Mtf Of Blu-ray/hd-dvds

[momaw]

The above is insulting. And the first sentence is so loosely worded it is unclear what it means.

It may may pass as an acceptable or normal post in the home thetare subforum, from some of what I haves seen posted there, but it is out of place here.

Please try to be constructive and specific, not vague and insulting.

:lol: :lol:

That's an insult????

Your premise is faulty to start with. You have provided no evidence. Your assumptions are based on the wrong technology. You seemingly know nothing about HD DVD or Blu-ray technology. You attribute MTF to things it is not applicable too (ie CGI involving no optics). You don't seem to understand movies ar enot shot at 1920x1080 or the process to get them transfered.

Seriously, what's the agenda?

And now your called out on it, you cry insult????

[MLXXX]

Momaw, I edited out my remarks when you amplified your post. In its original form your post was incomprehensible and insulting. Your amended verion was still insulting, but at least you had made some effort to explain the point you were making. My amended post is at #74 above.

I cannot respond to a mish mash of allegations. If you (and it appears Alebonau) have not understood, or have understood but do not agree, that is that.

I cannot go back and repeat myself endlessly.

(The calmness was nice while it lasted.)

Cheers,

MLXXX

Edited October 2, 2007 by MLXXX

[betty boop]

:lol: :lol:

That's an insult????

Your premise is faulty to start with. You have provided no evidence. Your assumptions are based on the wrong technology. You seemingly know nothing about HD DVD or Blu-ray technology. You attribute MTF to things it is not applicable too (ie CGI involving no optics). You don't seem to understand movies ar enot shot at 1920x1080 or the process to get them transfered.

Seriously, what's the agenda?

And now your called out on it, you cry insult????

well summed up momaw.

what I infact find an insult is that their personal & subjective opinion has been put forward as fact with no evidence or knowledge or experiece to back that up when challenged.

one day mlx, for owen and you the penny indeed does drop..

[momaw]

Momaw, I edited out my remarks when you amplified your post. In its original form your post was incomprehensible and insulting. Your amended verion was still insulting, but at least you had made some effort to explain the point you were making. My amended post is at #74 above.

Interesting as I didn't amend my post at all. Why is it now comprehensible when it wasn't earlier? And as to the insulting..... Good god boy. Even the thinest of skin could not find that insulting. But if you insist it is insulting, why not report it and let the mods be the judge.

Interesting that you claim we (Al and myself) have misunderstood when you attribute a problem with optical capture to something that has no optical capture involved at all i.e. cgi. Nor do you address the fact that movies are filmed at such high resolution that scaled down any inefficiencies can be filtered out.

[MLXXX]

Interesting as I didn't amend my post at all. Why is it now comprehensible when it wasn't earlier?

Perhaps I am going blind or perhaps there was some quirk in the system but when I saw your post it lacked the middle para. What I saw was as follows:

you do realize 1920x1080 actually refers to the number of lines of resolution and has nothing to do with the number of pixels - which your whole argument is based on (well that and technology that is completely irrelevant)?

I wondering what agenda you and owen actually have to be spreading such mis-information?

The gist I'm getting is that "full HD" doesnt have 1920x1080 resolvable resolution? How does this affect viewing distances (discounting PC's and CGI's I think :/ ) if there isnt really the resolution there to resolve?

Short answer: the charts for viewing televison video already take the limitation into account.

Longer answer:

The viewing distance charts were principally developed based on principles of human vision, and by actual use have been confirmed to work.

Real world video in a 1920x1080 pixel format (the format used by Blu-ray, HD-DVD, and some of the free to air high-definiton digital channels) never ever contains 1920xx1080 fully resolved lines of resolution, as explained earlier in this thread.

This is normal, standard, and par for the course.

Accordingly, the viewing distance charts would already take this characteristic into account [whether by explicit design, by empirical testing, or perhaps a bit of both; depending on who designed the chart].

Edited October 2, 2007 by MLXXX

[betty boop]

Short answer: the charts for viewing televison video already take the limitation into account.

Longer answer:

The viewing distance charts were principally developed based on principles of human vision, and by actual use have been confirmed to work.

really now, can we actually see some evidence to back up those statements

I find it quite interesting a chart would be developed for the whole reason to fully resolve 1920x1080 with a basis that you actually never can huh? do you understnd what you have even posted ?

Real world video in a 1920x1080 pixel format (the format used by Blu-ray, HD-DVD, and some of the free to air high-definiton digital channels) never ever contains 1920xx1080 fully resolved lines of resolution, as explained earlier in this thread.

This is normal, standard, and par for the course.

really now can we actually see some specific evidence from you to actually back that up, as we have seen nothing from you or anyone else justifying anything like that.

