"Club Penton" - Ask questions to Hollywood insider "Penton-Man"

[Penton-Man]

Originally Posted by madshi

Agreed. I was just saying that colors are harder to judge than some other things.

madshi, I posted this hue test here back in 2009 and I believe that everyone who took the test (despite not being professional digital colorists) did pretty damn good…
http://www.xrite.com/custom_page.aspx?PageID=77

In fact, several people scored either perfect scores (JimSD, DenonCI?) or else near perfect scores.

Regarding Robin Hood for enthusiasts who demand accuracy of their Blu-rays PQ to the original theatrical presentation (from A.C.) -

“In terms of a color palette, Mathieson says he and Scott wanted to avoid the rich greens and reds that characterize one of the most famous Robin Hood adaptations, Michael Curtiz’s early Technicolor production The Adventures of Robin Hood (1938). “There are a lot of burned browns and dark colors in our costumes and the sets,” says Mathieson. “It’s a pretty mucky-looking film, but we did shoot in the summer, and our ideal would have been to shoot right through the winter. I just love the winter light in the U.K. — the slow, long sun. It was a very wet summer, so we did get a lot of overcast skies, but an overcast sky in summer is not very interesting because there’s no movement in it. It’s often just a flat white, and you want more gray.”

I wonder if it will be acceptable to videophiles to make the Blu-ray rendition of Robin Hood with more rich and vivid colors as long as it is sharp as a tack?

You’re an algorithm guy right? Well, those downsampling digital filters are based on algos and those filters don’t do that great with tiny fractions (ratios, i.e. 2048 ->1920))

I find it just slightly disingenuous when last year some screenshot scientists b*tched and moaned about the slight cropping of a few Blu-ray titles as compared to the theatrical presentation despite the fact that by utilizing that mastering procedure it noticeably increases the clarity and detail of a 2k sourced Blu-ray image but.....then, they are fine with reduxed colors of a Blu-ray…as long as it is not soft due to overzealous dirt and scratch removal automatic settings with compensatory sharpening thrown in.

[Penton-Man]

Originally Posted by DenonCI

My dog barks at every freaking animal in any movie we are watching. Granted, he likes to yap and I have a front projector so they probably scare the shit out of him

Rumor has it, the dogs loved the Coyote Falls 3D trailer that preceded the main feature featuring their wild cousin Wile E. Coyote and that damn road runner….

http://www.youtube.com/watch?v=8vou4AhXHgA

[Doctorossi]

Originally Posted by Penton-Man

I find it just slightly disingenuous when last year some screenshot scientists b*tched and moaned about the slight cropping of a few Blu-ray titles as compared to the theatrical presentation despite the fact that by utilizing that mastering procedure it noticeably increases the clarity and detail of a 2k sourced Blu-ray image but.....then, they are fine with reduxed colors of a Blu-ray…as long as it is not soft due to overzealous dirt and scratch removal automatic settings with compensatory sharpening thrown in.

Disingenuous? Just sounds uneducated and misguided to me- you know, typical *science*.

That said, we all have a slightly different list of priorities about which aspects of picture-quality are most important to image-fidelity in our own eyes and I have a hard time begrudging any passionate fan having a slightly different priority list than mine.

Just for argument's sake, exactly how much cropping are we talking about here?

[madshi]

Originally Posted by Penton-Man

madshi, I posted this hue test here back in 2009 and I believe that everyone who took the test (despite not being professional digital colorists) did pretty damn good…
http://www.xrite.com/worldwide.aspx

Got a score of 4. But what does the ability to sort hues have to do with being or not being able to remember how a movie looked like in movie theaters many many years ago? Personally, I haven't even seen Gladiator in a movie theater at all.

Originally Posted by Penton-Man

I wonder if it will be acceptable to videophiles to make the Blu-ray rendition of Robin Hood with more rich and vivid colors as long as it is sharp as a tack?

No, it will not be acceptable, if the modified colors do not match the director's intent. The big problem is how to know the director's intent for sure.

Originally Posted by Penton-Man

You’re an algorithm guy right? Well, those downsampling digital filters are based on algos and those filters don’t do that great with tiny fractions (ratios, i.e. 2048 ->1920))

Depends on the algorithm. E.g. a proper written Lanczos filter handles such a downscale just fine. Unfortunately it adds ringing to the image. A simple Bilinear filter will not add ringing, but make the image softer. I'd probably use Mitchell-Netravali for a 2048 -> 1920 downscale as a good compromise. Or I'd modify a Lanczos filter to not ring.

Originally Posted by Penton-Man

I find it just slightly disingenuous when last year some screenshot scientists b*tched and moaned about the slight cropping of a few Blu-ray titles as compared to the theatrical presentation despite the fact that by utilizing that mastering procedure it noticeably increases the clarity and detail of a 2k sourced Blu-ray image

I agree with you that a crop is prefered to a downsize, if the lost image area is relatively small and if it noticeably improves image quality. Of course the best solution would be a 4K or 6K scan, downscaled directly to Blu-Ray resolution.

