HDV on the Big Screen

[Erkki Halkka]

Shoot 35mm film as DOP: No.

 

...it might also be interesting to know how much experience with HDV the participants have.

 

So, how many of us here has shot HDV?

Done post on it?

Worked on HDV to film project(s)?

Seen HDV to 35mm print footage in a theater?

 

"Yes" to all accounts here ;-)

[Phil Rhodes]

Hi,

 

No, Max, I'd quite like to find out who you've persuaded to pay to to shoot your personal glory projects on 35 - mainly in the sincere hope that I'm not inadvertently contributing myself!

 

Although given the way the EU works, that's entirely likely, I guess...

 

Phil

[Guest Charlie Seper]

16mm looks good projected onto a 10ft screen, not so hot on a 30ft wide screen.

 

Well, as somebody who's watched movies for 46-years and still has good far sightedness, I can only say that I've seen great stuff on a 30' screen shot on 16mm, especially super16.

 

Why do musicians still use valves in their guitar amps? Answer is gives a warmer sound.

 

I can duplicate any degree of muddiness, i.e.--warmth--with solid state compontents. Boosting lower mids is hardly rocket science. Some guys like the sound of distorted tubes. I normally don't. Actually, I don't think too many people do, they just think they do. There's a lot of guys my age who grew-up learning to play in the 60's/70's that will tell you that what people "think" they like about distorted tubes in those old amps is really the sound of distorted speakers. I can take any old amp that you think you like the distrotion characteristics of and replace the speakers with an ultra clean sounding one, like a typical JBL, and produce a sound you'll quickly regret. The best thing about the sound of old Fender Amps was those low wattage Blue-Bell speakers. When you over-drove those speakers you got a beautiful sound. Depending on the tubes you put in the pre-amp section you could get more or less distortion and the way in which you biased them had a lot to do with the clarity or mudd in the sound, but it was still the speaker that was producing most of the distortion characterists. Now, why people in-the-know still like tube amps isn't for their distorted sound. In fact, we'd rather have a solid state front end. But we like having tubes in the power section because they give you volume dynamics that solid state amps don't. That is, if I pick a note harder I'll get more volume and more distortion even though the power amp section isn't producing that distortion. Tube distortion is actually rather hard edged and nasal toned. I've never liked it. Solid state distortion sounds more like a broken up speaker--now that's what I'm after. That's why so many people like the Ibanez Tube Screamer. Despite its name its a solid state distortion pre-pre-amp devise that sounds very much like an over-driven speaker.

 

In PAL you can tell the difference between 4.1.1 and 4.2.2, the colours tend to be more blocky with the former and not as smooth.

 

No way, no how. 4:1:1 and 4:2:2 would have nothing to do with blockiness. Taking some color out would not produce blockiness. It would simply swap one color shade with a different one that's so close to the original that no one could tell the difference. Eki has suggested that you can tell when green screening and I'll take his word for it. But in normal viewing you shouldn't be able to tell one from the other.

 

35mm as against HD is petty cash

 

Hardly, we're talking about tens of thousands of dollars in savings on most movies. As Eki suggested, you can easily produce an entire HD movie in a computer work station environment and a 35mm print for $20,000. As to actors being over-paid, sure they are and yes, thats where the majority of the money goes. But that's only a small percentage of films that are made. MOST films are produced by independants on meager salaries, including those of the actors. Some of us have what it takes to work for ourselves and make a living. For the rest, there's always twinkyland.

 

I'm sorry, but this just isn't true. Video has much less colordepth than film. This is especially obvious on skintones, where video/digital simply doesn't look good. I can tell every time whether something is shot on video just by the way the skintones look.

 

On a monitor skin never looks as good as projected. Let's not forget that film and video have two different colorspaces.

 

Horse hockey. But nothing less than I've come to expect from you. The difference between 4:4:4 on down to 4:1:1 are differences you simply can in no way detect. This is also true for most forms of compression. They start by removing color info that you'd never miss. For instance, in the pic below you can see I've made a circular figure composed of various shades of red.

