Thoughts on WebM
The announcement of WebM, in general, was very positively accepted by the habrasociety, but here most of the conclusions that are viewed in the comments to the relevant topics, in my opinion, show excessive optimism and misunderstanding by many readers of many important points that can affect success (or failure) and WebM deployment speed. In this short article I will try to briefly describe the points that should be taken into account for a greater understanding of this whole situation with WebM, H.264, and so on.
Although these parameters, paradoxically as it may sound, are not decisive, they nevertheless play a significant role in the “struggle of formats”. If you look objectively and do not start from fanatical dogmas, Theora was frankly weaker than H.264 and although supporters of Theora constantly said that “it can be finished and done better”, everyone understood that in this case it’s not clear to all of this who, how and how much will be engaged and who will ultimately pay for it, so this will stretch out for unclear periods. With WebM, a similar situation is potentially possible. Despite the statements of On2, there is no reliable and verifiable evidence that WebM (or rather VP8) is better than H.264. But there are the words of the developer x264, which, although it can be considered an interested party, is nevertheless quite competent in this area,
This parameter is very critical, primarily for mobile gadgets, consumer electronics and weak netbooks. Although they promise us a low CPU load when playing WebM, it’s honestly said that so far it’s never weak, but they promise to fix it in the future. True, the reliability of these promises is in great doubt - although I am not an expert on video codecs, it is in principle clear that a processor will have a weak load only if the video compression ratio is weak, which, accordingly, will result in the large amount of data that the video will occupy. For the good, this problem can be solved quite simply if the decoding of the WebM format will be done in hardware, for example, with a computer video card or SoC gadget, but the following problem comes up from here.
One of the factors due to which the H.264 format began to rapidly gain popularity is the support of its hardware acceleration, not only by computers, but also by a variety of consumer electronics and gadgets, including well-known Apple devices. If existing generations of computer graphics cards are likely to be able (if manufacturers wish) to support WebM hardware acceleration (for example, at the driver level), then in the case of integrated SoCs this is practically impossible. To support hardware acceleration WebM will have to make a new player / smartphone, based on the new SoC, which still need to be designed and put into production, and this is far from the smallest cost for their manufacturers. Moreover, the costs of implementing WebM support in many SoCs can be regarded by manufacturers as such which do not make sense, because for this it is necessary to do unnecessary work (and this is the cost), and the meaning for them is less than obvious - at what pace will WebM gain popularity, will it be, who will release the content in it, and whether the ordinary consumer needs it ? As a result, this may result in a significant decrease in the rate of distribution of the new format.
Probably the most unpleasant topic that few people want to raise. The fact is that, as has already been said, “just because something is open source, it doesn't mean or guarantee that it doesn't infringe on others patents” - as if this were not the case, but these are cruel realities that also need to be considered in this matter. As far as I know, no claims have been made against WebM (or rather VP8) so far, but given that this standard was developed by On2, this cannot be ruled out in the future. At least claims against Theora, which was based on VP3 (as you know, this is also a child of On2), have been heard several times, although no legal proceedings have yet been followed, but this example is indicative.
Of course, developers are already actively working to implement WebM in libraries of decoders, encoders, browsers, and so on, but it’s worthwhile to understand that the statuses “almost ready” and “ready” are, as they say, two big differences. In the end, the web video is not limited to YouTube (otherwise it could be said that H.264 won a long time ago, although it was hiding, in most cases, behind a Flash player), but how convenient is it to use WebM for third parties? What software is it supported in? When it will be? At the same time, “long-ready” H.264 continues to implement in full swing many companies and projects, because everything is clear with it and everything is ready, recently even one large domestic social network introduced it :)
It is about the attractiveness of introducing the format for third parties. We must try to answer the question “Why should you implement WebM? What are its advantages? ”By the companies. From the obvious, we can recall that it is open and free from paying royalties, but the fact is that these factors in the world of large business have little effect on it simply because it is sometimes easier to pay and get a ready-made solution. Many should remember OGG Vorbis, which, in fact, was neglected in favor of MP3, although it was also open and free from royalties. Such facts “Promoted by Google” and “Supported by multiple browsers” should significantly increase the attractiveness - the first gives hope that the whole thing will not be abandoned after the premiere, but the second is already practical. But just with the second fact, problems may arise. No matter how I or someone else personally relate to the IE browser, the fact that this browser continues to be the leader in its share and occupies more than 50% in the world (although it has been steadily losing it lately) still remains a fact. If the new version of IE does not, as they say, support WebM out of the box (and the MS statement regarding WebM support, if such a codec is present in the system, this can be interpreted in this way), then this will be a very negative factor for the speed of WebM implementation by others market participants (not browsers, but projects).
