Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > VP9 and AV1

Reply
 
Thread Tools Search this Thread Display Modes
Old 9th November 2018, 20:16   #1201  |  Link
Mr_Khyron
Member
 
Mr_Khyron's Avatar
 
Join Date: Nov 2002
Posts: 203
Microsoft release AV1 Video codec for Windows 10

https://mspoweruser.com/microsoft-re...or-windows-10/
Quote:
Microsoft has released support for the new AV1 royalty-free video codec for Windows 10 via the Microsoft Store.

AOMedia Video 1 (AV1), is an open, royalty-free video coding format designed for video transmissions over the Internet. It is being developed by the Alliance for Open Media (AOMedia) and is meant to be a successor to VP9 without relying on any MPEG patents.

The AV1 extension in the Microsoft Store is an early beta version of the AV1 software decoder. Since this is an early release, users may see some performance issues when playing AV1 videos.

Microsoft says they will be regularly updating the codec via automatic store updates.

Find the new codec in the Microsoft Store here.
Mr_Khyron is offline   Reply With Quote
Old 10th November 2018, 10:08   #1202  |  Link
hydra3333
Registered User
 
Join Date: Oct 2009
Location: crow-land
Posts: 540
he he, clicked on "Get" 30 times in the microsoft store and nothing happens ... that may be saying something about quality.
hydra3333 is offline   Reply With Quote
Old 10th November 2018, 13:51   #1203  |  Link
v0lt
Registered User
 
Join Date: Dec 2008
Posts: 1,959
Quote:
Originally Posted by Mr_Khyron View Post
I tried bencmarking with ffmpeg 4.2
from 16fps with libaom to 77fps with Dav1d
Where can I download ffmpeg with libdav1d library?
v0lt is offline   Reply With Quote
Old 10th November 2018, 14:22   #1204  |  Link
lvqcl
Registered User
 
Join Date: Aug 2015
Posts: 293
Quote:
from 16fps with libaom to 77fps with Dav1d
AFAICS dav1d has only x86-64 AVX2 assembly code, right?
I wonder what's their plans about older hardware...
lvqcl is offline   Reply With Quote
Old 10th November 2018, 15:04   #1205  |  Link
SmilingWolf
I am maddo saientisto!
 
SmilingWolf's Avatar
 
Join Date: Aug 2018
Posts: 95
Quote:
Originally Posted by Clare View Post
aomenc -v --threads=8 --cpu-used=4 --row-mt=1 --lag-in-frames=25 --auto-alt-ref=1--passes=2 --pass=2 --bit-depth=10 --input-bit-depth=10 --end-usage=q --cq-level=28 -o Chimera_DCI4k2398p_HDR_P3PQ.ivf Chimera_DCI4k2398p_HDR_P3PQ.y4m
That worked, thanks!

Quote:
Originally Posted by v0lt View Post
Where can I download ffmpeg with libdav1d library?
Win64 GCC 8.2 static build:
ffmpeg-4.2-92396-g55e021f39b: https://mega.nz/#!IgAAVayA!jpzHzBaE6...FnbjR-ruOD8lCI
- libaom 1.0.0-902-g03d8ebedc
- libdav1d 58fc516

Last edited by SmilingWolf; 10th November 2018 at 15:06.
SmilingWolf is offline   Reply With Quote
Old 10th November 2018, 18:58   #1206  |  Link
NikosD
Registered User
 
Join Date: Aug 2010
Location: Athens, Greece
Posts: 2,901
In order to GET the new MS AV1 codec from MS Store, you need to install the forbidden (banned) Windows October 2018 Update.

Test:
MS Windows October x64
Core i3 4170
DXVA Checker (new beta version)

Sample:
Chimera AV1 1080p 8bit (Netflix free sample)

LAV x64 0.73.1 vs MS MFT AV1

LAV x64 19/34/144 (min/avg/max fps) CPU Usage: 57/70/83 (%)

MS MFT AV1 15/26/156 CPU Usage: 50/68/81

It seems that AOM AV1 codec is ~30% faster than MS MFT AV1 on average fps
__________________
Win 10 x64 (19042.572) - Core i5-2400 - Radeon RX 470 (20.10.1)
HEVC decoding benchmarks
H.264 DXVA Benchmarks for all
NikosD is offline   Reply With Quote
Old 10th November 2018, 19:16   #1207  |  Link
v0lt
Registered User
 
Join Date: Dec 2008
Posts: 1,959
@SmilingWolf
Thank.
But my results are different from those that were announced here.
I ran the following tests:
Code:
ffmpeg -hide_banner -t 10 -c:v libaom-av1 -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
ffmpeg -hide_banner -t 10 -c:v libdav1d -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
ffmpeg -hide_banner -t 10 -c:v libdav1d -threads 4 -tilethreads 4 -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
And got the following results:
libaom-av1 - max 14 fps
libdav1d - max 7.4 fps
libdav1d -threads 4 -tilethreads 4 - max 9.7 fps

Added:
Intel i5-3570k, Windows 7 Sp1 x64.