 

 

I used only 24-shades to do it. I don't give a damn if this is on a monitor (where you can see MORE info--not less) or a movie screen, 24-shades are 24-shades. Yet I'm confident that I could swap any two shades around in this pic that are right next to one another and you'd never know that I had done so. In other words, I could drop out half of these 24-shades and you probably wouldn't see the difference. But lets say you could. How about I use 24,000 circles with 24,000 shades in that same area to get from the beginning to the end. To think that I couldn't remove thousands of shades without you knowing it would be absurd! And that's exactly what changing color from 4:4:4 to 4:1:1 mostly entails. NOBODY could begin to see a difference. You might as well tell me you can broadjump the Atlantic. Here's another pic with 5 different shades in it. Do they really look any different? And again, this is from an already limited color space nowhere near 4:1:1. I could have picked 5 shades that were much closer to one another in degree than this. But I'm confident that you can't even tell from this pic that the shades are different.

 

 

I always thought this graph was interesting. The RGB color space is a small percentage of what our eyes see. I have no idea where film falls into this graph, but I think it demonstrates why it's harder to get things to look natural with HD/HDV.

 

See above pics. Plus (and this is huge) when you look at a photo your seeing colors distributed all over the place. The pics above are from one specific color in various shades right next to each other and even there its hard to tell one from the one next to it. If you don't think I could swap out the color shade in someone's iris or the zipper on their jacket wihtout you knowing the difference then you're just kidding yourself.

 

Skin tones fall within a certain range of hue, saturation and luminosity, when you watch them from film, in a theater.. As far as i know, this range can represented in 8-bit YUV, or 8-bit RGB without problems, as well as displayed by RGB devices, easily.

 

Exactly.

 

If it really were as simple as you say, then how come I ALWAYS recognize whether something is shot on video just by the look of the skintones? You claim that skintones shot on video can be made to look like those on film, but I have never and I mean NEVER seen that. Hell, I even recognize whether a film went through a DI or not by looking at the skintones.

 

Check your pants because there's definately smaoke in the house.

 

I mean, don't you think Rembrandt cared more about color and shade and composition than the consumers of his paintings? That's why he was Rembrandt and that's why the people buying his paintings didn't just paint them themselves.

 

Rembrandt was a pretty smart fella. I'm sure he couldn't care less if he had 25,000 shades of red or 250 since he would likely have been bright enought to know that no one could distinguish the difference, including himself unless he suffered from dillusiuons.

 

Most people can't tell a difference between the quality of an audio recording on a CD versus an MP3 recording; does that mean that professional audio people should only record at MP3 quality from now on even if THEY can hear a difference?

 

They can't if the mp3 is of a high enough bitrate/resolution/sample. I've run numerous tests on this over the years and no one has been successful at it as of yet. There's no reason they should be. Once you get up above about 256 bitrate you're mostly removing dead space from sound files. There's nothing to miss.

 

I never put my own money in my films.

 

Oh I would guess you'll be working for someone else the rest of your life. You'd never make it on your own.

[Erkki Halkka]

No way, no how. 4:1:1 and 4:2:2 would have nothing to do with blockiness. Taking some color out would not produce blockiness. It would simply swap one color shade with a different one that's so close to the original that no one could tell the difference. Eki has suggested that you can tell when green screening and I'll take his word for it. But in normal viewing you shouldn't be able to tell one from the other.

 

Charlie, you've totally misunderstood this 4:4:4 / 4:1:1 color space thing. 4:1:1 has exactly the same amount of color SHADES asd 4:4:4 - the amount of available colors depends on bit depth, not chroma sampling.

 

Lower chroma sampling reduces SPATIAL chroma resolution, giving blocky images. In 4:1:1, color is saved at only 1/4 resolution - or in other words, color information is saved only for each 4th pixel.

 

It's not much different than sampling audio: sampling resolution tells how many samples are taken (i.e. 48 000 samples of audio / second - every pixel of luminosity channel [4] and every 2nd pixel of both color channels [2:2] in 4:2:2 color sampling), bit depth tells how many diffrent levels those samples can be at (256 colors / channel in 24 bit RGB images or 256 audio levels at 8 bit audio).

 

I.e 4:2:2 could just as well be marked as 1:1/2:1/2.

