Problem is compression

[santosh]

Dear Friends can you tall me how can i calculate video compression like this is 8bit and this is 10bit what is posses plz tall me how can i calculate and what is posses OK thank you

[tingsern]

I don't think anybody understood your question at all.

What video compression are you asking? There are 100's of compression algorithms in use today - MPEG1, MPEG2, MPEG4, CANOPUS HQ, DVCPRO HD, etc, etc.

In your original question - 8 bits and 10 bits have nothing to do with video compression. They are the number of bits per colour (one interpretation) - hence, in the RGB world, 8 bits per colour = 256 shades of RED, 256 shades of BLUE, 256 shades of GREEN. 256x256x256 = 16 million colours. 10 bits = 1024 shades of each colour.

Video on the other hand - don't use RGB as colour space.

I know I am not answering your question, but, I have problems understanding your question in the first place.

[antonsvideo]

8bit is for 90% of the world's video users

10bit is for George Lucas shooting Starwars on video (or other similar high end productions)

note: you will need ten (10) SCSI hard disks connected in Raid0 in order to work with 10bit uncompressed HD video in realtime successfully

check out Matrox Axio HD system requirements when working in 10bit mode

[tingsern]

Anton,

That's a good one - ha ha ha .... :-) Most likely scare him off ....

But, AVC-Intra (not AVCHD) can go 4:2:2 at 10bits at HD resolution at 100Mbits/sec (same as DVCPRO-HD today). Nobody does uncompressed video in the camera anyway. Some Panasonic shoulder mount cameras (HPX2000, HPX3000) today can do that already - recording to P2 cards. If you are referring to digital cinema (RED, for example) - that's a different story.

[STORMDAVE]

And the Viper ...and HDCAM SR. :)

Don't think anyone of us is using these cams...but 10 bit IS important...especially for motion graphics people. It's not only for "video".

[tingsern]

Especially for AE folks - 10 bits is important - less artifacts from filters.

I suppose if we are using RED, VIPER, etc ... we won't be in this forum anyway :-).

[santosh]

Dear Friends what my question you can,t understand see i give you example i capture one min footage form Canopus HD system in uncompressed(YUV) format it size was 1.16GB and i have capture uncompressed RGB it size was 2.23GB so i want to calculate what compression of uncompressed(YUV) and uncompressed(RGB)how can i calculate this compression ratio can you tall me this footage compression is 10bit ya 8bit if you know so tall me what is proses can i calculate it

[santosh]

Anton,

That's a good one - ha ha ha .... :-) Most likely scare him off ....

But, AVC-Intra (not AVCHD) can go 4:2:2 at 10bits at HD resolution at 100Mbits/sec (same as DVCPRO-HD today). Nobody does uncompressed video in the camera anyway. Some Panasonic shoulder mount cameras (HPX2000, HPX3000) today can do that already - recording to P2 cards. If you are referring to digital cinema (RED, for example) - that's a different story.

first think freinds 4:2:2 is not a 10bit it is 8bit 4:4:4 it is 10bit i thing

[antonsvideo]

there is no 10bit card or optional upgrade for any Edius related product

unless I am mistaken, which can happen

[STORMDAVE]

Yeah, EDIUS itself only supports 4:2:2 color space and upto 8bit. Any 10bit clip loaded on the timeline will downconvert in realtime to 8bit.

[santosh]

Yeah, EDIUS itself only supports 4:2:2 color space and upto 8bit. Any 10bit clip loaded on the timeline will downconvert in realtime to 8bit.

but HQ can do the upconvert it na 4:4:4 if yes so how

[santosh]

there is no 10bit card or optional upgrade for any Edius related product

unless I am mistaken, which can happen

i think edius HD is passable you can captuer 10bit if you captuer HD-SDI it captuer 1.5gbps 10bit

[antonsvideo]

i think edius HD is passable you can captuer 10bit if you captuer HD-SDI it captuer 1.5gbps 10bit

no, I really don't know any Edius Canopus product in the HD world that can do 10bit

once again, I maybe mistaken, maybe someone from GV can confirm

[GrassValley_KH]

Anton is correct - we are 8-bit only, for the moment.

[GrassValley_BH]

Don't confuse sampling and sample precision with data compression.
Sampling is a data reduction or quantization. It is independent of compression.
Precision of each sample is also a separate issue.

So first, how many samples are being taken, and second, what is the precision of each sample?

Sampling:
Sampling is sort-of related to resolution. You can have an image that is a set resolution, but there might not be discrete data for each pixel of that image.
See Adam Wilt's DV FAQ Pix on sampling and Charles Poynton's document on Chroma subsampling notation

Precision:
Precision is directly related to the data-size of each sample.
For example, a 8-bit BMP versus a 16-bit or 24-bit BMP, they are all uncompressed. But they have different precision on the sample size. An 8-bit BMP can show significantly fewer colors than a 24-bit BMP, and a 24-bit BMP (because it's 3 bytes per pixel) is 3 times larger than an 8-bit BMP (with only one byte per pixel).

FourCC.org lists common compression formats and sample sizes.