Posts Tagged ‘review’

MPEG 4 Part -2 / AVI video codec performance review: Divx vs Xvid which is better?

MPEG-4 Part 2 is a video compression technology developed by MPEG. It belongs to the MPEG-4 ISO/IEC standard (ISO/IEC 14496-2). It is a discrete cosine transform compression standard, similar to previous standards such as MPEG-1 and MPEG-2. Several popular codecs including DivX, Xvid and Nero Digital are implementations of this standard.

If you were looking for MP4/Mpeg4 Part 10  CoreAVC vs FFmpeg video codec performance review then click here

Codecs:

Though there are several codecs that perform the encoding owing to the universal use of this encoding in dvds, common implementations are Nero, Divx, Xvid etc. Among these arguably the most popular codecs are Xvid and Divx. We will be seeing side by side comparison of both in the article. You can find both codecs individual history on their wiki pages. I find the history of Divx quite amusing,They adopted their name Divx to mock a company that used to charge users for viewing by the hour and themselves used spyware in their later version hence becoming a subject of mockery. You might have already noticed that “Xvid” is “Divx” backwards.

I will talk of decoding performance here. Decoding probably makes much more sense to a normal user as encoding is done by few but everybody decodes videos. Decoding is quite a subjective criteria hence besides my own comments I have also attached screenshots that allow the user to make his/her own judgment. The codecs were taken from Final build site.

Xvid
Xvid Logo
There is a very good chance that any video that you come across is encoded with Xvid. Across my search of encoders I found Xvid is the most preferred one for encoding. Its opensource hence free, much more configurable hence lets you draw the last byte’s worth. Though it seems logical that if video is encoded with encoder its respective decoder should also be the best, however I found the results didn’t fare as I expected.

Divx
Divx Logo

The good old closed source version. Divx is costly and costing is on per PC basis hence encoding costly. Decoding can be done for free using their web player. The pro version is not free however there are other ways of using the codec in Media Player classic shipped with Klite Codec as explained below.

Here is the screenshot of Gspot codec analyzation. This provides the codec information with which the file was encoded in:

GSPOT Screenshot of Sherlock Holmes Important things that can be highlighted from the Gspot codec is

1. Video Encoding was done by Divx version 5.1.1

2. Audio Encoding was done by MPEG-1 Layer 3 more popularly known as MP3

3. The default resolution is 352×272 and should be judged at that resolution

Further information can also be seen from the screenshot

PS: you can also use other codec information utils like AVI codec etc.

Following are the screenshots taken from the video with different filters(All the left screenshots are Divx and right ones are Xvid)
Divx Filter with Sherlock Holmes

Xvid filter with Sherlock Holmes

Screenshot with Divx Filter

Screenshot with Xvid Filter

Click the pics and view them at full-size and compare them. As the videos clearly depicted the Divx Filter far outperforms Xvid Filter in visual apeal.

There is a very important decoder that I have not mentioned here, that is the FFmpeg decoder(default decoder in most open players like mplayer, media player classic and vlc). This decoder generally uses libavcodec to decode the media files and even though the performance is not as good as the above decoders but it is good enough but the difference in CPU usage is extreme. Ex for a normal movie where Divx and Xvid take as much as 25-30% CPU usage, FFmpeg will take barely 10-15%. So if its quality you are looking for then look no further than Divx but for performance nobody matches FFmpeg

So overall the results are quite ambigous. If you are looking for quality then Divx, for free encoding/decoding Xvid and for decoding performance FFmpeg . You can download the filters from here, and you can learn how to switch your filter from here. So what are you going to do with your player???

MP4/ HD video codec performance review: CoreAVC vs FFmpeg which is better ?

MPEG-4 is used for AV data for web (streaming media) and CD distribution, voice (telephone, videophone) and broadcast television applications.MPEG-4 adds new features such as (extended) VRML support for 3D rendering, object-oriented composite files (including audio, video and VRML objects), support for externally-specified Digital Rights Management and various types of interactivity. AAC (Advanced Audio Codec) was standardized as an adjunct to MPEG-2 (as Part 7) before MPEG-4 was issued.The key parts to be aware of are MPEG-4 part 2 (MPEG-4 SP/ASP, used by codecs such as DivX, Xvid, Nero Digital and 3ivx and by Quicktime 6) and MPEG-4 part 10 (MPEG-4 AVC/H.264, used by the x264 codec, by Nero Digital AVC, by Quicktime 7, and by next-gen DVD formats like HD DVD and Blu-ray Disc).

If you were looking for MPEG 4 Part 2/ AVI video codec performance review: Divx vs Xvid then click here

Codecs:

As told previously 2 parts of MPEG-4 format are popular. There are several codecs available that implement  MPEG-4 with different level of perfection, popular ones are Quicktime(Part 10), FFmpeg(Part 10), Xvid(Part 2), Divx(Part 2) and CoreAVC(Part 10). After consulting several forums, I found 2 of those very talked about CoreAVC and FFmpeg. I have compared the 2 codecs below.

FFmpeg
FFmpeg Logo
Whether you are a codec geek or not, this is the most probable codec that you have been using since time  immemorial.  Its free, opensource, can allow a lot of tweaking, plays a major number of containers and codecs. I will try not to go beyond H.264 in this article but one thing is for sure, whether you are using CoreAVC or not, you definitely have to use FFmpeg due to its overwhelmingly large number of codec compatibility.

CoreAVC
CoreAVC or Corecodec Logo
Its a closed source alternative for decoding H.264, infact its one of the accepted formats for Blue-ray. When it came out it astonished everyone with the sheer speed and performance outputs. Its known to exceed several hardware implementations (reminds me of John Carmack implementation of square root in Quake 3 which was faster than FPU). Corecodec people call it the fastest codec on earth.(and I am nobody to argue).

Left one is CoreAVC and right is FFmpeg

x264 encoded movie Sunshine decoded with CoreAVC

x264 encoded movie Sunshine decoded with CoreAVC

x264 encoded movie Sunshine decoded with CoreAVC

x264 encoded movie Sunshine decoded with CoreAVC
x264 encoded movie Sunshine decoded with CoreAVC

x264 encoded movie Sunshine decoded with CoreAVC

Doesnt take a rocket scientist to notice that better one. However when I discussed this on outside forums, I faced considerable shouting and screaming so I did more tweaking and testing and found another bit of information that convinced me to use CoreAVC for good.
Left is CoreAVC and right is FFmpeg.
Top is W/O post processing and bottom is with Post Processing

CPU usage between FFmpeg and CoreAVC
Cpu Cycles with CoreAVC
CPU cycles consumption when using FFmpeg codec
Difference between CPU usage when using FFmpeg and CoreAVC is huge. It can be seen from the above CPU usage or by seeing CPU cycles consumption . Considering that I have a Core 2 duo system with 1.8Ghz this difference is a substantial one, infact with preprocessing the player hung at 100% cpu usage in FFmpeg.

I found out to my amazement that Corecodecs are actually the fastest codecs in the world. I used all possible codecs at my disposal(VLC, ffmpeg, quicktime, windows media player and Nero) but CoreAVC not only gave better performance compared to all of them but also astonishingly smooth one at that. I ran it on 1080p trailer of 10000BC on MPC using quicktime alternative(had to rename .mov to .hdmov to use coreAVC) and performance was mind boggling.
Bottomline, I’d suggest you one thing. just go right now to Final build site and download the coreavc format and start using it. You can find out how to do that here.