HDV on the Big Screen

[Erkki Halkka]

Its hard to imagine how one could be blockier than the other. I don't know how you produced the photo you did but I made mine with Photo Impact working in 24-bit (total) and mine showed no blockiness at 4:1:1 compared to the same original photo at 4:2:2

 

I could send you audio test files with 8kHz and 44kHz sampling, where you probably couldn't hear the difference (say, a bass track without much treble in it). That doesn't mean there are no differences in other cases.

 

As said many times before, the difference doesn't show much in all pictures. On yours, the red logo on the hat is the best candidate for seeing color sampling errors - and it in fact does look a tiny bit mushier on the 4:1:1 image. If you zoom into the logo and A/B compare the images, you can see the difference. But it's very, very small in this case.

 

Try making red text over black background, at 4:4:4, 4:2:2 and 4:1:1. Or look at the examples i posted earlier, to see the color sampling differences better.

 

That red hat artifacting shot i posted is a direct .png screen grab from PAL MiniDV footage i shot on vacation, with the enlarged portion done in photoshop. It's actually 4:2:0 - but 4:1:1 should look rather similar.

 

All the color sampling examples with text were made by creating a title in SonyVegas, saving that to various formats (Uncomp, DV, DVCPro50, HDV) and then taking a direct .png screen grab of those.

 

PS, I don't know if you're intested but Adam Wilt had some info on his website about working with green-screen effects in HDV and 4:1:1 here: Chroma Keying

 

The heading of the page is "Steppy Edges" - more or less exactly the same thing as the "blocky colors" we've been talking about...

 

He's trying to get rid of it by filtering - i use more or less the same methods if i need to reduce color sampling errors when doing keying work, or mastering in higer sampling formats with DV/HDV footage. Instead of just blurring the color channels, i prefer median filtering - it gives slightly sharper edges.

 

As said before, on most raw footage, the difference isn't big enough to be worried about - if it was, low color sampling wouldn't be used. The problem raises it's ugly head with original footage IN SOME CASES, and quite a lot more often, if the footage is manipulated a lot (color correcting individual colors, keying etc.).

 

I too have a page about keying compressed / low color sampling video, have a looksee:

 

http://www.kolumbus.fi/erkki.halkka/HDVKey...and_keying.html

Edited December 17, 2005 by Eki Halkka

[Erkki Halkka]

I also never understood why the more high-end hi-def systems that record an mpeg signal do it at 4:2:0 but the HDV spec is 4:1:1. Again, I may be wrong but that's my understanding of the HDV spec. Eki can correct me if I'm wrong.

 

As far as i know, HDV is 4:2:0. I've got a web page on the differences of HD formats too - here:

 

http://www.kolumbus.fi/erkki.halkka/HDform...HD_formats.html

[Brian Wells]

I'll just go ahead and post the pics from the test outside of Flash where even Twinkle Toes can put it under a microscope and possibly even manage to see a difference there.

Please, take a look at the samples I'm offering. The file sizes are huge (1MB per frame) but you can clearly see the difference of 4:1:1 and 4:2:2. I created the samples in Final Cut Pro using the included DV25 and DV50 Codec's and exported them to uncompressed tif's for viewing.

 

http://homepage.mac.com/thisiswells/DV25.tif

http://homepage.mac.com/thisiswells/DV50.tif

 

*Note: I tried uploading these files to this site and it says, "Upload Failed. You are not permitted to upload a file with that file extention." Just a friendly FYI...

[Max Jacoby]

Pretty big difference indeed.

[Stephen Williams]

Please, take a look at the samples I'm offering. The file sizes are huge (1MB per frame) but you can clearly see the difference of 4:1:1 and 4:2:2. I created the samples in Final Cut Pro using the included DV25 and DV50 Codec's and exported them to uncompressed tif's for viewing.

 

http://homepage.mac.com/thisiswells/DV25.tif

http://homepage.mac.com/thisiswells/DV50.tif

 

*Note: I tried uploading these files to this site and it says, "Upload Failed. You are not permitted to upload a file with that file extention." Just a friendly FYI...

 

Hi Brian,

 

Thank you for the examples. I think it's very easy to see the difference, however some of the posters here today will not agree with me!

 

Cheers,

 

Stephen

[Guest Charlie Seper]

Fast motion can catch codecs out pretty easily, especially on the cheap-and-quick hardware which must by definition be in things like the Z1, making hard edges and ninety-degree corners visible in the video.

