Recherche avancée

Sur d’autres sites (11237)

• Revision 869d4ca519 : Add adaptation option for VBR. Allow min and maxQ to creep when the undershoot

  9 octobre 2014, par Paul Wilkins

  Changed Paths :
   Modify /vp9/encoder/vp9_firstpass.c

  
   Modify /vp9/encoder/vp9_firstpass.h

  
   Modify /vp9/encoder/vp9_ratectrl.c

  
   Modify /vp9/encoder/vp9_ratectrl.h

  
  
  Add adaptation option for VBR.

  Allow min and maxQ to creep when the undershoot
  or overshoot exceeds thresholds controlled by the
  command line under_shoot_pct and over_shoot_pct
  values.

  Default is 100%,100% which disables adaptation.

  Derf results for example undershoot% / overshoot% :-

  Head :- Mean abs (%rate error) = 14.4%

  This check in :-
  25%/25% - Mean abs (%rate error) = 6.7%
  PSNR hit -1% SSIM -0.1%

  5% / 5% - Mean abs (%rate error) = 2.2%
  PSNR hit -3.3% SSIM - 1.1%

  Most of the remaining error and most of the quality hit is
  at extreme data rates. The adaptation code still has an
  exception for material that is in effect static so that we
  don’t over adjust and over spend on YT slide show type
  content.

  Change-Id : If25a2449a415449c150acff23df713e9598d64c9

• Revision 0e1068a4bd : Improve two pass VBR accuracy. Adjustments to the GF interval choice and minimu

  26 septembre 2014, par Paul Wilkins

  Changed Paths :
   Modify /vp9/encoder/vp9_encodeframe.c

  
   Modify /vp9/encoder/vp9_firstpass.c

  
   Modify /vp9/encoder/vp9_ratectrl.c

  
  
  Improve two pass VBR accuracy.

  Adjustments to the GF interval choice and minimum boost.
  Adjustment to the calculation of 2 pass worst q.
  Compared to 09/29 head there is metrics hit on derf of
  (-0.123%,-0.191%)

  Compared to the September 29 head and a baseline on
  September 18 baseline the accuracy of the VBR rate control
  measured on the derf set is as follows :-

  Mean error % / Mean abs(error %)
  Sept 18 baseline (-7.0% / 14.76%)
  Sept 29 head (-15.7%, 19.8%)
  This check in (-1.5% / 14.4%)

  The mean undershoot is reduced slightly but the
  worst case overshoot on e.g. harbour/highway is
  increased. This will be addressed in a later patch.

  Change-Id : Iffd9b0ab7432a131c98fbaaa82d1e5b40be72b58

• Evolution of Multimedia Fiefdoms

  1er octobre 2014, par Multimedia Mike — General

  I want to examine how multimedia fiefdoms have risen and fallen through the years.

  
  
  

  Back in the day, the multimedia fiefdoms were built around the formats put forth by competing companies : there was Microsoft/WMV, Apple/MOV, and Real/RM as the big contenders. On2 always wanted to be a player in this arena but could never quite catch a break. A few brave contenders held the line for open source and also for the power users who desired one application that could handle everything (my original motivation for wanting to get into multimedia hacking).

  The computer desktop was the battleground for internet-based media stream. Whatever happened to those days ? Actually, if memory serves, Flash-based video streaming stepped on all of them.

  Over the last 6-7 years, the battleground has expanded to cover mobile devices, where Flash’s impact has… lessened. During this time, multimedia technology pretty well standardized on a particular stack, namely, the MPEG (MP4/H.264/AAC) stack.

  The belligerents in this war tried for years to effectively penetrate new territory, namely, the living room where the television lived. This had been slowgoing for years due to various user interface and content issues, but steadily improved.

  Last April, Amazon announced their entry into the set-top box market with the Fire TV. That was when it suddenly crystallized for me that the multimedia ecosystem has radically shifted. Now, the multimedia fiefdoms revolve around access to content via streaming services.

  Off the top of my head, here are some of the fiefdoms these days (fiefdoms I have experience using) :

  • Netflix (subscription streaming)
  • Amazon (subscription, rental, and purchased streaming)
  • Hulu Plus (subscription streaming)
  • Apple (rental and purchased media)

  I checked some results on Can I Stream.It ? (which I refer to often) and found a bunch more streaming fiefdoms such as Google (both Play and YouTube, which are separate services), Sony, Xbox 360, Crackle, Redbox Instant, Vudu, Target Ticket, Epix, Sony, SnagFilms, and XFINITY StreamPix. And surely, these are probably just services available in the United States ; I know other geographical regions have their own fiefdoms.

  What happened ?

  When I got into multimedia hacking, there were all these disparate, competing ecosystems. As a consumer, I didn’t care where the media came from, I just wanted to play it. That’s what inspired me to work on open source multimedia projects. Now I realize that I have the same problem 10-15 years later : there are multiple competing ecosystems. I might subscribe to fiefdoms X and Y, but am frustrated to learn that something I’d like to watch is only available through fiefdom Z. Very few of these fiefdoms can be penetrated using open source technology.

  I’m not really sure about the point about this whole post. Multimedia technology seems really standardized these days. But that’s probably just my perspective because I have spent way too long focusing on a few areas of multimedia technology such as audio and video coding. It’s interesting that all these services probably leverage the same limited number of codecs. Their differentiation comes from the catalog of content that each is able to license for streaming. There are different problems to solve in the multimedia arena now.