Last edited by v0lt; 13th November 2018 at 19:22.
v0lt is offline   Reply With Quote
Old 10th November 2018, 20:20   #1208  |  Link
richardpl
Registered User
 
Join Date: Jan 2012
Posts: 272
Probably because you are not using right arch and CPU combo.
richardpl is offline   Reply With Quote
Old 11th November 2018, 04:36   #1209  |  Link
Wolfberry
Helenium(Easter)
 
Wolfberry's Avatar
 
Join Date: Aug 2017
Location: Hsinchu, Taiwan
Posts: 99
Quote:
Originally Posted by v0lt View Post
I ran the following tests:
Code:
ffmpeg -hide_banner -t 10 -c:v libaom-av1 -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
ffmpeg -hide_banner -t 10 -c:v libdav1d -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
ffmpeg -hide_banner -t 10 -c:v libdav1d -threads 4 -tilethreads 4 -i Stream2_AV1_4K_22.7mbps.webm -benchmark -f null -
I ran the same test as above and get 16/38/46 fps.
What is the CPU you use for testing?
It might be related to the AVX2 code used in dav1d.
__________________
Monochrome Anomaly
Wolfberry is offline   Reply With Quote
Old 11th November 2018, 05:12   #1210  |  Link
v0lt
Registered User
 
Join Date: Dec 2008
Posts: 1,959
Quote:
Originally Posted by wolfberry View Post
it might be related to the avx2 code used in dav1d.
sse2, sse4.1?
v0lt is offline   Reply With Quote
Old 11th November 2018, 07:37   #1211  |  Link
Aleksoid1978
Registered User
 
Aleksoid1978's Avatar
 
Join Date: Apr 2008
Location: Russia, Vladivostok
Posts: 2,783
Very "good" optimisation dav1d - much slower on my system...
__________________
AMD Ryzen 5 3600 /GIGABYTE B450 Gaming X /Patriot 32Gb@3200 /Kingston 500Gb M.2 /RTX 4060 /Samsung U28R550UQI /OLED Philips 55OLED707 /Yamaha RX-V471 + NS-555 + NS-C444 + NS-333 + YST-SW215

Last edited by Aleksoid1978; 11th November 2018 at 08:59.
Aleksoid1978 is offline   Reply With Quote
Old 11th November 2018, 09:04   #1212  |  Link
Nintendo Maniac 64
Registered User
 
Nintendo Maniac 64's Avatar
 
Join Date: Nov 2009
Location: Northeast Ohio
Posts: 447
Quote:
Originally Posted by lvqcl View Post
AFAICS dav1d has only x86-64 AVX2 assembly code, right?
I wonder what's their plans about older hardware...
Don't forgot that Pentiums and Celerons don't support AVX, and this includes the models that use full-fat Sky/Kaby/Coffee cores such as the ever-popular 2c/4t G4560 and its successor the G5400 (as well as the variants with the faster iGPU like the G4600 and G5500).

And of course, it's those very same AVX-lacking Celerons and Pentiums and such that would stand to gain the biggest benefit from any such software decoder optimizations because those processors simply lack the raw "moar cores!" computational grunt that their i7 and Ryzen brethren have for brute-forcing their way through.

So needless to say, it'd be pretty disappointing to me if dav1d pretty much required having an AVX-capable CPU in order to have any benefit.


Quote:
Originally Posted by Aleksoid1978 View Post
Very "good" optimisation dav1d - mush slower on my system...
...that's not a Fernando Alonso reference, is it?
Nintendo Maniac 64 is offline   Reply With Quote
Old 11th November 2018, 13:14   #1213  |  Link
Mystery Keeper
Beyond Kawaii
 
Mystery Keeper's Avatar
 
Join Date: Feb 2008
Location: Russia
Posts: 724
I wish aomenc/vpxenc had GOP-level parallelism. When each thread is encoding one GOP, and then they are stitched together. That would make use of all CPU power without compromising quality/compression.
__________________
...desu!
Mystery Keeper is offline   Reply With Quote
Old 11th November 2018, 13:25   #1214  |  Link
Selur
Registered User
 
