RGB24, YUYV, YUY2 ... What does it all mean?

[WarwickG]

Is there a one-stop shop of information that can explain (in lamen's terms) what each of these settings are meant to do for me when capturing analog video? I keep reading bits and pieces about the use of YUV-based or RGB-based options for different applications (like post-processing and filters etc), but it all still has me completely bewildered.

If anyone can jot down a quick summary of why I must choose one over the other, that'd help immensely.

btw, I'm also confused over the many elements that go into analog video (like luminance, chroma, etc) but I don't know what any of them really do and how I am meant tweak each one to improve on the video that I have digitised. Is there any info on this as well? Many thanks.

[thecoalman]

Try the glossary to the left, below What Is.

[fmctm1sw]

I keep reading bits and pieces about the use of YUV-based or RGB-based options for different applications (like post-processing and filters etc), but it all still has me completely bewildered.

If anyone can jot down a quick summary of why I must choose one over the other, that'd help immensely.

You may still be bewildered but... www.fourcc.org

[trevlac]

A challange! Quick and clear.... I'll try.


Color - A Color can be represented by a 3 component formula. RGB is an example. By 'mixing' Red Green & Blue, you can get all other colors.

Color Space - A color space is really the formula of 3 components used to represent colors. RGB is an example, YUV is a general example. CMY, is another example.

4CC - For windows video, the pixel colors are stored using a given colorspace formula, and an order of the 3 component numbers as they are saved in a file. RGB24 probably stores the red value, the green value, and the blue value in that order and using 24 bits. YUY2 & YUYV both use the YUV type colorspace. For this colorspace formula, there is generally 1 Y for each pixel and then 2 or more pixels share the U&V values. So if the values YUYV were stored in a file, they would cover 2 pixels, and usually this is done in 32 bits. I'm not sure of the difference between YUYV and YUY2. Programs that except 1 probably are ok with the other. YV12 is an example where 4 pixels share a given UV.

-------------------------------------
So knowing that .... Let's talk about YUV color in more detail.

In digital, YUV is properly called YCbCr. Y is the black and white info also known as the luma. Our eyes are more sensitive to black and white, so YCbCr takes advantage of this and throws out 1/2 the CbCr to save space. Cb & Cr are known as color difference signals or chroma or the color signal. Cb is a blue difference signal (it's relationship to RGB is Y-B), Cr is the red difference. These are not as easy to visualize as RGB mixing. But, you can work back and forth between RGB and YCbCr with some simple math.

Some practical info ....

Most video signals we work with, start as an analog form of YCbCr. YIQ for NTSC and YUV for PAL. This is because to save on transmission bandwidth, they use this weird colorspace. So, using a YUV type colorspace is ok. In fact, because of the built in compression of the chroma, it makes for smaller files and faster processing.

Tweaking the luma usually means making the picture sharper. Tweaking the chroma usually means making the colors more vibrant. It is easy to overdo both of these.

I'll stop there ....