 

If you have i.e. 100*500 pixel image, you'll get

 

4:4:4

 

1000*500 luminosity resolution

1000*500 color resolution

 

4:2:2

 

1000*500 luminosity resolution

500*500 color resolution

 

4:1:1

1000*500 luminosity resolution

500*250 color resolution

 

Please, have a look at this image again, it shows the artifacts well.

 

http://eki.pp.fi/temp/Eki/ColorSampling/Madeira_411.png

 

So does this:

 

http://eki.pp.fi/temp/MiniDV_uncomp_from_HDV.png

Edited December 14, 2005 by Eki Halkka

[Phil Rhodes]

Hi,

 

> No way, no how. 4:1:1 and 4:2:2 would have nothing to do with blockiness.

 

Not so. The attached JPEG clearly demonstrates blockiness due to colour subsampling. You're confusing spatial resolution with quantisation resolution. For the record I wouldn't be too horribly depressed if a final show format for digital cinema was 4:2:2 subsampled, but it isn't really something we want to see on an origination format.

 

Image provided at 400% actual size. PAL Panasonic AG-DVC200; Microsoft Windows XP DV decoder; no overriding degradation by JPEG compression.

 

As for all the skintone issues, I call "drivel". Because I'm polite. If you shoot an 8-bit format and colour correct the hell out of it, yes, it'll start contouring, but that's not an problem inherent to all video any more than the super-size grain of super-8 is a problem inherent to all film. The problem I do see time and again on otherwise nice video is skin shininess - video sees this and picks it up like nothing on earth, so ensure makeup are aware and get a polariser in the way!

 

But "electronically reproduced skintones are inherently bad?" Balls. I have very little time for the circular logic of "I can always tell." Because if you can't tell, you wouldn't know to gripe about it, now would you?

 

Phil

[Joshua Provost]

No way, no how. 4:1:1 and 4:2:2 would have nothing to do with blockiness. Taking some color out would not produce blockiness. It would simply swap one color shade with a different one that's so close to the original that no one could tell the difference. Eki has suggested that you can tell when green screening and I'll take his word for it. But in normal viewing you shouldn't be able to tell one from the other.

 

Horse hockey. But nothing less than I've come to expect from you. The difference between 4:4:4 on down to 4:1:1 are differences you simply can in no way detect. This is also true for most forms of compression. They start by removing color info that you'd never miss. For instance, in the pic below you can see I've made a circular figure composed of various shades of red.

 

I'm glad Charlie finally exposed himself, and that Eki was quick on calling it out. 4:2:2 vs. 4:1:1 has nothing to do with color depth. It's 24-bit (8-bit/channel) color regardless of 4:2:2 vs. 4:1:1. It's the same number of "shades."

 

Charlie, have you ever worked with 4:1:1 video? It's obvious to anyone who works with this stuff. My wife walks in the room and remarks "Why is that blocky?"

 

I think the downsmapled/upsampled images you have posted are dubious. I have examined them closely, and believe them to be identical. It was convenient of you to encapsulate them in Flash, so it would be harder to critically analyze them. I know exactly where to look to find the 4:1:1 artifacts. These artifacts must exist, but they don't in your images, so your images aren't 4:1:1.

 

You are an audio guy, you should know that there have been test of 48kHz vs. 96kHz vs. 192kHz audio. Even though the human ear is not supposed to be able to hear the higher frequencies the larger formats capture, in blind A/B tests subjects could tell the difference, they could feel the difference.

 

The more information you can cpature and present, the higher the perceived quality. Compression techniques are based on achieving acceptable levels of quality. The uncompressed originals still retain higher quality. Even if I look at a 4:1:1 image and say, "sure, that looks good," a 4:2:2 image might look "great," and 4:4:4 image "exceptional."

 

Man, you are in the deepest video denial I have ever witnessed. Try jumping over to dvxuser.com or dvinfo.net, where thousands of professionals who use 4:1:1 and 4:2:0 every day are hopeful of moving to reasonable-cost 4:2:2 formats, so they can improve on the quality compromise they currently operate under.