Originally Posted by Penton-Man

but.....then, they are fine with reduxed colors of a Blu-ray…

They would probably not be fine with it, if it was clear that the colors are changed away from director's intent. We do not really know that for sure, though.

Look, for us "normal" consumers it's really hard: We are very used to getting multiple releases of movies which differ in colors. There are even Blu-Rays released in different countries, based on different masters, with different colors. How the heck are we consumers expected to know which colors are "correct" and which are not? There is no way for us to know! Probably because of that we've grown more or less used to colors changing between different releases, without asking why or how.

At least that's my explanation for why there's not so much discussion about the Gladiator color change...

*Again*: That doesn't mean that we don't care about changing colors away from director's intent. It's just that we have no way to know...

[Doctorossi]

Originally Posted by madshi

I agree with you that a crop is prefered to a downsize, if the lost image area is relatively small and if it noticeably improves image quality.

I'm not sure I agree. In either case, the viewable resolution is still probably going to exceed most theatrical showings. Meanwhile, framing can be very rich with meaning and has an extremely limited resolution tolerance, if you follow my conceptual wrangling (what's the effective MTF of "wrong"?).

[Penton-Man]

Originally Posted by Doctorossi

Disingenuous? Just sounds uneducated and misguided to me- you know, typical *science*.

That said, we all have a slightly different list of priorities about which aspects of picture-quality are most important to image-fidelity in our own eyes and I have a hard time begrudging any passionate fan having a slightly different priority list than mine.

Just for argument's sake, exactly how much cropping are we talking about here?

It’s very small. Within SMPTE recommendations, well within all scanner tolerances, except maybe the Northlight which is more precise and probably, most importantly, within the gate inaccuracies of your average film projectors, thru which people are viewing the original theatrical presentation at many of their Multiplexes.

Doc (and to anyone else who is not following the discussion) the main point is that a correct (maintaining 1:1 pixel) 2048 ->1920 crop will result in more obvious increase in resolution and clarity than the difference in resolution which people are seeing with those screenshots which Oliver and madshi posted showing the increased detail of the 2010 Blur-ray of Gladiator compared to that of the color correct, albeit softer looking 2009 edition of the Gladiator Blu-ray.

[Penton-Man]

Originally Posted by madshi

Got a score of 4. But what does the ability to sort hues have to do with being or not being able to remember how a movie looked like in movie theaters many many years ago? Personally, I haven't even seen Gladiator in a movie theater at all.

Well, I was referring more to general online audiences and their duly capable ability to differentiate colors when their memories of such hotly debated motion pictures are intact. I was answering your question about colors with this past post as a subtext…..

https://forum.blu-ray.com/insider-di...ml#post3542230

[Penton-Man]

Originally Posted by madshi

How the heck are we consumers expected to know which colors are "correct" and which are not?

Do some research as to what the Director and/or his DP have said in the past regarding any post production facilities’ inability to reproduce the correct color scheme on the release prints. In this case, there is no evidence to suggest the Director being unhappy that his color palate was not successfully maintained with the typical release print. In fact, on the contrary, you'll find more than one complement as to Technicolor's work.

In lieu of that, attend a local screening featuring one of the filmmakers intimately involved in the production where no such ‘color discrepancy’ between the typical release print and the “Director’s intent” was debated …..much less entertained …..as such a hypothetical is a nonissue in this case.

[Penton-Man]

Originally Posted by madshi

Depends on the algorithm. E.g. a proper written Lanczos filter handles such a downscale just fine. Unfortunately it adds ringing to the image. A simple Bilinear filter will not add ringing, but make the image softer. I'd probably use Mitchell-Netravali for a 2048 -> 1920 downscale as a good compromise. Or I'd modify a Lanczos filter to not ring.

It would seem that your mathematical prowess and ability to implement such a thing exceeds that of anyone else in the business, including Galt. I would suggest that you implement your algorithmic changes and bring it to market in the form of a digital filter because me thinks you could make yourself a lot of money in the process.

Madshi, read this article, and if it’s a bit long and not completely relevant to our discussion, if memory serves, use the key word “horrible” to find the relevant paragraph. Unfortunately, I’m not at liberty to give you a transcript of his presentation at last years The Reel Thing.
http://magazine.creativecow.net/arti...ture-of-pixels

[Penton-Man]

Originally Posted by Penton-Man

The Reel Thing

Speaking of which, this year’s Preliminary Schedule has just been posted (I love these tie-ins ) -
http://www.amianet.org/events/theree...10/program.htm

Be there or be square.

[madshi]

Originally Posted by Penton-Man

the main point is that a correct (maintaining 1:1 pixel) 2048 ->1920 crop will result in more obvious increase in resolution and clarity than the difference in resolution which people are seeing with those screenshots which Oliver and madshi posted showing the increased detail of the 2010 Blur-ray of Gladiator compared to that of the color correct, albeit softer looking 2009 edition of the Gladiator Blu-ray.