It looks very positive that so many companies from various industries have expressed their support for the new format, but this fact should not be overestimated. The fact is that in the field of IT, like any other field of activity, “promising does not mean getting married” and often support is almost “for the company”, after which, in fact, there is little to be done if something is done - then generally.
How can I get around this point? :) There is every reason to believe that MPEG LA is not full of fools, although it seems to many different :) Therefore, if they do not want the positions of one of their standards to be shaken (and they hardly want it), they can have quite strong resistance on two main fronts.
The first is judicial claims regarding the violation by the VP8 format of MPEG LA patents. Despite the fact that software patents are far from being available in all countries of the world, problems in the USA alone will be enough to “cripple” the introduction of a new format.
The second is the increase in the attractiveness of H.264 compared to its competitor, and for this, you will not have to do anything with the format itself, but it will be enough to liberalize the conditions for its use. What do you think, how many will be worn with WebM if MPEG LA removes the most acute claims to the format? And this, I recall, is paid for by manufacturers of free browsers and, so far, theoretical, the need to pay for its use to end users. MPEG LA could very well do it under the threat of “losing web video”, because they still get the bulk of the money and will not receive it from Mozilla and Opera.
I hope that this small note written in the middle of the night will be useful and understandable to you :) If I remember something else during the day, I will update the note.
My personal opinion is that WebM will ultimately take a decent share in the web video market, for there is still a lot of effort to advance (and Google is unlikely to score on this project, otherwise they would not buy On2) :) but its pace for some reason, the implementations do not seem so rosy to me, and indeed the continuation of the “format war” is not the most pleasant and useful for end users.
PS: in the end, the format that the producers of porn will choose anyway will win anyway :)
Quality and Efficiency
Although these parameters, paradoxically as it may sound, are not decisive, they nevertheless play a significant role in the “struggle of formats”. If you look objectively and do not start from fanatical dogmas, Theora was frankly weaker than H.264 and although supporters of Theora constantly said that “it can be finished and done better”, everyone understood that in this case it’s not clear to all of this who, how and how much will be engaged and who will ultimately pay for it, so this will stretch out for unclear periods. With WebM, a similar situation is potentially possible. Despite the statements of On2, there is no reliable and verifiable evidence that WebM (or rather VP8) is better than H.264. But there are the words of the developer x264, which, although it can be considered an interested party, is nevertheless quite competent in this area,
CPU load
This parameter is very critical, primarily for mobile gadgets, consumer electronics and weak netbooks. Although they promise us a low CPU load when playing WebM, it’s honestly said that so far it’s never weak, but they promise to fix it in the future. True, the reliability of these promises is in great doubt - although I am not an expert on video codecs, it is in principle clear that a processor will have a weak load only if the video compression ratio is weak, which, accordingly, will result in the large amount of data that the video will occupy. For the good, this problem can be solved quite simply if the decoding of the WebM format will be done in hardware, for example, with a computer video card or SoC gadget, but the following problem comes up from here.
Hardware acceleration
One of the factors due to which the H.264 format began to rapidly gain popularity is the support of its hardware acceleration, not only by computers, but also by a variety of consumer electronics and gadgets, including well-known Apple devices. If existing generations of computer graphics cards are likely to be able (if manufacturers wish) to support WebM hardware acceleration (for example, at the driver level), then in the case of integrated SoCs this is practically impossible. To support hardware acceleration WebM will have to make a new player / smartphone, based on the new SoC, which still need to be designed and put into production, and this is far from the smallest cost for their manufacturers. Moreover, the costs of implementing WebM support in many SoCs can be regarded by manufacturers as such which do not make sense, because for this it is necessary to do unnecessary work (and this is the cost), and the meaning for them is less than obvious - at what pace will WebM gain popularity, will it be, who will release the content in it, and whether the ordinary consumer needs it ? As a result, this may result in a significant decrease in the rate of distribution of the new format.
Patent Security
Probably the most unpleasant topic that few people want to raise. The fact is that, as has already been said, “just because something is open source, it doesn't mean or guarantee that it doesn't infringe on others patents” - as if this were not the case, but these are cruel realities that also need to be considered in this matter. As far as I know, no claims have been made against WebM (or rather VP8) so far, but given that this standard was developed by On2, this cannot be ruled out in the future. At least claims against Theora, which was based on VP3 (as you know, this is also a child of On2), have been heard several times, although no legal proceedings have yet been followed, but this example is indicative.
Ready for use
Of course, developers are already actively working to implement WebM in libraries of decoders, encoders, browsers, and so on, but it’s worthwhile to understand that the statuses “almost ready” and “ready” are, as they say, two big differences. In the end, the web video is not limited to YouTube (otherwise it could be said that H.264 won a long time ago, although it was hiding, in most cases, behind a Flash player), but how convenient is it to use WebM for third parties? What software is it supported in? When it will be? At the same time, “long-ready” H.264 continues to implement in full swing many companies and projects, because everything is clear with it and everything is ready, recently even one large domestic social network introduced it :)
Attractiveness
It is about the attractiveness of introducing the format for third parties. We must try to answer the question “Why should you implement WebM? What are its advantages? ”By the companies. From the obvious, we can recall that it is open and free from paying royalties, but the fact is that these factors in the world of large business have little effect on it simply because it is sometimes easier to pay and get a ready-made solution. Many should remember OGG Vorbis, which, in fact, was neglected in favor of MP3, although it was also open and free from royalties. Such facts “Promoted by Google” and “Supported by multiple browsers” should significantly increase the attractiveness - the first gives hope that the whole thing will not be abandoned after the premiere, but the second is already practical. But just with the second fact, problems may arise. No matter how I or someone else personally relate to the IE browser, the fact that this browser continues to be the leader in its share and occupies more than 50% in the world (although it has been steadily losing it lately) still remains a fact. If the new version of IE does not, as they say, support WebM out of the box (and the MS statement regarding WebM support, if such a codec is present in the system, this can be interpreted in this way), then this will be a very negative factor for the speed of WebM implementation by others market participants (not browsers, but projects).
Manufacturers support
It looks very positive that so many companies from various industries have expressed their support for the new format, but this fact should not be overestimated. The fact is that in the field of IT, like any other field of activity, “promising does not mean getting married” and often support is almost “for the company”, after which, in fact, there is little to be done if something is done - then generally.
Resist Competitors
How can I get around this point? :) There is every reason to believe that MPEG LA is not full of fools, although it seems to many different :) Therefore, if they do not want the positions of one of their standards to be shaken (and they hardly want it), they can have quite strong resistance on two main fronts.
The first is judicial claims regarding the violation by the VP8 format of MPEG LA patents. Despite the fact that software patents are far from being available in all countries of the world, problems in the USA alone will be enough to “cripple” the introduction of a new format.
The second is the increase in the attractiveness of H.264 compared to its competitor, and for this, you will not have to do anything with the format itself, but it will be enough to liberalize the conditions for its use. What do you think, how many will be worn with WebM if MPEG LA removes the most acute claims to the format? And this, I recall, is paid for by manufacturers of free browsers and, so far, theoretical, the need to pay for its use to end users. MPEG LA could very well do it under the threat of “losing web video”, because they still get the bulk of the money and will not receive it from Mozilla and Opera.
Instead of conclusions
I hope that this small note written in the middle of the night will be useful and understandable to you :) If I remember something else during the day, I will update the note.
My personal opinion is that WebM will ultimately take a decent share in the web video market, for there is still a lot of effort to advance (and Google is unlikely to score on this project, otherwise they would not buy On2) :) but its pace for some reason, the implementations do not seem so rosy to me, and indeed the continuation of the “format war” is not the most pleasant and useful for end users.
PS: in the end, the format that the producers of porn will choose anyway will win anyway :)