Edited December 14, 2005 by Joshua Provost

[Erkki Halkka]

You are an audio guy, you should know that there have been test of 48kHz vs. 96kHz vs. 192kHz audio. Even though the human ear is not supposed to be able to hear the higher frequencies the larger formats capture, in blind A/B tests subjects could tell the difference, they could feel the difference.

 

I don't think it's necessarily about the higher than 22K frequencies - it's about the high but hearable frequencies (near 20kHz) having too few samples for each wave length, causing distortion.

 

4:4:4 - 4:2:2 - 4:1:1 is analogue to 96, 44 and 22 kHz audio sampling... you can't always hear the difference between 96 kHz and 22 kHz sound, but sometimes you do. It totally depends on the source.

 

As far as mp3's go - 128 Kbit sounds usully rather OK. But in SOME cases...

 

That said, i've destroyed my hearing by playing in rock bands in the 80's, i couldn't hear the difference at all - hasn't stopped me from mixing stuff for video though, i just need a pair of good ears to come and check the mixes for me in case of nasty errors in the high end ;-)

Edited December 14, 2005 by Eki Halkka

[Erkki Halkka]

How about I use 24,000 circles with 24,000 shades in that same area to get from the beginning to the end. To think that I couldn't remove thousands of shades without you knowing it would be absurd! And that's exactly what changing color from 4:4:4 to 4:1:1 mostly entails.

 

Even if the above was true (if you weren't wrong about how color spaces work making your point 100% moot), you'd still be wrong.

 

 

See the above gradient. The top portion of it uses 64 shades (same as 7 bit), the bottom portion uses 128 shades (8 bit). Unless you have a very bad monitor, you should see banding only in the top half of the image, because there's not enough shades to show the gradient properly. If your monitor is bad, both show banding.

Edited December 14, 2005 by Eki Halkka

[Tenolian Bell]

The beauty of it is that it uses actual film grain. The grain looks real simply because it IS real.

 

OK, never heard of that. Yes it is a better option than grain look plug in's. But still not quite the same as actually shooting film.

 

Just a quick question - who among the posters in this thread has actually shot 35mm film and | or been in a telecine transfer with one of the big 10 post houses.

 

Yes to both questions. Have shot and telecined in big post houses in NY and LA.

 

So, how many of us here has shot HDV?

Done post on it?

Worked on HDV to film project(s)?

Seen HDV to 35mm print footage in a theater?

 

Yes just recently shot HDV.

The two shorts I shot are in post now.

The two shorts will live on TV and never go to film.

I have never seen HDV to 35mm blow up.

Or even know anyone who has attempted it.

 

I can only say that I've seen great stuff on a 30' screen shot on 16mm, especially super16.

 

City of God

Ballad of Jack and Rose

The Constant Gardner

 

All shot on super 16. I have seen them all on screens larger than 30 feet.

 

Hardly, we're talking about tens of thousands of dollars in savings on most movies. As Eki suggested, you can easily produce an entire HD movie in a computer work station environment and a 35mm print for $20,000. As to actors being over-paid, sure they are and yes, thats where the majority of the money goes. But that's only a small percentage of films that are made. MOST films are produced by independants on meager salaries, including those of the actors. Some of us have what it takes to work for ourselves and make a living. For the rest, there's always twinkyland.

 

Charlie you need to keep things in a perspective. If your film budget is $60 million. The price of film stock, processing, telecine is a very small percent of that. If your film budget is $500,000 the percentage those variables take from the budget are considerably more. Whether you believe it or not Hollywood is not that concerned about the price of film. Seeing the fact that HDCAM has been around for five years but nearly all studio films continue to be shot on 35mm. More money can be spent on food than is spent on film stock.

 

I'm sure someone can deliver you HD to 35mm print for $20,000. But I doubt it would be all that great.

 

Most films being produced by independents. I guess technically that's true.

 

Most films that a large audience of people actually see and actually produce sizable profit, pay their actors a lot of money.

 

 

I never put my own money in my films.

 

Oh I would guess you'll be working for someone else the rest of your life. You'd never make it on your own.

 

Within the studio production and distribution system, no one pays for their own film. All films have multiple backers and investors.

[Brian Drysdale]

Well, as somebody who's watched movies for 46-years and still has good far sightedness, I can only say that I've seen great stuff on a 30' screen shot on 16mm, especially super16.

 

.

 

 

A 16mm print projected onto a 30ft screen doesn't look good.

 

A Super 16 print blown up to 35mm looks good when projected 30 ft wide. Assuming equal quality at the lab a standard 16mm blow up won't be quite so good. However, a film shot on 35mm will look much better.

[Phil Rhodes]

Hi,

 

> Within the studio production and distribution system, no one pays for their own film. All films have multiple

> backers and investors.

 

Yes, obviously, but that's not really "his" film then, is it?

 

Phil

[Tenolian Bell]

We want to argue semantics?

 

If someone else puts up most of the money, then they reap most of the profit.

 

If he directed it then creatively yes it is his work, and he gets the imdb credit.

[Sam Wells]

A 16mm print projected onto a 30ft screen doesn't look good.

 

A Super 16 print blown up to 35mm looks good when projected 30 ft wide. Assuming equal quality at the lab a standard 16mm blow up won't be quite so good. However, a film shot on 35mm will look much better.

 

I've seen 16mm projected on ~ 32' screen which looked quite good. Depends on the projector, optics etc.

 

The largest I've seen my prints was about 25' - I was quite happy. The *sharpest* projection of one of my prints I've seen was with an Eastman 25 and a hand picked ISCO lens. This was 20' wide or so but I had the feeling there was room to expand :)

 

The *big* advantage of 35mm projection in comparison and this would include 16/S16 blowups is even illumination across the screen and edge to edge sharpness.

 

-Sam

[Brian Drysdale]

I've seen 16mm projected on ~ 32' screen which looked quite good. Depends on the projector, optics etc.

 

The largest I've seen my prints was about 25' - I was quite happy. The *sharpest* projection of one of my prints I've seen was with an Eastman 25 and a hand picked ISCO lens. This was 20' wide or so but I had the feeling there was room to expand :)

 

The *big* advantage of 35mm projection in comparison and this would include 16/S16 blowups is even illumination across the screen and edge to edge sharpness.

 

-Sam

 

The projector is the key, but how often do you get high quality 16mm projection with a great lens? I've been at a number of film festivals where you can see the drop in quality straight away (even on a smaller picture) compared to the blow ups.

 

However, the biggest problem with projecting 16mm is the quality of the optical sound.

[Max Jacoby]

No, Max, I'd quite like to find out who you've persuaded to pay to to shoot your personal glory projects on 35 - mainly in the sincere hope that I'm not inadvertently contributing myself!

What the fu** is your problem Phil? Just because the UK's financing of short films is pretty non existent does that mean that no one else around the world is allowed a budget for short films either?

 

As for the financing, most of the budget (which was ?150K) came from the Filmfund of Luxembourg, but my production company put some of their money in it as well and another part came through taxshelter. The Filmfund is financed by the government and their money comes from the taxpayers, so in a way every taxpaying person in Luxembourg, and that includes myself, contributed to the film.

[Max Jacoby]

Check your pants because there's definately smaoke in the house.

Charlie

 

There is no need to insult people. I have given up trying to argue with you, because it is rather like trying to explain colors to a blind person. As you can see from other people's post, I am certainly not the only one who feels that way about you.

[Sam Wells]

The projector is the key, but how often do you get high quality 16mm projection with a great lens? I've been at a number of film festivals where you can see the drop in quality straight away (even on a smaller picture) compared to the blow ups.

 

However, the biggest problem with projecting 16mm is the quality of the optical sound.

 

It requires Vigilance as the former Attorney General liked to say.

 

Well I've done DTS in 16mm, that worked :) (not mass distributable......)

 

16mm projection seems like a sidetrack from the issues here..

 

-Sam

[Stephen Williams]

...it might also be interesting to know how much experience with HDV the participants have.

 

So, how many of us here has shot HDV?

Done post on it?

Worked on HDV to film project(s)?

Seen HDV to 35mm print footage in a theater?

 

"Yes" to all accounts here ;-)

 

 

Hi,

 

Yes I have shot on HDV with a Z1. I was very impressed with the pictures, for a low cost video camera.

 

If I knew I wanted a 35mm print at the start of the project I would not master on HDV. I would shoot 35mm 1:185.

 

Stephen

[Brian Drysdale]

It requires Vigilance as the former Attorney General liked to say.

 

Well I've done DTS in 16mm, that worked :) (not mass distributable......)

 

16mm projection seems like a sidetrack from the issues here..

 

-Sam

 

 

Interesting about the DTS.

 

Agree about the 16mm projection being a side track in this thread.

[Guest Charlie Seper]

In 4:1:1, color is saved at only 1/4 resolution - or in other words, color information is saved only for each 4th pixel.

 

Yes, but with 180 color samples per scan line in 4:1:1, half that of 4:2:2, the amount of luma you have in the graphic is going to play into the the percieved color. The colors (in theory) have the ability to be richer in 4:2:2. However, if we're still talking about 25mbps either way, you'd still have the same amount of info available. 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 or even at 4:1:1 after compressing at 75%. You'll also not see any banding in my photo of the circle where the dots overlap.

 

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

 

He also has published some info on audience viewing tests from the EBU concerning diefferences between 4:1:1 DV at both 50mbps and 25mbps and a few other here: EBU Test

 

I'm glad Charlie finally exposed himself, and that Eki was quick on calling it out. 4:2:2 vs. 4:1:1 has nothing to do with color depth. It's 24-bit (8-bit/channel) color regardless of 4:2:2 vs. 4:1:1. It's the same number of "shades."... Charlie, have you ever worked with 4:1:1 video?

 

A 4:1:1 is the only cam I own. See above comments.

 

I think the downsmapled/upsampled images you have posted are dubious. I have examined them closely, and believe them to be identical.

 

Thanks, that's the best argument in my favor you could have given me. Tell you what kid, I'll send a PM to Eki telling him which is which in the four photos. I don't know him well but I'll trust him to be honest with us all. In the meantime why don't you simply download one of the many free flash (swf) extractors/decompilers that are available and have a closer look yourself. I'm amazed that no one had the brains to do that already. How anyone expects to work in video in this day and age, and not know Flash inside out, is beyond me. Extracting swf files (except for audio) in a no-brainer. The photos are each named for what they represent i.e.--4:2:2.jpg and so on. If you still can't figure it out, ask Eki and mabe he'll tell you which is which by PM or email. I'd rather not post the answers here in case anyone else takes the test.

 

It was convenient of you to encapsulate them in Flash, so it would be harder to critically analyze them.

 

Like I said, its a no-brainer extracting swf files, or at least it should be for anyone with a background in video. The fact that you haven't been able to is rather telling in itself....

 

You are an audio guy, you should know that there have been test of 48kHz vs. 96kHz vs. 192kHz audio. Even though the human ear is not supposed to be able to hear the higher frequencies the larger formats capture, in blind A/B tests subjects could tell the difference, they could feel the difference.

 

That's simply incorrect. I've done test after test after test on this and no one has ever been able to tell which is which. Here's a link to an audio test I posted at this website back in August. I even gave the answers at the end. My Audio Test

 

If you think you can tell the difference between those then I'll be happy to produce a few more tests where I WON'T give you the answers, and will in fact lock them in a flash file where you can't put the files through a spectrum analysis graph to cheat, because given your comments in this thread so far, I have every reason to think you're dishonest from A to Z. At any rate, I don't know of any swf extractors that can extract the audio, so they should be safe locked in a swf file.

 

As far as mp3's go - 128 Kbit sounds usully rather OK. But in SOME cases...

 

Most of the problems stem from a narrowing of the stereo spectrum. This is especially evident on stereo miked acoustic instruments. You get a sort of out-of-phase sound just like you would have if you had your mics too close together. WMA files are much better at producing the same sounds in the same bandwidth. In fact, even at 160mbps they sound quite good.

 

That said, i've destroyed my hearing by playing in rock bands in the 80's, i couldn't hear the difference at all

 

Audio guys very often suffer from the same misconceptions as video guys in that they like to think they have some kind of super-human abilties in their sense perceptions. I've even heard some say that they liked the sound of one "wave engine" over another. Of course a wave is a wave is a wave. No matter what proggy you use they all record the same exact way. Moreover, your hearing is the first thing to go. That 20 to 20k humans are capable of hearing is pretty well gone by the time your 15. Few people can hear anything above 16k by the time they're 20.