I'm sorry, but this is absolutely and totally incorrect. A proper downscale from 2048 -> 1920 will reduce resolution only very slightly, while the difference in detail/resolution between the two Gladiator BDs is like night and day. If you want, I can create & upload a demo 1920 -> xxx (choose any resolution you want) downscale to proof my point.

I guess you haven't seen a proper downscaling algorithm yet?

[madshi]

Originally Posted by Penton-Man

It would seem that your mathematical prowess and ability to implement such a thing exceeds that of anyone else in the business, including Galt. I would suggest that you implement your algorithmic changes and bring it to market in the form of a digital filter because me thinks you could make yourself a lot of money in the process.

Sounds like you're having fun with me. Anyway, maybe this could help convince you that I'm not talking out of my a..., ehm, bottom:

http://forum.doom9.org/showthread.php?t=145358

[Penton-Man]

Originally Posted by madshi

I'm sorry, but this is absolutely and totally incorrect. A proper downscale from 2048 -> 1920 will reduce resolution only very slightly, while the difference in detail/resolution between the two Gladiator BDs is like night and day. If you want, I can create & upload a demo 1920 -> xxx (choose any resolution you want) downscale to proof my point.

I guess you haven't seen a proper downscaling algorithm yet?

Off the top of my head, I would say that a "proper" crop of a 2k source will give you as much as a 20% (or more, maybe 30%) increase in resolution than any "proper" downscaling algorithm I've seen (by guys here in L.A. who have done proprietary algorithms for certain IMAX features, past and present).

Is the 2010 Blu-ray of Gladiator giving you that much increased clarity over the 2009 version?

Maybe you haven't seen a proper 2K -> HD crop done...without some subpixel resizing thrown in the mix?

[Penton-Man]

Originally Posted by madshi

Sounds like you're having fun with me.

Yup.

[madshi]

Originally Posted by Penton-Man

Off the top of my head, I would say that a "proper" crop of a 2k source will give you as much as a 20% (or more, maybe 30%) increase in resolution than any "proper" downscaling algorithm I've seen

Ok, let's take that to the test. Are those Pixar movies rendered directly to Blu-Ray resolution? If so, if I take a screenshot of e.g. Ratatouille and downscale that, that should give us a good indication of how much resolution gets lost, do you agree?

Originally Posted by Penton-Man

Is the 2010 Blu-ray of Gladiator giving you that much increased clarity over the 2009 version?

It's hard to quantify the increased clarity. But for me the 2010 BD is at least twice as detailed as the 2009 BD, maybe more.

Originally Posted by Penton-Man

Maybe you haven't seen a proper 2K -> HD crop done...without some subpixel resizing thrown in the mix?

Well, there probably are some Blu-Rays which are 2K -> HD cropped. Name one, and if I have that in my collection, I can try downscaling it !

I want to put your claim of a 20-30% resolution loss on the test bench. Do you bite?

[Penton-Man]

Originally Posted by madshi

Anyway, maybe this could help convince you that I'm not talking out of my a..., ehm, bottom:

http://forum.doom9.org/showthread.php?t=145358

I believe that you are a very smart guy and if you think that you can do in the real world what I've failed to see others do, then I would suggest sending in a resume to David K. here....
http://www.imax.com/corporate/theatr...ostProduction/

If you're from the U.K. then contact my old friend Mr.D over on AVS....as he too might be interested in your expertise.

[Penton-Man]

Anyway, I don’t want to dominate this thread with downsampling algo talk; otherwise I'll fall asleep at the keyboard and we'll probably drive everyone but Spears away so, speaking of other expertise, I’ve just been informed of the credentials of the guy that did the Gladiator review which I linked here -
https://forum.blu-ray.com/insider-di...ml#post3546616

Apparently, he has been reviewing home entertainment software media for years, if not decades , so he’s not a duffer and his name is John Puccio, if anyone is familiar with him from the days of DVDs and such.

[madshi]

Originally Posted by Penton-Man

I believe that you are a very smart guy and if you think that you can do in the real world what I've failed to see others do, then I would suggest sending in a resume to David K. here....
http://www.imax.com/corporate/theatr...ostProduction/

Thank you. I do like to be my own boss, though...

Originally Posted by Penton-Man

If you're from the U.K. then contact my old friend Mr.D over on AVS....as he too might be interested in your expertise.

I've already had a few chats with Mr.D. He's helped out now and then with calibration related hints that helped during development of my HTPC software.

[Penton-Man]

I foresee a new use of digital manipulation for 3D post production specialists, i.e. quelling the fears of some actresses….http://www.stuff.co.nz/entertainment...tmans-3D-worry

[Vincent Pereira]

Originally Posted by Penton-Man

...the 2010 Blur-ray of Gladiator...

"Blur-ray"? I just had a flashback to the format wars...

Could it be that Penton-Man is a double-agent and is, in fact, a red ant???!?!?!



Vincent