Accordingly, the viewing distance charts would already take this characteristic into account [whether by explicit design, by empirical testing, or perhaps a bit of both; depending on who designed the chart].

in other words, i'll give you the

short answer...

YOU Don't know !.

As you don't really know about any of the other statements you have made and are pushing as fact.

this is plain ridiculos mlx, I know the visual aquity charts are the holy of holy's, the santity of santity for owen, hell his whole world revolves around it !, but to shoot your self in the foot in trying to defend its premise, is a bizarre act !

by the way i'll be posting this and any of your answers to this in the viewing distance thread pinned up at the top of the forum for all to see.

[momaw]

so an inability of the human eye to fully resolve the resolution on a certain size screen at certain distances means the resolution is not in the source ?

If a tree falls in the forest and no-one is around to hear it, does it make a sound?

*Firmly tongue in cheek* ------>

No one is disputing that MTF is a very real phenomenon. Camera and lens designers go to great lengths to compensate for it. The problem is it is just not relevant to motion pictures on Blu-ray or HD DVD. It does become relevant if the source material is from the same or lower resolution - ie video source masters such as some old TV shows, but not for Studio Masters sourced from film or highdef digital cameras. Now as technology progresses and we get closer to bit for bit identical trandfers to the master, MTF will start to become more of an issue.

However, unless we all have imax at home I believe the human eye will have a greater limiting factor than MTF.

The original premise that Al and I have taken issue with is that 1920x1080 can not be fully resolved due to MTF issues. I think we have demonstrated that is not the case. In some circumstances (as illustrated above) it may be an issue. However, movies masters are of such a high resolution that when converted to blu-ray or HD DVD, these issues can be corrected for and removed. To do this care must be taken with the transfer process and admittedly, the proper care is not always taken. However, studios do put alot of effort into most releases.

[MLXXX]

so an inability of the human eye to fully resolve the resolution on a certain size screen at certain distances means the resolution is not in the source ? That statement is complex and I'll just comment on one aspect of it. Even if detail is not resolved at 100% MTF it may well still be visible.

If a tree falls in the forest and no-one is around to hear it, does it make a sound? Yes (!).

*Firmly tongue in cheek* ------>

No one is disputing that MTF is a very real phenomenon. Camera and lens designers go to great lengths to compensate for it. MTF drop-off can be reduced by good physical design, and partially compensated for by image sharpening.

The problem is it is just not relevant to motion pictures on Blu-ray or HD DVD. It does become relevant if the source material is from the same or lower resolution - ie video source masters such as some old TV shows, but not for Studio Masters sourced from film or highdef digital cameras. Now as technology progresses and we get closer to bit for bit identical trandfers to the master, MTF will start to become more of an issue. What you have written here I'd tend to go along with broadly. But although Owen raised doubt about a number of aspects of cinema quality, his principle issue was the MTF for the 1920x1080 format. And in particular whether that MTF is 100%. As he knew it was nowehere near 100% he knew he had a case. The trouble is that you, Momaw, and many others have concentrated your energies on defending the standards of cinematographic technology rather than focussing [no pun intended] on the specific issue of inter-pixel MTF of the 1920x1080 format.

However, unless we all have imax at home I believe the human eye will have a greater limiting factor than MTF.

The original premise that Al and I have taken issue with is that 1920x1080 can not be fully resolved due to MTF issues. I think we have demonstrated that is not the case. In some circumstances (as illustrated above) it may be an issue. However, movies masters are of such a high resolution that when converted to blu-ray or HD DVD, these issues can be corrected for and removed. To do this care must be taken with the transfer process and admittedly, the proper care is not always taken. However, studios do put alot of effort into most releases.

Mo,

I agree with most of what you have written in the post above. And I can see that you have thought about the issues involved.

Unfortunately, at its core, the point Owen was making was what could be called a pedantic technical point namely that the 1920x1080 format [like any sampling format] cannot have adjacent pixels with full contrast, if it is intended to represent video smoothly. And as a supplement to that, even if the need for filtering/anti-aliasing did not require that result, various technical limitations (in even the best lenses) would also contribute to some diminution in inter-pixel contrast.

In supporting Owen's very technical argument I have exposed myself to criticism, but that's life.

by the way i'll be posting this and any of your answers to this in the viewing distance thread pinned up at the top of the forum for all to see.

Please don't, Al. I think it would simply create confusion. It would have no positive benefit.

[Owen]

I have not been around for the last couple of days, so rather the respond to individual posts I shall contribute the following.

There is the MTF data for Kodak’s top of the range Vision 500T 35mm cinema film.