 

That's especially evident on big screen projection, and its seems much more noticable if the camera moves (as in hand held video) rather than if something moves in front of the camera. But either way compression artifacts really aren't a big problem unless there's some quick motion or a lot of post correction work is done. I said a long time ago (not in this thread) that, if you want to project a miniDV production, something like, My Dinner With Andre, would be the best way to go, where almost everything is done indoors with closer/mid shots and very little movement. An XL2 could pull that off quite well.

 

If you only own a camera that is capable of 4:1:1, then how did you create this comparison of 4:2:2 and 4:1:1.

 

I said that I used a still camera for it. It was from an inexpensive 2-megapx Olympus that shoots native 4:2:2 uncompressed jpegs at 72-dpi. Well, technically all jpegs are compressed but it was at 100% compression. I opened it in Photo Impact and saved a copy in 4:1:1 also at 100% compression, plus another copy at 4:1:1 with 75% compression.

 

As said many times before, the difference doesn't show much in all pictures. On yours, the red logo on the hat is the best candidate for seeing color sampling errors - and it in fact does look a tiny bit mushier on the 4:1:1 image. If you zoom into the logo and A/B compare the images, you can see the difference.

 

Yeah, I zoomed in on that logo too but to be honest, even at 1200% magnification I couldn't see any difference. You could obviously see the difference in the 75% compressed one though. But if you have to zoom in 800% or more to see a difference, as far as I'm concerned its not worth thinking about. Its like dust on top of the fridge: Our wives, mothers, and girlfriends may not agree but my working relationship with cleaning has always been that, if I can't see it--it isn't there. ;)

 

As far as i know, HDV is 4:2:0. I've got a web page on the differences of HD formats too

 

That's why I put the ad-hoc disclaimer on that post. I have a magazine from last year that said the HDV spec was 4:1:1 and I didn't trust it because they've been wrong on several issues before.

 

Please, take a look at the samples I'm offering. The file sizes are huge (1MB per frame) but you can clearly see the difference of 4:1:1 and 4:2:2. I created the samples in Final Cut Pro using the included DV25 and DV50 Codec's and exported them to uncompressed tif's for viewing.

 

Well sure, one was a 50mbps mock-up and the other 25mbps. Of course you've got twice as much info in one. No one's debating that. I said that 4:1:1 and 4:2:2 didn't look any different to me at 25mbps. And like I told Phil above, I don't know of any cams that shoot 4:2:2 at 25mbps. Maybe there is one but I'm not aware of it. But I can't imagine that it would look very different, if at all, given the same amount of info in the data stream and both using the same DV codec. Now if one used DV and the other Mpeg or something else, then that might be worth taking a closer look at.

 

Another thing to consider is that in my photo test I used a photo from a still camera and I can't really see any difference at all when dropping it to 4:1:1. Everything was calm outside and the guy in the pic was motionless. Like Phil said above here, compression really gives itself away in motion. The wmv codec is a great example. It looks outstanding until you put some motion in it. Maybe Eki could test this sometime since he works with HDV now and then and uses green screen effects sometimes too but, I would guess that if he put someone in front of a green screen that was motionless and the images(s) composited behind him was motionless as well, that the effect could be pulled off with very few artifacts showing up even working with HDV.

 

Hey, no one's debating that 35mm doesn't have everything in the world going for it. The debate is whether or not anyone could tell that it looks different from HD or even MiniDv and when or how. I say that on standard def TV they look the same almost every time. On the big screen the differences are mostly just one of resolution, which in turn depends a lot on the size of the screen, the type and condition of the projector, and Eki brought something else into the frey with the green screen effects problems he's had with HDV (or I assume any 4:1:1 format at 25mbps or under). And I will admit there can be some motion artifacts in highly compressed video. But I still say that video looks exactly like film if shot right, and motion is kept to reasonable limits (you're obviously not going to shoot The Matrix on DV). The difference on the big screen is almost entirely one of resolution. November looked really good, not great, but quite well done. And part of that was because it was shot right. 90% of it was done indoors with close to mid shots and no ridiculously fast movements. Not more than 5-minutes (I'm guessing) was done with hand held shooting. They used a 35mm anamorphic lens (if what I've read was true), etc. And the color was perfectly fine. Now with HDV we get a bigger, clearer frame and might give up a bit of color in the process. I must say that even at the 8000 data rate which DVD Mpeg2 codecs like TMPGEnc can generate, I can barely see any difference between my original miniDV footage and the mpeg2 copy. At the 18000 to 25000 data rates HDV generates there ought not to be any differences at all that anyone will be able to see. It ought to look quite a bit better on the big screen than November did. And November looked quite good as it is.

