Why 4:2:0 ?

[naveenmedia]

I feel your pain. It's gotten so whacked it's as if DVXuser needs it's own specific forum where people can compare/argue/bicker and anything else about tech talk on cameras ad infinitum, and leave the rest of the site to more meaningful discussion.

Hey, I didn't mean to start an argument, I was just asking questions as someone who's not familiar with the technical issues, particularly on the acquisition end of things. No harm intended!

I got into video through the back-end: education in graphic/web design, job with production company doing graphics for post, moved into editing, then got focused on workflow issues that led me to acquisition (i.e. shooting).

So I have a lot of basic tech questions and I appreciate all the help from those here at DVXUser who seem so willing to share! This site is a gem of a resource, and much better than some other forums, where people are often non-responsive or cocky know-it-all's.

My thanks to Barry Green, Stevet, Philip Bloom, puredrifting (Dan), MattinSTL, and any others I know I'm forgetting for the massive amount of good info I've gotten from this site in the last few weeks alone. I've tried to contribute as much as I've been able, and hope to offer more in the future.

And to the specifics of this debate, different cameras will be better than each other in different areas, some will be just "better" overall, the only way to know is to actually try the gear. I wholeheartedly agree!

But it helps to at least have a handle on basic knowledge of what's better, at least on paper, to help evaluate any new product and to get the best results, which is why I asked.

Naveen

 

[Barry_Green]

I feel your pain. It's gotten so whacked it's as if DVXuser needs it's own specific forum where people can compare/argue/bicker and anything else about tech talk on cameras ad infinitum, and leave the rest of the site to more meaningful discussion.

Agreed. Problem is, the tech talk is like 80% of the discussion here, and "the rest of the site" (excluding the fun nonsense in the cafe) is pretty minimal. I'd love to see more participation in the screenwriting and directing and cinematography forums, but those sections don't get nearly as much traffic.

 

[Barry_Green]

Hey, I didn't mean to start an argument, I was just asking questions as someone who's not familiar with the technical issues, particularly on the acquisition end of things. No harm intended!

There is no harm done -- our site excels at explaining the issues, and you asked a specific issue (and this is kind of why I've taken a hard line in this discussion on trying to keep the issue centered on what you asked, instead of letting it get derailed into some manufacturer vs. manufacturer debate).

A simple understanding of 4:2:0, or long-GOP, or rolling shutter, or CCD smear, or backfocus, or any of the other myriad of technical issues we face, is what many of our users need and search for. And we're glad to help explain these items. Your question was entirely appropriate for this forum.

While it'd be nice if we got more participation in the art & craft sections, a large part of this forum's value is in our helping people to understand those things that they don't currently understand.

for the massive amount of good info I've gotten from this site in the last few weeks alone. I've tried to contribute as much as I've been able, and hope to offer more in the future.

That's exactly what we're about here. We've had many who started as relative newbies and who have shared much of what they've learned. We like that.

 

[Stevet]

Agreed!

"Can't we all just get along". I'd bet if we all got together over a few beers, there would be FAR less debate and more creative discussions. I really believe most of this stems from having to write your thoughts in a couple sentences. It really makes it hard to convey we all just want the same dang thing.

Also, which has been brought up in this thread and a million others, it's really how we use these tools. All of these cameras ar very capable of producing great images.

Having something worthy shooting is 80% of it. I could take a 200K camera into my kitchen, shoot a quick round. I'm sure I would be very uninterested in the footage. Yep, there's my kitchen.... boring....

We all certainly agree on having decent images for composition.

 

[nyvz]

So wait, now wouldnt the HVX have a maximum theoretical useable color resolution of around 640x540 since one cannot get more than 540 lines of vertical res in one color channel from a 960x540 single color ccd even if the color channel is recorded as 640x1080? and wouldnt this resolution, where the ccd resolution is the limiting factor vertically and codec is limiting horizontally, be less resolution than the ex1's 960x540 color resolution which is a resolution limited by the 4:2:0 sampling in the codec? furthermore, wouldnt the the 4:2:0 of the ex1 be superior even further since 2:1 scaling (1920x1080 chip to 960x540 output) is less likely to create artifacts and soften the image than 1.5:1 horizontal scaling (960x540 chip to 640x540 usable output resolution). not to mention a 960x540 codec is more likely able to represent 960x540 resolution than a 960x540 ccd is able to capture 960x540 actual resolution. does that make sense? in this way, would it be reasonable to say that despite the alure of 4:2:2 dvcprohd in the hvx, the ex1 is more likely to produce color channels with more useable resolution in most situations?