The test is under ideal conditions without the losses imposed by a camera and lens.

http://www.kodak.com/global/images/en/moti...ive/f0267ac.gif

Full info sheet here:

http://www.kodak.com/US/en/motion/products....20.4&lc=en

Kodaks best film has a 50% MTF point of about 40 cycles per mm (or line pairs per mm) for Red and 80 cycles per mm for Green and Blue.

For lower grade Kodak film 50% MTF value is obtained at 30 cycles per mm or line pairs per mm for Red and 50 cycles per mm for Blue and Green.

To convert this to something more meaningful to readers here we need to convert to lines per picture height, which is effectively the same as pixel resolution that people are used to seeing.

First we must multiply by 2 to get lines per mm, them multiply by the size of the active 35mm film frame.

35mm Cinema film used in anamorphic wide screen mode has an active area of about 22mm wide by 18mm high (at best)

http://en.wikipedia.org/wiki/35_mm_film

If we apply this to Kodaks best film under ideal conditions without a lens or camera, we get:

40x2x22=1720 lines across the frame for Red at 50% MTF

80x2x22=3520 lines across the frame for Green and Blue at 50% MTF

40x2x18=1440 lines up and down the frame for Red at 50% MTF

80x2x18=2880 lines up and down the frame for Green and Blue at 50% MTF

For the standard grade of Kodak film we get.

30x2x22=1320 lines across the frame for Red at 50% MTF

50x2x22=2200 lines across the frame for Green and Blue at 50% MTF

30x2x18=1080 lines up and down the frame for Red at 50% MTF

50x2x18=1800 lines across the frame for Green and Blue at 50% MTF

Remember that 50% MTF represents resolution that has only 50% of its original contrast, and is therefore not even close to “fully resolved”. Black and white adjacent lines will be displayed as low contrast grey lines.

To find out how cinema film behaves in practice when used in a typical Panavision camera with prime lenes we need to download and look at these articles. Copy and past URL’s below into a new browser.

www.cst.fr/IMG/pdf/35mm_resolution_english.pdf

www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

I linked to these in my original post but the link did not work. I have now fixed my original post.

The above .pdf white papers provide very interesting info on the REAL resolution of a Panavison cinema film camera. MTF graphs are in Lines per picture height, an expression used for both horizontal and vertical resolution. This is equivalent to pixel resolution that we are all accustomed to. Data is provided for original negative as well as all stages of duplication, right up to and including on screen cinema projection presentation.

For the original negative 50% MTF occurs at only 1000-1100 vertical lines across the frame and about 900-1000 horizontal lines up and down the frame.

At a horizontal resolution of 1920 line per picture height, cinema film has an MTF of only 10-15% which is effectively nill, and represent virtually no visible detail.

Vertical resolution is better, with an MTF of about 50-57% at 1080, which is good and very useable but not remotely close to fully resolved.

Remember that the above results are the combined effect of Red, Green and Blue,

the Kodak raw film data separates colour.

Also remember that when digitizing film at 4k resolution, there will be further loss of MTF at high special frequencies. When these digital masters are downscaled to 1920x1080 a further but smaller drop in MTF will occur. The affect of MFT is cumulative at every step in the chain.

Digital image compression will further degrade MTF and the sharpness and detail we are able to perceive on our display screens.

Excellent background info on MTF, film and lens can be read here:

http://www.normankoren.com/Tutorials/MTF1A.html

Nice little article about MTF and digital cameras here:

http://www.wyofoto.com/Digital_Image_Quali...age_Quality.htm

For those who did not read this link, I strongly suggest you do. It ties in cinema film with HD video. The comparison is very informative.

Copy and past into a new browser:

bssc.sel.sony.com/Professional/production/productsite/files/24PTechnicalSeminar2.pdf

The effects of MTF loss on digital cameras is quite consistent, no matter if it is a still, video or a digital cinema camera like Genesis, the same governing laws apply. The higher the number of pixels the higher the Nyquist limit, and the higher the 50% MTF point. It’s pretty much a linear relationship, and film digitising equipment has the same limitations.

It’s therefore easy to extrapolate the MTF curve of different digital capture devises with higher pixel counts, although lens performance becomes increasingly important at high spatial frequencies.

The Sony article also explains that the area under the MTF curve is what dictates perceived image sharpness. This is why digital images usually look shaper then film, even though the film has higher absolute detail.

1920x1080 images with 100% MTF out to the resolution limit are not possible from any source other then PC generated images, which are always perfectly resolved with 100% MTF. However if the image is scaled, MTF will again be degraded to less then 100%

Displays also have an MTF curve, which explains why one display can look sharper then another, or why displays of identical pixel count can look sharper or softer then one another.