 

What bothers me is that there are some people in this industry (and probably in this forum) who have a lot to lose when film sales go through the floor. They'd like you to think that anything less than super16 isn't worth using and that video has a certain "look" to it that is nothing like film. But I know perfectly well that video shot at 24p with the right gamma etc., and under the right lighting looks exactly like film. There's no reason that it shouldn't. Film isn't magic. There's nothing going on there that can't be digitized into 1(s) and 0(s) quit easily. The differences that matter mostly deal with resolution, plain and simple, unless you're shooting an action yarn for kids. HDV is already taking off like a San Diego fire and I think HDV at 24p will be around for a long time even though I have yet to see it projected. I also think that guys who now shoot on film and have never done any editing on an NLE will REALLY appreciate it once they get into it. Bring in Flash for its wonderful compositing, photo manipulation/movments and scalable vector drawing, SwishMax for it's photo and especially it's lettering effects, a good multitrack music editor with the ability to sync to a video track, and incredibly good mics, mixers, preamps and effects for audio recording into a DAW setup at rock bottom prices, and a guy can do almost anything he sets his mind to all by his lonesome at a reasonable price. The film guys ought to be worried....

Edited December 19, 2005 by Charlie Seper

[Brian Wells]

Charlie,

 

I, too, have given up on trying to prove my point to you.

[Guest Charlie Seper]

From Brian Wells:

 

What Codec did you use to create a 25Mbps 4:2:2 sample?

 

Like I said, I used an Olympus still camera for it. The argument wasn't about whether or not 4:2:2 looked different than 4:1:1 given different situations or codecs, but just in general. You could bring any 4:2:2 still into a good graphics proggie and output a copy at 4:1:1 and get the same results I did. I've never been able to tell the difference. Try it yourself. The argument is that 4:2:2 has more color available to it than 4:1:1 does. I say that given the same bitrate they won't look any different. At least they sure look the same to me.

 

Just how many times are you going to change your message kid?

Edited December 19, 2005 by Charlie Seper

[Brian Wells]

say that given the same bitrate they won't look any different.

Unfortunately, that isn't an option in the video world. There aren't any 4:2:2 Codec's that are also 25Mbps.

[Guest Charlie Seper]

Unfortunately, that isn't an option in the video world. There aren't any 4:2:2 Codec's that are also 25Mbps.

 

Which I also said...twice.

 

However, I belive there may be cams that shoot both 4:2:2 and 4:1:1 at 50mbps. But it doesn't matter. All things being equal, I don't see 4:2:2 looking any different from 4:1:1. Like I said, bring any 4:2:2 pic into a graphics proggie and output it at 4:1:1 using the same compression (or none) and see for yourself.

[Brian Wells]

I don't see 4:2:2 looking any different from 4:1:1... bring any 4:2:2 pic into a graphics proggie and output it at 4:1:1 using the same compression

Here is an Uncompressed 4:2:2 version of the same sample I posted earlier if you'd like to try it...

http://homepage.mac.com/thisiswells/10-bit.tif

 

All others... Get your popcorn, candy bars, and over-sized Coca-Cola!

[Tim Tyler]

Like I said, bring any 4:2:2 pic into a graphics proggie and output it at 4:1:1 using the same compression (or none) and see for yourself.

 

Please correct me if I'm wrong, but...

 

Unfortunately, this is not a valid way to test or prove your theory.

 

The test files you post on the 16th are 1.35MB (megaBYTES) each. HDV allows 25Mbs (mega BITS per second). That means 60i HDV allows each frame to be about 53KB.

 

Post some 1440x1080 53KB files in 4.1.1 and 4.2.2 and see how they compare.

[Guest Charlie Seper]

I'll try and do that Tim. I've got to be away for a few days with work though. Maybe by Friday or Saturday.

 

That's a point well taken though. I don't know a way to shoot 4:2:2 at 25mbps with an actual camcorder though so the best I could do would be to bring a still pic into Photo Impact and resize it to 720 X 540 at 72-dpi and choose to "compress by size" and output it at both 4:2:2 and 4:1:1 at a size of 53k. Are you sure about that being 53k though? I'll run it through a calculator later but that seems too low a file size off the top of my head.

 

Later....

[Stephen Williams]

I'll try and do that Tim. I've got to be away for a few days with work though. Maybe by Friday or Saturday.

 

That's a point well taken though. I don't know a way to shoot 4:2:2 at 25mbps with an actual camcorder though so the best I could do would be to bring a still pic into Photo Impact and resize it to 720 X 540 at 72-dpi and choose to "compress by size" and output it at both 4:2:2 and 4:1:1 at a size of 53k. Are you sure about that being 53k though? I'll run it through a calculator later but that seems too low a file size off the top of my head.

 

Later....

 

Hi,

 

Thats part of the problem, a very small data rate.

 

Stephen