GH2 posterization, banding, gradients

nickolaos

I have a suspicion that is due to the fact Panasonic implement route CAVLC instead of CABAC on AVC codec. http://en.wikipedia.org/wiki/Context-adaptive_binary_arithmetic_coding , http://en.wikipedia.org/wiki/Context-adaptive_variable-length_coding Solutions high ISO settings that introduce noise in gradients, ND filters to compensate at day light, high f number but the negative, in this, is wider DOF, but then again with high ISO the DR drops considerably.

Wises 1- CABAC implementation 2- 4:2:2 color subsampling 3- Unlock SH BURST RATE more than 1sec and 40frames 4- Raw video mode

Vitaliy_Kiselev

Please just understand one thing (valid for both coding techniques).

It is a lossless compression technique.

So thing you wrote is just bullshit, sorry.

nickolaos

Well maybe, but I have example of MTS 8bit 4:2:0 17mbit CABAC file without banding. I record side by side to compare them.

Vitaliy_Kiselev

@nickolaos

It is not maybe. It is loseless compression techniques.

driftwood

VK is right they are just coding techniques. For example GH3 uses both CABAC and CAVLC techniques for the different recording modes. CAVLC is being employed by All-I 24p/25p. 1080p IPB is using CABAC.

nickolaos

http://books.google.gr/books?id=RG5YUOAzRsEC&printsec=frontcover#v=onepage&q&f=false and http://www.google.gr/url?sa=t&rct=j&q=encoding%20h264%20-%20smeurope%20-%20short&source=web&cd=2&ved=0CCsQFjAB&url=http%3A%2F%2Fstreaminglearningcenter.com%2Fattachments%2Fdownload%2F43%2FEncoding%2520H264%2520-%2520SMEurope%2520-%2520short.pdf&ei=vki2UJ6SHMbSsgaHp4DIDg&usg=AFQjCNEMCd3hG2EQfio5LJrO3s60dqUkZg&cad=rja Page 10

driftwood

All you have to know is:-

CAVLC offers lower compression produces bigger filesizes but doesn't need as much processing power to decode. CABAC offers higher compression and yields smaller filesizes but needs more processing to decode.

nickolaos

http://en.wikipedia.org/wiki/High_Efficiency_Video_Coding Entropy coding

nickolaos

Analysing with Elecard Streameye and Codecvista ... will post results

nickolaos

Please take a look this sample and the codec efficiency I can only dream what if this codec was possible to implemented on a GH2 LOL http://x264.nl/x264/10bit_01-ateme_pierre_larbier_422_10-bit.pdf http://x264.nl/x264/10bit_02-ateme-why_does_10bit_save_bandwidth.pdf http://x264.nl/x264/10bit_03-422_10_bit_pristine_video_quality.pdf

Sample http://s.greenqloud.com/forpitch/%5BForPitch%5D%20Usagi%20Drop%20-%2003%20%28640x360%20h264%20Hi444PP%20OGG%29%20%5B264C0DEC%5D.mkv

nickolaos

Video #1
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High 4:4:4 Predictive@L5.0
Format settings, CABAC : Yes
Format settings, ReFrames : 5 frames
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 4s 295ms
Bit rate : 37.5 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate mode : Constant
Frame rate : 23.976 fps
Color space : YUV
Chroma subsampling : 4:4:4
Bit depth : 10 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.754
Stream size : 19.6 MiB (100%)
Writing library : x264 core 129 r2230+696+35 805d074 tMod [10-bit@all X86_64]

Upsampled GH2 ...must figure out a way for NLE (Remind me of 2007 AVCHD files)

nickolaos

http://rapidshare.com/files/3828302090/00022.mp4

It is a lossless compression technique.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters