is also known as a keyframe... this frame contains a full frame.. or an entire picture.
P-Frames(Predicted Frames) and B-Frames(Bi-Directional Frames) draw on the information from an I-Frame to make up parts of their own frame.
4:2:2 just refers to the colour space of the codec, It doesn't have much to do with the I-Frame... (although I am sure algorithm wise, it has a lot to do.. but thats a whole other can of worms..)
Can anyone give me a reference on the meaning of the three numbers separated by colons are (e.g., 4:2:2). I see this used all the time but cannot find a specific reference on what the meaning of the numbers is? Thanks much.
The numbers stand for the components in a component video signal: Y, Cr, and Cb. They refer to how frequently those elements are sampled per scan line, per group of four pixels.
4:1:1 has one of each color sample for every group of 4 luminance samples, and that happens on every scan line.
4:2:0 has two color samples of one of the components on each scan lines per group of four luminance samples, but *no* color samples of that component on the odd scan lines. So on an even 4:2:0 scan line you'll find four luma (Y) samples, and two Cr samples, but no Cb samples. On an odd 4:2:0 scan line you'll find four Y samples, no Cr samples, but two Cb samples.
Both methods have 2 color samples per 4x2 block of pixels, just in a different pattern. 4:2:0 is used by PAL DV and by MPEG-2 and HDV, 4:1:1 is used by NTSC DV.
Both methods have 2 color samples per 4x2 block of pixels, just in a different pattern. 4:2:0 is used by PAL DV and by MPEG-2 and HDV, 4:1:1 is used by NTSC DV.
How "badly" is the colour samples effected when converting PAL DV to/from NTSC DV with software?
There's theory, and then there's reality. I suggest if you want to see, download the demo version of DVFilm's Atlantis software, and then you can see what the reality is.
Thanks all for the answers. I can't tell you how useful I find this site. I really appreciate "those in the know" taking time out to answer the newbie questions.
Thanks again Barry and Loki for your replies. Now that I sort of understand the Luma and color sampling designations, this brings another question to mind. Suppose you recorded video in DVCPRO50 format (4:2:2). This is presumably "better" because of the higher frequency sampling of Cr and Cb. Can this be edited on a PC using Vegas just like NTSC DV (4:1:1)? And then, when the video is burned to dvd MPEG2 (4:2:0 ??), is there a problem with rendering the 4:2:2 format as opposed to the usual NTSC DV 4:1:1 format? Maybe another way to pose this question: Is the "better" video quality of DVCPRO50 lost if you ultimately edit and then burn to dvd?
I don't believe Vegas currently supports the DVCPRO50 codec, but maybe it does. I'm sure other editors (like FCP) do.
When burned to DVD, it's likely that you'll get a better-looking final product from 4:2:2 down to 4:2:0 than you would from 4:1:1 down to 4:2:0, but that's just speculation, I'd like to run some real-world results before declaring that it would be so.