Difference between revisions of "Talk:Hardware video acceleration"

From ArchWiki
Jump to navigation Jump to search
(cooment)
(→‎Vp9 in haswell: re, close)
 
(5 intermediate revisions by 3 users not shown)
Line 1: Line 1:
== <s>chromium developpers and vaapi</s> ==
+
== <s>Vp9 in haswell</s> ==
  
chromium developpers don't want to trust linux drivers, neither apply the patch that enables vaapi.
+
Vp9 is not broken in my haswell desktop, it is working well. I have seen the GitHub page and there is anybody complaining this bug. May we delete this sentence ?
have we got to put this link as a source here?
 
https://bugs.chromium.org/p/chromium/issues/detail?id=463440#c65
 
  
like with firefox I think
+
{{unsigned|01:03, 13 February 2020‎|Elcompartidor}}
what do you think?
 
  
PD: sorry for my level of english, I will improve my english with the time
+
:Removed. Maybe fixed by https://github.com/intel/intel-vaapi-driver/commit/71f77c6 ? -- [[User:Svito|Svito]] ([[User talk:Svito|talk]]) 00:41, 14 February 2020 (UTC)
 
 
{{Unsigned|11:30, 19 June 2019 (UTC)|Elcompartidor}}
 
 
 
:I don't think there is much value in reading that thread. It was much better if people actually reported and documented their buggy hardware.
 
:I reverted your recent edits because they don't have associated issue reports to VLC. As for VP9 CPU load compared to H.264 this should be measured on shittiest hardware, because that is where it makes all the difference between playable video and a slideshow. "works for me" and "I think" does not give me any confidence that these should stay.
 
:-- [[User:Svito|Svito]] ([[User talk:Svito|talk]]) 17:34, 19 June 2019 (UTC)
 
 
 
:
 
I have tested it in shittiest harware, I have an dell inspiron 1545 with core 2 duo T6600, Vp9 uses less cpu in firefox than H264 now, it changed recent.
 
It has not been reported because VLC uses vdpau instead of vaapi if auto option is marked, the people don't know about hadware video accel, it is the cause that is not reported, and much people don't see if there is vaapi acceleration in fact, by example, the people thinks that firefox can accelerate video with GPU and.. they don't ... VLC at least in my hardware and 2 others can't use vaapi, and in my shittiest hardware and many other in firefox the use of this complement is for nothing, if you don't trust me I let you to test, but I think that to delete something because you don't find anything about it is an error, and at the same way that we have post the link about the bug in firefox, why not to add the same bug in chromium? both are the same, one is added and the other ones not...
 
:-- [[User:elcompartidor|elcompartidor]] ([[User talk:elcompartidor|talk]]) 17:34, 19 June 2019 (UTC)
 
 
 
::If you followed VLC link it does say how to switch API. "Does not work" without reasoning and proof is useless and incorrect information. If you follow Chromium link in "Documentation column" it has link to mentioned bug report and packages that add support for VA-API. As for firefox bug report there is no such page on the wiki, because feature is not implemented at all, not unlike chromium where it is disabled on Linux, but is fully implemented in project code. -- [[User:Svito|Svito]] ([[User talk:Svito|talk]]) 18:51, 19 June 2019 (UTC)
 
 
 
:
 
It is not working because I have tried to make it work with many video outputs and many vaapi options and anyone was working, all got the same CPU use, I have read all the documentation about it, VLC has old docs, like saying that you need to put xvba for amd to use vaapi... old like catalyst drivers.
 
:-- [[User:elcompartidor|elcompartidor]] ([[User talk:elcompartidor|talk]])
 

Latest revision as of 00:41, 14 February 2020

Vp9 in haswell

Vp9 is not broken in my haswell desktop, it is working well. I have seen the GitHub page and there is anybody complaining this bug. May we delete this sentence ?

—This unsigned comment is by Elcompartidor (talk) 01:03, 13 February 2020‎. Please sign your posts with ~~~~!

Removed. Maybe fixed by https://github.com/intel/intel-vaapi-driver/commit/71f77c6 ? -- Svito (talk) 00:41, 14 February 2020 (UTC)