Selur's Avatar
 
Join Date: Oct 2001
Location: Germany
Posts: 7,259
Quote:
I wish aomenc/vpxenc had GOP-level parallelism.
Which would require 2pass encoding and a fixed gop structue (in regard to the gop sizes), iirc 2nd pass normally should be able to overwrite GOP to archive vbv limits (not totally sure).
__________________
Hybrid here in the forum, homepage
Selur is offline   Reply With Quote
Old 11th November 2018, 13:49   #1215  |  Link
Gravitator
Registered User
 
Join Date: May 2014
Posts: 292
ffmpeg -hide_banner -t 10 -c:v libaom-av1 -i 1.mp4 -benchmark -f null - (43 fps)
ffmpeg -hide_banner -t 10 -c:v libdav1d -i 1.mp4 -benchmark -f null - (52 fps)
ffmpeg -hide_banner -t 10 -c:v libdav1d -threads 1 -tilethreads 2 -i 1.mp4 -benchmark -f null - (61 fps)
ffmpeg -hide_banner -t 10 -c:v libdav1d -threads 2 -tilethreads 2 -i 1.mp4 -benchmark -f null - (65 fps)
Gravitator is offline   Reply With Quote
Old 11th November 2018, 14:08   #1216  |  Link
lvqcl
Registered User
 
Join Date: Aug 2015
Posts: 293
Quote:
Originally Posted by v0lt View Post
sse2, sse4.1?
It seems that one of dav1d developers said: "we don't care about mmx/sse2 support anyway" (link). Have no idea about sse4.1.
lvqcl is offline   Reply With Quote
Old 11th November 2018, 15:12   #1217  |  Link
SmilingWolf
I am maddo saientisto!
 
SmilingWolf's Avatar
 
Join Date: Aug 2018
Posts: 95
Quote:
Originally Posted by lvqcl View Post
It seems that one of dav1d developers said: "we don't care about mmx/sse2 support anyway" (link). Have no idea about sse4.1.
BBB is part of TwoOrioles, so it might have been referred to the company based on its userbase.
Still, MMX is hardly relevant nowadays. SSE4.1 as the lowest bar doesn't sound too unreasonable

Also relevant: https://code.videolan.org/videolan/d.../15#note_22262

Last edited by SmilingWolf; 11th November 2018 at 18:34.
SmilingWolf is offline   Reply With Quote
Old 11th November 2018, 19:06   #1218  |  Link
NikosD
Registered User
 
Join Date: Aug 2010
Location: Athens, Greece
Posts: 2,901
Quote:
Originally Posted by lvqcl View Post
It seems that one of dav1d developers said: "we don't care about mmx/sse2 support anyway"
Have no idea about sse4.1.
Quote:
Originally Posted by SmilingWolf View Post
Still, MMX is hardly relevant nowadays. SSE4.1 as the lowest bar doesn't sound too unreasonable.
MMX is too old and not that beneficial as it can reach only 64bits (maybe 80bits max)
SSEx should be the base as it is 128bit with very fast implementation on all CPUs of the last 10 years.
Especially SSE2 is mandatory for x64 architecture.
From the last link it's obvious that dav1d developers targeted AVX2 for 256bit acceleration using ASM, but not exclusively.
They are going to optimise for SSEx later.
So no worries, I think.
__________________
Win 10 x64 (19042.572) - Core i5-2400 - Radeon RX 470 (20.10.1)
HEVC decoding benchmarks
H.264 DXVA Benchmarks for all
NikosD is offline   Reply With Quote
Old 11th November 2018, 20:50   #1219  |  Link
marcomsousa
Registered User
 
Join Date: Jul 2018
Posts: 80
if they want to go with 4k and 8k videos they have to use AVX2.
__________________
AV1 win64 VS2019 builds
Last build here
marcomsousa is offline   Reply With Quote
Old 12th November 2018, 00:30   #1220  |  Link
Nintendo Maniac 64
Registered User
 
Nintendo Maniac 64's Avatar
 
Join Date: Nov 2009
Location: Northeast Ohio
Posts: 447
Quote:
Originally Posted by NikosD View Post
Especially SSE2 is mandatory for x64 architecture.
You can also usually safely target SSE3 (no, not SSSE3) as well since it's supported on all DDR2-capable 64bit x86 CPUs and newer.

(the only 64bit x86 CPUs that don't support SSE3 are some socket 754 and 939 Athlon 64s which used DDR1)
Nintendo Maniac 64 is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 18:41.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.