 

[David Jimerson]

So wait, now wouldnt the HVX have a maximum theoretical useable color resolution of around 640x540 since one cannot get more than 540 lines of vertical res in one color channel from a 960x540 single color ccd even if the color channel is recorded as 640x1080? and wouldnt this resolution, where the ccd resolution is the limiting factor vertically and codec is limiting horizontally, be less resolution than the ex1's 960x540 color resolution which is a resolution limited by the 4:2:0 sampling in the codec? furthermore, wouldnt the the 4:2:0 of the ex1 be superior even further since 2:1 scaling (1920x1080 chip to 960x540 output) is less likely to create artifacts and soften the image than 1.5:1 horizontal scaling (960x540 chip to 640x540 usable output resolution). not to mention a 960x540 codec is more likely able to represent 960x540 resolution than a 960x540 ccd is able to capture 960x540 actual resolution. does that make sense? in this way, would it be reasonable to say that despite the alure of 4:2:2 dvcprohd in the hvx, the ex1 is more likely to produce color channels with more useable resolution in most situations?

Whatever the math, this misses the point of the thread.

Imagine the EX1 with 4:2:2 instead of 4:2:0. It would have twice the color info that it currently has. What the HVX or DVCPRO does doesn't matter to that point.

 

[harddrive]

But you do realize you're bringing in an argument that had nothing to do with the original question, right?

No - because whilst you may understand the wider issues, many don't, and think ANY 4:2:2 system is better than ANY 4:2:0 system. That's not true, and bringing in the wider differences, and using DVCProHD resolutions as an example to illustrate the point, I would call valid.

As an absolute answer: 4:2:2 is better than 4:2:0. Unquestionably.

With the qualifier "like for like" I agree. As a blanket statement, it could be misleading.

And I reject the classification anyway because the HVX is undoubtedly sharper in 1080 mode than it is in 720, so trying to characterize it as less really just has no place in this discussion.

Err, I don't think I said anything of the sort, look again before dismissing it.

I actually said: "Use the EX at 720 and there is considerable detail loss relative to 1080 - with the others the loss is much less, certainly in the vertical direction." which clearly implies my belief that the HVX200/500 are sharper in 1080 mode than 720. How much of this is due to 720 DVCProHD downsampling to 960 I don't know - it would be interesting to see the difference between the HVX 1080 mode and a 720 mode recording the full 1280x720 raster, rather than 960x720.

But this isn't a number crunching argument anyway. We've seen very good examples of 1080 images from the HVX, the EX and others, and the EX was unarguably much sharper. The conclusion I draw is that the overall luminance resolution is way more important than any 4:2:2, 4:2:0 arguments.

-- look at DVCPRO50 vs. DV. Read some of the reports, terms like "amazing" do get used.

I believe you are referring to reports within the US. I don't disagree with the conclusions - but they are therefore comparing 4:2:2 with 4:1:1, not 4:2:0, and there's a world of difference. DVCpro50 also has much milder overall compression, so the 'amazing' difference isn't solely down to colour space anyway.

In 50Hz countries 4:1:1 only exists at 25Mbs within DVCPro, and that's far less common than DVCAM. The very widespread use of 2/3" DVCAM here for broadcast use (even within an interlace system) tends to make me feel that whilst noone should claim 4:2:0 as good as 4:2:2, the practical differences are less than sometimes made out.

 

[Barry_Green]

harddrive, I'm being patient but you're bordering on trolling with this. You seem so hell-bent on discrediting anything and everything Panasonic that you're flinging it into a thread that had nothing to do with Panasonic. Why not just give it a rest?

If the question had been "is Sony's 4:2:0 as good as Panasonic's 4:2:2 from the HVX" then maybe you'd have a leg to stand on. As it is, you're dredging up numbers in a vain attempt to discredit Panasonic's efforts, in a thread that had nothing to do with them.

So let's try again, and let's try to stay on track: is 4:2:2 better than 4:2:0? Yes. Absolutely, unquestionably, without a doubt. You will never, under any circumstances, ever find an uncompressed video system where the engineers say "y'know what? Let's just subsample to 4:2:0 'cause I think it looks better." That won't happen because the very notion is patently absurd. If you want to argue the particulars of tradeoffs involved in compression algorithms, that's a different discussion. If you further want to muddy the waters by bringing in long-GOP unfiltered vs. intraframe prefiltered, you're just confusing the issue.

For all you know, the original poster could have been talking about 4:2:2 from DigiBeta. Or MPEG-IMX. Or HDCAM-SR. Or BetaSX. Or, basically, any and every professional format -- they're almost all exclusively 4:2:2, only HDCAM differs (it's 3:1:1, which is still a darn sight better than 4:2:0).

So, back on topic: Sony could have made the EX1 4:2:2. It would have been better had it been 4:2:2. They didn't do it. We don't know why. End of story.

 

[ffaf07]

/thread.