bad MP4 and DVD are not a better match.

[HitTheRoad]

Plus, if the industry has already created something like MPEG7 (which is expansive and extremely capable of ANY metadata you might think to include), and it supports just about ALL the various MPEG formats (MPEG1, 2, 4, h.264, h.265, etc), why reinvent the wheel? Just find or make an app that takes advantage of that.

I don't no anything about MPEG7. I'll have to dig into that subject. Don't you regard it as a weakness that MPEG7 partly depends on an external database?

I thought you might be interested in reading this http://forum.xbmc.org/showthread.php?tid=12721

Interesting indeed. All people trying to find a solution to make their vast growing AVI and MPEG collection accessible trough a tool that writes and reads meta data.

MOST PEOPLE don't keep track of "Artist", "Album" and "Genre" when describing/catalogging Video files, so your interest/need is not standard, in this respect.

This may be true. But MOST PEOPLE with huge video collections want to keep track in a way that suits them best. For example, collectors of TV Shows may want to categorize their files using atoms like Show Name, Episode Number, Season Number, Director, Genre, and so on.

I never did any programming but I would like to try to build me a tool that suits me best:

* I already have a player that plays MPEG2: VLC.

* I need an interface/organiser. Maybe vHelp can help me with this? vHelp, are you still there?

* The bottleneck will be to build a tag tool, that can write tags into the USERDATA field of MPEG2 (using MPEG7 or not) and that can read them as well. Of course the tag tool has to be integrated into the interface.

What skills and what software do I need to build an editor like this? Can I do this with GStreamer?

[usually_quiet]

Plus, if the industry has already created something like MPEG7 (which is expansive and extremely capable of ANY metadata you might think to include), and it supports just about ALL the various MPEG formats (MPEG1, 2, 4, h.264, h.265, etc), why reinvent the wheel? Just find or make an app that takes advantage of that.

I don't no anything about MPEG7. I'll have to dig into that subject. Don't you regard it as a weakness that MPEG7 partly depends on an external database?

I thought you might be interested in reading this http://forum.xbmc.org/showthread.php?tid=12721

Interesting indeed. All people trying to find a solution to make their vast growing AVI and MPEG collection accessible trough a tool that writes and reads meta data.

MOST PEOPLE don't keep track of "Artist", "Album" and "Genre" when describing/catalogging Video files, so your interest/need is not standard, in this respect.

This may be true. But MOST PEOPLE with huge video collections want to keep track in a way that suits them best. For example, collectors of TV Shows may want to categorize their files using atoms like Show Name, Episode Number, Season Number, Director, Genre, and so on.

I never did any programming but I would like to try to build me a tool that suits me best:

* I already have a player that plays MPEG2: VLC.

* I need an interface/organiser. Maybe vHelp can help me with this? vHelp, are you still there?

* The bottleneck will be to build a tag tool, that can write tags into the USERDATA field of MPEG2 (using MPEG7 or not) and that can read them as well. Of course the tag tool has to be integrated into the interface.

What skills and what software do I need to build an editor like this? Can I do this with GStreamer?

While others are interested in using metadata embeddied in files, the one XBMC developer who showed any interest in the project basically said the only video standard he would consider making a tag reader for was MP4 with iTunes atoms or ID3 data. The reason being that no formal standards exist for any other kind of metadata embedded in media files commonly used by consumers. This confirms everything that Cornucopia already said.

In other words, you can make up your own standards for MPEG-2 metadata, but no programmer in his right mind will write a program to use them because that is a lot of effort to put into something that is not already used by large numbers of people, and that is not controlled by any published standard. MPEG-7 is more practical, because it is controlled by a standard, but it isn't popular with consumers yet, so there is no demand for a consumer product that uses it.

As has also already been said, the other problem is that using the GOP user data is an imperfect solution. If you use the GOP user area for your tags, that will erase any N. American style closed captions already there. (VLC and some other players can display them.) There are people who use those closed captions to be able to see the lyrics and dialog for performances they recorded from TV. ...and some NTSC DVDs use them as well. To use your idea, they would have to convert closed captions to subtitles before tagging their files.

Since you are the only one who will use your tagging method, it probably will be up to you to implement it. I have no experience with GStreamer, so I can't answer your question about whether it can be used to do what you want.

[jagabo]

MPEG-7 is more practical, because it is controlled by a standard, but it isn't popular with consumers yet, so there is no demand for a consumer product that uses it.

And it is likely covered by many patents.

[HitTheRoad]

While others are interested in using metadata embeddied in files, the one XBMC developer who showed any interest in the project basically said the only video standard he would consider making a tag reader for was MP4 with iTunes atoms or ID3 data. The reason being that no formal standards exist for any other kind of metadata embedded in media files commonly used by consumers.

MPEG2 files ARE commonly uses by consumers (The ones who rip DVD's and don't want to do conversions, because they want the real thing (DVD quality). There may be no formal standards to embed metadata in MPEG2 files, but this doesn't mean there couldn't be such a formal standard and it doesn't mean there's no demand for this.

MPEG-7 is more practical, because it is controlled by a standard, but it isn't popular with consumers yet, so there is no demand for a consumer product that uses it.

Are you sure there isn't a demand for MPEG-7? It's new and few have ever heard of it.

As has also already been said, the other problem is that using the GOP user data is an imperfect solution. If you use the GOP user area for your tags, that will erase any N. American style closed captions already there. (VLC and some other players can display them.)

Are there GOP user data inside MPEG2's that have been ripped from DVD? In the MPEG2's I opened in VLC, I couldn't find any data, except the file title.

In other words, you can make up your own standards for MPEG-2 metadata, but no programmer in his right mind will write a program to use them because that is a lot of effort to put into something that is not already used by large numbers of people, and that is not controlled by any published standard.

I don't care about the programs programmers will write. All I care for is a tool that suits me best. Since there's a least tens of thousands consumers that want the same, I think there IS a demand.

[usually_quiet]

While others are interested in using metadata embeddied in files, the one XBMC developer who showed any interest in the project basically said the only video standard he would consider making a tag reader for was MP4 with iTunes atoms or ID3 data. The reason being that no formal standards exist for any other kind of metadata embedded in media files commonly used by consumers.

MPEG2 files ARE commonly uses by consumers (The ones who rip DVD's and don't want to do conversions, because they want the real thing (DVD quality). There may be no formal standards to embed metadata in MPEG2 files, but this doesn't mean there couldn't be such a formal standard and it doesn't mean there's no demand for this.

...and who will write, publish, and oversee this MPEG-2 standard? Go ahead and write your own, but chances are few will be willing to use it without official sanction from the Motion Picture Expert's Group (MPEG), who controls the MPEG-2 standard. [Edit]or support from some other body, like the ISO or the CEA, which maintain and/or publish standards governing the use of MPEG-2 user data for other purposes.[End Edit]

MPEG-7 is more practical, because it is controlled by a standard, but it isn't popular with consumers yet, so there is no demand for a consumer product that uses it.

Are you sure there isn't a demand for MPEG-7? It's new and few have ever heard of it.

No third parties are going to be willing to take a chance on supporting something that isn't already being widely used by the target market for their software. Apple designed and made iTunes as a consumer product, and marketed it to create demand. Third parties followed only after demand was established.

As has also already been said, the other problem is that using the GOP user data is an imperfect solution. If you use the GOP user area for your tags, that will erase any N. American style closed captions already there. (VLC and some other players can display them.)

Are there GOP user data inside MPEG2's that have been ripped from DVD? In the MPEG2's I opened in VLC, I couldn't find any data, except the file title.

You live in PAL land. Trust me, in NTSC land things are very different. By law, the majority of TV broadcasts in N. America must be closed captioned. I have dozens of recordings stored on my computer as .mpg or .ts files that have them. Many older commercially produced NTSC DVDs have closed captions, but not English subtitles. I own at least one of those DVDs as well, and my DVD collection is tiny.

In other words, you can make up your own standards for MPEG-2 metadata, but no programmer in his right mind will write a program to use them because that is a lot of effort to put into something that is not already used by large numbers of people, and that is not controlled by any published standard.

I don't care about the programs programmers will write. All I care for is a tool that suits me best. Since there's a least tens of thousands consumers that want the same, I think there IS a demand.

Until those those tens of thousands of consumers are organized and willing to pay development costs for the product they want, it probably won't happen.

[HitTheRoad]

Usually_quiet and Cornucopia,

Thank you for your patience. As I said before, I'm like Alice in Wonderland. I'm quite new to the world of video codecs, etcetera. So I keep on asking, and keep on learning from what you're telling me.

...and who will write, publish, and oversee this MPEG-2 standard? Go ahead and write your own, but chances are few will be willing to use it without official sanction from the Motion Picture Expert's Group (MPEG), who controls the MPEG-2 standard.

Never heard of any Motion Picture Expert's Group. But I can imagine they don't like people to mess around with MPEG's. The only thing they want us to do is to buy and play their DVD's in stand alone DVD-players.

No third parties are going to be willing to take a chance on supporting something that isn't already being widely used by the target market for their software.

If I have what a want (a MPEG2-player/tag editor/interface), what do I need third parties for?

You live in PAL land. Trust me, in NTSC land things are very different. By law, the majority of TV broadcasts in N. America must be closed captioned.

I have quite a lot NTSC DVD's that I ripped to MPEG2. How can I see/find the GOP user data? You said something about VLC being able to visualize this?

Until those those tens of thousands of consumers are organized and willing to pay development costs for the product they want, it probably won't happen.

There's freeware everywhere. Nobody has paid development costs for this.

[usually_quiet]

...and who will write, publish, and oversee this MPEG-2 standard? Go ahead and write your own, but chances are few will be willing to use it without official sanction from the Motion Picture Expert's Group (MPEG), who controls the MPEG-2 standard.

Never heard of any Motion Picture Expert's Group. But I can imagine they don't like people to mess around with MPEG's. The only thing they want us to do is to buy and play their DVD's in stand alone DVD-players.

Sorry, I misremembered the name. It is "Moving Pictures Experts Group" (MPEG), as in MPEG-1, MPEG-2, MPEG-4, MPEG-7.

No third parties are going to be willing to take a chance on supporting something that isn't already being widely used by the target market for their software.

If I have what a want (a MPEG2-player/tag editor/interface), what do I need third parties for?

You don't have one yet, do you? That is why you need third parties (developers) to create a MPEG2-player/tag editor/interface. ...unless you plan to do all the work yourself.

You live in PAL land. Trust me, in NTSC land things are very different. By law, the majority of TV broadcasts in N. America must be closed captioned.

I have quite a lot NTSC DVD's that I ripped to MPEG2. How can I see/find the GOP user data? You said something about VLC being able to visualize this?

In the Windows version of VLC you would use click Video, then Subtitles, then Closed captions 1. The other closed caption fields are not used as often.

Until those those tens of thousands of consumers are organized and willing to pay development costs for the product they want, it probably won't happen.

There's freeware everywhere. Nobody has paid development costs for this.

Not really true. A lot of "freeware" projects are abandoned if not enough users donate money. The developers get tired of working endless hours on the project and providing support for free. Either that or the project turns into adware or nagware.