Talk:VP9

History section

The history section of this article reads like a list from a diary and is overly precise and detailed. Look at WP:NOTDIARY. It should be rewritten as prose such that the key events are described in relation to their continued importance over time. Editors wanting to add to the history section should consider updating or replacing existing sentences rather than just adding yet another line on the end. Rincewind42 (talk) 07:23, 8 April 2014 (UTC)

Some of it could be moved to an Implementations section, like on VP8. At least the "implementation" details. 84.214.220.135 (talk) 19:06, 3 May 2016 (UTC)

For example, I would like to add Eve, but it doesn't belong in history. 84.214.220.135 (talk) 21:54, 3 May 2016 (UTC)

VP9 may not be royalty free and open.

As of the time of this writing, Google has not released any patent grants for VP9, and the statement that it is an "Open and royalty-free" codec based on a marketing statement doesn't really give any proper reference to claim this.

In fact, it's been a point of discussion at the Mozilla legal department, see https://bugzilla.mozilla.org/show_bug.cgi?id=1026707 A Google representative has made a statement that Google "should" release proper licensing "soon", but nothing concrete has happened in the month or so since this was stated.

With the wide-spread pushing of Google through Google-owned products that obviously don't have to worry about patent grants (YouTube, Chrome, ChromeOS, etc.) and the lack of clear licensing, it's also a point to consider for everyone incorporating/using VP9 in the library included in many Open Source operating systems (libvpx includes it since 1.3.0, a silently rolled out point release).

In short, the codec itself is patent-encumbered without a patent grant at this time. The implementation is proprietary and copyrighted by Google Inc. The implementation Source is released, but the actual use of VP9 is not within the realm of FOSS without a proper patent grant.

Perhaps someone can investigate this and adapt this Wikipedia article accordingly to make an accurate representation about the current state of licensing and about the codec not necessarily being open, nor royalty-free. The current state of affairs is that Google may at any time claim royalties if they so choose...

Wolfbeast (talk) 23:01, 28 July 2014 (UTC)

Google is a reliable source and their deal with the MPEG LA did cover both VP8 and one next generation video standard. --GrandDrake (talk) 04:31, 19 August 2014 (UTC)

Only one, does that spell disaster for VP10 then? 84.214.220.135 (talk) 21:51, 3 May 2016 (UTC)

The whole article seems to be skewed to be pro VP9, and the numbers that are presented are really meaningless because since Chrome, a google product, is one of the most common browsers, that would influence the numbers greatly. 216.191.105.146 (talk) 22:49, 15 November 2017 (UTC)

Which numbers are meaningless? Just because google owns chrome that doesn't make all numbers meaningless. It does not read particularly "pro" compared to other articles like High Efficiency Video Coding. Of course you are free to add a "Criticisms" section and add any reputably published criticism -- Q Chris (talk) 08:57, 16 November 2017 (UTC)

I don't see how the fact that Chrome is a common browser skews any numbers in this article. (Which ones, actually? The only references to Chrome that mention performance are actually *negative*.) To me this article seems fairly neutral, and all the numbers given related to performance and efficiency seem well-cited. I'm removing that notice, as there is no proper reason given.Slhck (talk) 19:46, 16 November 2017 (UTC)

Hardware → Hardware/GPU ?

Some of those "hardware" codecs are OpenCL implementations – software. Is it worth pointing out such an implementation detail for the sake of absolute correctness? In all fairness, there are firmware blobs in H.264 "hardware" too! 84.214.220.135 (talk) 22:15, 16 June 2016 (UTC)

AMD Polaris does not support VP9 Encoding

For some reason, this wiki page states that Polaris supports VP9 encoding. It does not, or at least not natively.

--Xaymar (talk) 14:28, 28 October 2016 (UTC)

References

Neutrality

While I personally support the use of webm, the style of the article feels at times like an advertisement of VP9 to me, especially passages like "with the clear development perspective and support from the industry demonstrated by the founding of the Alliance for Open Media, as well as the pricey and complex licensing situation of HEVC it is expected that users of the hitherto leading MPEG formats will often switch to the royalty-free alternative formats of the VPx/AVx series instead of upgrading to HEVC." I went ahead an tried to make the introduction a bit more neutral, but right now I'm leaving the rest as is, with the addition of a warning template. --Yury Bulka (talk) 20:10, 9 February 2020 (UTC)

Accelerated hardware encoding / decoding table

J._M.

Karakarga70

I think I know where the issue is: the table is formatted in a really weird way. It is only for devices with accelerated decoding, but the additional column is asking for encoding (it's not a mistake, all references inside point to encoding information). Which means that those Gforces can be added to the list, because the referenced | support table from nVidia tells under decoding table the state of VP9 decoding.

But for the check in the last column, they must support encoding. I hope I didn't mix up anything. Does that make sense? — K4rolB (talk) 20:00, 17 March 2021 (UTC)

Yes, the table includes products that support VP9 decoding. The "Encoding" column is an extra column that says which of those listed products that support decoding support encoding, too. There is nothing weird about it. While decoding support is quite common, encoding support is much more rare. You probably won't find products that support encoding, and not decoding. Unfortunately, Karakarga70 completely destroyed the table with their last edit that completely changed the meaning for all products, and therefore products that supported decoding were listed as products that do not support decoding.

If the column said "Decoding", it would of course, logically, be completely redundant and pointless, because the table itself only lists products that support decoding. The columns then add additional information, like encoding support. Unfortunately, Karakarga70 is either not capable of understanding anything, despite repeated explanation, and despite even themselves admitting they were wrong, or keeps pretending that they don't understand anything, and keeps trolling and edit warring.—J. M. (talk) 20:10, 17 March 2021 (UTC)

I added missing Nvidia products to the table, according to the GPU support matrix page. Feel free to correct it if I made some mistakes. And I hope this will put an end to this affair.—J. M. (talk) 21:52, 17 March 2021 (UTC)

Hardware implementations

The table under "Hardware implementations" is outdated, not properly maintained and has become redundant because almost every GPU and SoC released in the past five years supports VP9 decoding and encoding hasn't taken off and unlikely it ever will considering AV1 is all the rage nowadays. I suggest we delete this table altogether and maybe mention only holdouts, i.e. companies/chips which do not support VP9 which I can't think there are any nowadays. VP9 has become sort of commodity. Artem S. Tashkinov (talk) 11:18, 17 May 2023 (UTC)