I have never said that no visible detail is visible at 1920x1080 from the best sources, I have only ever argued that this low level detail is in no way “fully resolved”, which would require 100% MTF at the resolution limit both vertically and horizontally.

I keep using the term “fully resolved at 100% MTF” but this has been continually misunderstood or ignored.

[momaw]

oh Owen, if you are going to keep popping in after not partaking in the discussion at least have the courtesy to review the material in the thread before posting more of the same. I would love to see you post your theory in the "Industry Insiders" thread over at AVS, but somehow I don't see that happening.

[Owen]

I have read all posts.

If you have any constructive comment to make on the evidence I have provided please do so.

If you have any hard data such as MTF graphs to refute my evidence or the Kodak data sheets I would be interested to see it.

[betty boop]

I have read all posts.

If you have any constructive comment to make on the evidence I have provided please do so.

If you have any hard data such as MTF graphs to refute my evidence or the Kodak data sheets I would be interested to see it.

[MLXXX]

[This is not in response to posts #86, #87 and #89.]

Owen, you've assembled some highy relevant and informative reference material!

If anyone would prefer to just read one of the links Owen has provided, I'd suggest this one: bssc.sel.sony.com/Professional/production/productsite/files/24PTechnicalSeminar2.pdf

It starts with digital cameras and then moves on to film cameras, and compares the two. Fascinating stuff. Pages 11 and 12 show the MTF of low and high speed film. The MTF for green [and blue] is much better than for red.

If after reading that article people still cannot understand or accept the propositions Owen has been presenting, perhaps we should all call it quits. I know I cannot present the concepts more eloquently than that article.

Slightly off-topic

I think that if Hollywood wants to continue to use film for capture of movies, and yet wants to establish an edge in the high definition stakes for videophiles, Hollywood will need to consider film cameras that capture onto stock bigger than the standard 35mm size, appropriate lenses, and editing in high pixel count formats. The extra definition may not be all that useful now, but could be useful a few years into the future when the consumer videophile high definition standard evolves beyond 1920x1080.

Edited October 3, 2007 by MLXXX

[momaw]

I have read all posts.

If you have any constructive comment to make on the evidence I have provided please do so.

If you have any hard data such as MTF graphs to refute my evidence or the Kodak data sheets I would be interested to see it.

I take it you are not interested in presenting this evidence to the experts as suggested?

[Hosko]

I have read all posts.

If you have any constructive comment to make on the evidence I have provided please do so.

If you have any hard data such as MTF graphs to refute my evidence or the Kodak data sheets I would be interested to see it.

It would appear that your correct about Kodak film, you say the MTF is 50% for film which would be scanned in at 4096p. That means that when its encoded to 1080p every forth pixel would be used horizontally. It really doesn't matter then that adjacent pixels can't have 100% contrast because adjacent pixels will never make it to the 1080p transfer. Sure the full res master might suffer from this but not the 1080p

Edited October 3, 2007 by Hosko

[Owen]

Direct them here and tell them to bring their evidence.

So far you have yet to bring anything useful to the discussion, unless you consider bad attitude to be constructive.

[betty boop]

~

Slightly off-topic

I think that if Hollywood wants to continue to use film for capture of movies, and yet wants to establish an edge in the high definition stakes for videophiles, Hollywood will need to consider film cameras that capture onto stock bigger than the standard 35mm size, appropriate lenses, and editing in high data formats. The extra definition may not be all that useful now, but could be useful a few years into the future when the consumer videophile high definition standard evolves beyond 1920x1080.

hate to burst you bubble again, but I cant see it happening anytime soon, and reading your post I think you know the reason why

[Owen]

It would appear that your correct about Kodak film, you say the MTF is 50% for film which would be scanned in at 4096p. That means that when its encoded to 1080p every forth pixel would be used horizontally. It really doesn't matter then that adjacent pixels can't have 100% contrast because adjacent pixels will never make it to the 1080p transfer. Sure the full res master might suffer from this but not the 1080p

Scaling takes ALL data and interpolates a new image, no data is ignored.

A 4k image downscaled will have more detail and potentially higher MTF then an image scaned at 2k.

For digital imaging you must sample at no less the double the output resolution required, and even then you cant have an image with 100% MTF.

