4:2:2 vs 4:2:0

I

interelectronix

Active member
Is there any real frame by frame comparison between 4:2:2 and 4:2:0. Especially in cc.

I read the comparison article between the XH A1 and the HVX200. In the comparison the colors of the XH A1 where not impressive. A lot of comments said: " but you can do cc and then there will be no difference"

I have limited expirience because I don´t own anything with HDV printed on. Can anyone bring light to my underexposed thoughts.
 
interelectronix said:
A lot of comments said: " but you can do cc and then there will be no difference"
Click to expand...

Well, but that CC will be inside the limitations of a 4:2:0 color space (eg same as DV Pal...)
DVCproHD offers lower compression and more color information compared to HDV.
But there has been some nice footage shown here, filmed in HDV.
Nevertheless, I still prefer DVCproHD.
 
4:2:2 at full native 1080 is better than HDV at full 1080 for CC.

I wont go into particulars but I will say not all is equal in the summation.
 
The way you can look at it is that even though the HVX uses spatial offset to record a 1080P HD picture which may mean that it has slightly less resolution than the Canon HDV, the fact that the colour information is twice that of a HDV picture makes it heads and tails above for me.

Especially with regards to things such as chromakeying. I mean you try pulling a simple clean matte from DV footage and that gives you an idea about colourspace. As much as you have extra resolution in HDV over DV, the colourspace is the same. Thats not to say you can't pull a matte from HDV, but it will require more work.

If you have only 1 colour per 4 pixels in 4:2:0 [which means colour resolution is a quarter of the screen resolution] in the HVX you get 1 colour per 2 pixels in 4:2:2 so you get half the colour resolution of the screen resolution.

I would love to see a side by side comparisson anyway !

Thanks
Richard
 
Last edited:
One thing you could do to see the difference between the different color sampling is to insert a TIFF or other low-compression/uncompressed file from a DSLR into your NLE, and render that out in both DVCProHD and HDV format.

Open the resulting videos, and start pixel-peeping. This doesn't get into artifacts that may result from motion or complex scenes, but it will let you see how the two codecs handle fine detail.
 
4:2:2 has more color information, but it's important to note that this seldom translated into noticeable color difference in the raw image. The A1's color may seem muted to you, but this is due to the way the color matrix is programmed and not because of the 4:2:0 color space. Different cameras render color differently, and a Sony 4:2:0 camera is capable of giving extremely vibrant color on the same color space as the A1. The difference between a compressed codec like HDV and a less compressed one like DVCPRO HD is in the amount of data. Visually there is no difference, but it's when you start doing CC that the extra information from DVCPROHD comes in handy. Otherwise looking at the footage visually you'll be hard pressed to tell which is 4:2:2 and which is 4:2:0.

One area in which 4:2:2 holds a definite and clear visual advantage is in reduced "tearing of the reds" when pure red shows up in the frame. With DV 4:1:1 or 4:2:0, you can clearly see that the red areas will appear "torn" because the color intensity overwhelmed the codec. With 4:2:2 this is much less likely to happen.
 
I have made some tests and magnified some shots. I will post some pictures by tomorrow.

In brief:

In most cases it is more than obvious that 4:2:2 is far superior. And one very interesting further insight - the grainy look of the HVX is more tied to the codec then to the camera.

Nearly 2 a.m - I need some sleep.
 
THoff said:
One thing you could do to see the difference between the different color sampling is to insert a TIFF or other low-compression/uncompressed file from a DSLR into your NLE, and render that out in both DVCProHD and HDV format.

Open the resulting videos, and start pixel-peeping. This doesn't get into artifacts that may result from motion or complex scenes, but it will let you see how the two codecs handle fine detail.
Click to expand...

Thats very true if you started out with say footage only from the HVX and then compared the source footage with some converted to HDV (not entirely accurate but a roundabout comparison).

4:2:2 colour space in a HVX is not like 4:2:2 colour space in say a higher end HPX 3000 or the like. So it cant be argued purely on that point alone without looking at the other factors involved. And to do so IMO is misleading.

BUT, how much are you paying for the HVX in comparison to those higher end cameras? A hell of a lot less, and for good reason - its not as capable and if it were no one would buy anything else.

But for the price its as close as you can get. Now to the Canon - took a different route. No variable frame recording - the same old HDV tape drive = cheaper camera again. The canon did go for chips with more physical pixels. It still uses pixel shift but a lot less than the HVX.

If your going to argue colour space here: scientific results would probably put the HVX ahead for cc and chroma key - but really by such a small amount that the results become insignificant. You want better results you need to spend more, but even so I think you would find either of these could give you an acceptable result, especially if you plan the shot and take care to get it right.
 
You must log in or register to reply here.
Back
Top