 

Charlie

 

There is no need to insult people.

 

Given your previous message to Phil, "What the fu** is your problem Phil?" and looking back over your posts in this and other forums, you might want to get the log out of your own eye pal. You strike me as someone who'd bitch over a spec of pixie dust on his ballet slippers, even if they could only see it under a magnifying glass.

Edited December 16, 2005 by Charlie Seper

[Guest Charlie Seper]

What the heck, 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. I can always make another test.

 

4:1:1 Pic (bumped back up to 4:2:2)

 

4:2:2 Pic

 

4:1:1 75% Compressed Pic

[Phil Rhodes]

Hi,

 

> 4:1:1 DV at both 50mbps and 25mbps

 

25 megabit DV is 4:1:1 (or 4:2:0). The 50-megabit variants are, universally unless I'm missing something, 4:2:2.

 

Yes, they will look extremely similar on first generation. I have no problem with people using 25-meg DV for applications where it isn't going to be processed. It generally looks better than Beta SP. But it really does fall apart very quickly if you try to grade it, although only slightly worse than anything else.

 

Phil

[Guest Charlie Seper]

I made a mistake when I said, "4:1:1 Pic (bumped back up to 4:2:2)". It was NOT bumped back up, but left at 4:1:1.

 

A point I should clarify is when I said, "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 or even at 4:1:1 after compressing at 75%."

 

If you magnify the compressed photo several times you'll certainly see compression artifacts in the compressed photo. I don't think of it as blocky though, more of a blur or smear at this compression. Once you get down around 50% or worse compression, that's when you really get into the blocky look.

 

Also, We just had the St. Louis film festival a couple of weeks ago. Aside from HDV, I think I've seen just about every flavor of cam work on the big screen by now (although often on just a 20' screen) and even the worst DV movies never looked blocky to me. They were certainly soft and even blurry if really bad, but I wouldn't say they looked blocky anywhere. Sometimes they're a bit noisy, like a TV station that doesn't have the best reception.

~~~~~~~~~~~~~

~~~~~~~~~~~~~

 

25 megabit DV is 4:1:1 (or 4:2:0). The 50-megabit variants are, universally unless I'm missing something, 4:2:2.

 

Yeah, I know. I thought it was odd when he said that (I don't even remember who it was at this point) but I think maybe he meant 4:2:0 Pal DV and was comparing that to 4:1:1 NTSC, both at 25mbps. I don't know of any system that records 4:2:2 at 25mbps. I mean, there may be one or more that somehow do it for all I know, but I don't know of any.

 

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.

Edited December 16, 2005 by Charlie Seper

[Phil Rhodes]

Hi,

 

Frankly I think that 4:2:0 does look quite considerably better than 4:1:1. For chromakey it's only as blocky in the vertical direction as Digibeta is in the horizontal - it works a lot better. The dogma of "you can't chromakey DV" is much less true for PAL than it is for NTSC. The downside of 4:2:0 is that you can get chroma pollution between fields, which has caused me fairly severe problems in the past regarding solving field order issues.

 

No, DCT compression doesn't tend to look blocky with the coefficient quantisation at the sort of levels applied by DV. If you recompress it a few times and apply processing, particularly that which increases contrast, you can begin to see macroblock field boundaries. The artifact here is most often a repeating pattern of dots, rather than actual visible squares, rather like a random-dot stereogram. Obviously, projecting standard def on any scale larger than a TV screen tends to look objectionably soft.

 

IBP formats like MPEG-2 as applied to HDV are, in my experience, more likely to suffer block noise if only because the compression electronics are required to work in three dimensions rather than two. 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.

 

In my opinion, HDCAM burned out to 35 looks mushy in the same way that slight halation from a groundglass lens adaptor does. Not great, but overlookable in the context of a well-produced feature. The thing here is that it looks very much better projected electronically, without an extra layer of grain and instability pasted over the top of it.

 

Phil

[Andrew Koch]

Charlie said:

-A 4:1:1 is the only cam I own. See above comments.

 

 

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.