[betty boop]

It would appear that your correct about Kodak film, you say the MTF is 50% for film which would be scanned in at 4096p. That means that when its encoded to 1080p every forth pixel would be used horizontally. It really doesn't matter then that adjacent pixels can't have 100% contrast because adjacent pixels will never make it to the 1080p transfer. Sure the full res master might suffer from this but not the 1080p

yes quite right, one of my very early posts and links said that..

http://www.cintel.co.uk/dlfiles/techdocs/E...0Resolution.pdf

Firstly, the film format, 35mm can accomodate 6K resolution. Secondly the quality of taking lens has a dramatic effect on resolution. While today's zoom lenses are extremely good, the ultimate quality can often only be achieved with prime lenses. Good prime lens can acheive a solid 4K on the film and still resolve some information at 6K.

as it says 4K is a walk in the park and still resolve as high as 6K.

and really we have not seen any evidence as such to disprove that or the capabilities of modern day cameras capabilities, I listed a whole host of them includign the genesis and theres been nothing really said in regards CG sources.

I'd like to see that post owen, over on the insiders forum of avs...

oh and by the way we're still waiting on the massive contradictions on the little matter of the visual aquity charts which according to mlx are now not designed to fully resolve 1920x1080

[momaw]

Direct them here and tell them to bring their evidence.

So far you have yet to bring anything useful to the discussion, unless you consider bad attitude to be constructive.

Guess that's a no then.

I don't have a bad attitude. You just don't have a valid argument or any relevant evidence. But hey if that equals bad attitude on my part......

It would appear that your correct about Kodak film, you say the MTF is 50% for film which would be scanned in at 4096p. That means that when its encoded to 1080p every forth pixel would be used horizontally. It really doesn't matter then that adjacent pixels can't have 100% contrast because adjacent pixels will never make it to the 1080p transfer. Sure the full res master might suffer from this but not the 1080p

Thank you.

Scaling takes ALL data and interpolates a new image, no data is ignored.

A 4k image downscaled will have more detail and potentially higher MTF then an image scaned at 2k.

For digital imaging you must sample at no less the double the output resolution required, and even then you cant have an image with 100% MTF.

Oh my. Clearly you need to educate yourself on these processes.

[Owen]

yes well one of my very early posts and links said that..

http://www.cintel.co.uk/dlfiles/techdocs/E...0Resolution.pdf

as it says 4K is a walk in the park and still resolve as high as 6K.

The problem is the MTF at that solution. Please provide just one example of any camera that can provide 100% MTF at 1920 lines per picture width or 1080 lines per picture height.

and really we have not seen any evidence as such to disprove that or the capabilities of modern day cameras capabilities, I listed a whole host of them includign the genesis and theres been nothing really said in regards CG sources.

You obviously did not read my links, they provide ample evidance to suppot my case.

I have always maintained that the ONLY way to get 100% MTF at 1902x1080 is with computer generated images, (provided they have not been scaled). Unless we are only viewing CGI only content, 100% MTF images are not posible.

oh and by the way we're still waiting on the massive contradictions on the little matter of the visual aquity charts which according to mlx are now not designed to fully resolve 1920x1080

The Visual acuity model on which viewing distance calculators are based have proved to be very accurate in my experience with 1:1 pixel mapped computer generated test paterns. They are based on perfect video source with 100% MTF at the resolution specified, a situation that cannot exist in practice for none computer generated content.

Video never has the resolution its pixel count would suggest, so longer viewing distances can be use without perceptible loss of visible resolution for video sources.

[Owen]

Guess that's a no then.

I don't have a bad attitude. You just don't have a valid argument or any relevant evidence. But hey if that equals bad attitude on my part......

Thank you.

Oh my. Clearly you need to educate yourself on these processes.

You have not presented a shred of hard evidence to refute my claim that 1920 line per picture width by 1080 lines per picture height images with 100% MTF can be created outside a computer.

If you don’t believe any of the information I have linked, what point is there in any further discussion?

[momaw]

You have not presented a shred of hard evidence to refute my claim that 1920 line per picture width by 1080 lines per picture height images with 100% MTF can be created outside a computer.

If you don’t believe any of the information I have linked, what point is there in any further discussion?

You haven't presented a shred of evidence relating to MTF and HD DVD and Blu-ray 1920x1080 transfers from high resolution studio masters for me to refute

The Visual acuity model on which viewing distance calculators are based have proved to be very accurate in my experience.....

Bingo!!!! We have a winner folks. You are passing personal subjective experience off as fact. Is there an emoticon for shaking the finger?

[Owen]

The film negative from a Panavisoon camera cant provide 1920 lines per picture width by 1080 lines per picture height at 100% MTF, not even close, so how on earth is a HDDVD or BluRay disk mastered from that negative going to achieve it?

[Owen]

Bingo!!!! We have a winner folks. You are passing personal subjective experience off as fact. Is there an emoticon for shaking the finger?

Do you have any evidence to suggest that the human visual acuity model is inaccurate?