16mm generational loss?

[Filip Plesha]

Here is a nice little Kodak article that covers this:

 

http://www.kodak.com/US/en/motion/students...0.1.4.9.6&lc=en

 

 

Maybe I'm just not good at explaining things.

 

 

Here is one on the MTF:

http://www.kodak.com/US/en/motion/students...c=en#modulation

 

 

 

Here is the whole section, it contains most of what you need to begin understanding analog

photographic materials and how they work in real life:

 

http://www.kodak.com/US/en/motion/students....4.9.6.10&lc=en

 

 

good luck, hope you learn something new today

 

 

p.s.

 

Now I would have to do some maths for the lens using the rayleigh limit?

 

no you don't have to use that formula, I did a simplified math for you and gave you specific

diffraction limits for specific apertures. Either use that list, or

use the quick simplified calculation: divide 1500 with your aperture and you get the diffraction

limit for that aperture

 

Example:

 

f/4....

1500/4 = 375

 

375 is your maximum resolving power at f/4 for any perfect lens.

Edited March 3, 2006 by Filip Plesha

[Mark Williams]

Thanks Filip Although I doubt I will fully understand this I do get the drift and it looks to me as if super 16mm fits in about right with the HDV Standard.. Ordinary 16mm masked for 16:9 doesnt..

 

Also, there is actually a noticeable loss with each Film generation...

[John Pytlak RIP]

Another good Kodak tutorial on sensitometry and image structure:

 

http://www.kodak.com/US/en/motion/support/...uctureP.shtml#p

[David Mullen ASC]

Thanks Filip Although I doubt I will fully understand this I do get the drift and it looks to me as if super 16mm fits in about right with the HDV Standard.. Ordinary 16mm masked for 16:9 doesnt..

 

Also, there is actually a noticeable loss with each Film generation...

 

Hence the danger of just looking at charts and numbers, because you're misunderstanding the issues. I can't even begin to explain it to you, but the phrase "apples and oranges" comes to mind. Plus your conclusions don't really have any real-world relevance if you're suggesting the cropped regular 16mm fails to meet the "HDV standard". That doesn't even make sense.

[Mark Williams]

Hence the danger of just looking at charts and numbers, because you're misunderstanding the issues. I can't even begin to explain it to you, but the phrase "apples and oranges" comes to mind. Plus your conclusions don't really have any real-world relevance if you're suggesting the cropped regular 16mm fails to meet the "HDV standard". That doesn't even make sense.

You know Im sure your right david, I agree line resolution and film dont form part of the equation anyway..

 

But other things would like grain etc..

 

Im thinking that HDTV AND HDV uses 1080 lines? Super 16 only resolves 1k to 1.5k when 3 generations are used? so where does that leave standard 16mm?

[Filip Plesha]

You just did an unfair comparison...

 

You said Hd has 1080 lines and 4th generation super16 has 1K to 1.5K resolution.

IF you print HD you will probably print it to a negative or IP (like on Star Wars) and then

make further "preprint" copying

So your HD footage won't be as sharp as it was on tape. Your HD footage won't show 1080 clearly

visible lines at original contrast, it will lose definition just like film, so again you will have less

resolution than if you started with high-end super16

 

 

The popular thinking seens to be that as long as the copy medium outresolves the original, it wont degrade the image.

That's not true, even a 5000 lp/mm copy medium degrades the image, or 5000000 lp/mm.

So, while you may not lose valuable image information when printing negatives with recorded video information,

you are loosing detail defitinion (response) when copying.

 

You can't compare HD on a monitor to a 4th generation super16 print. That's not fair.

 

Thanks Filip Although I doubt I will fully understand this I do get the drift and it looks to me as if super 16mm fits in about right with the HDV Standard.. Ordinary 16mm masked for 16:9 doesnt..

 

Also, there is actually a noticeable loss with each Film generation...

 

 

I only made digital example equivalents of resolution, because that's the way your mind seems to work.

In reality, you can't just say, ok certain lp/mm equals certain pixel density, because it's like David says apples and oranges. You can use pixels to simulate analogue resolving power, the question is how low can you go in pixel density when sampling a certain film resolution and get away with it.

You can sample 50lp/mm film images at 4000dpi or 10 000 dpi. With 10 000 lp/mm it will be closer to how film looks like when magnified, with 4000 dpi it will be just on the boarder of distinguishing lines among pixels.

It's really relative what you consider to be the digital equivalent of a given analog resolution.

Edited March 3, 2006 by Filip Plesha

[John Pytlak RIP]

Super 16 only resolves 1k to 1.5k when 3 generations are used? so where does that leave standard 16mm?

 

Almost all 16mm and Super-16 used for HD production today has the camera original as the scanned element. No generational loss due to printing or intermediates.

 

Film grain (which can have implications for compression efficiency if excessive) is a controllable factor, even with the higher speed films (i.e., don't underexpose). And if you have the light, the new lower speed stocks are very low in graininess.

[Mark Williams]

OK I dont mean to frustrate through my lack of understanding here I may appear backward in this regard I dont even know the lingo its all NEW to me.. Sorry... I accept though that if I want to be a DOP This is an area I will have to learn at some point..

 

SO HOW does a fourth generation 16mm film exposed using a standard gate and framed for 16:9 stand up to THE HD standard of 1080i or 720p would I lose anything? Compared to super 16 The public are facing a digital revolution where all TVs are going to become HDTV As well as DVDS being released in this format so DV Camcorders will be to..

 

HOW does my Camera (standard 16 framed for 16:9) after four generations stand up against This new format..

[John Pytlak RIP]

OK I dont mean to frustrate through my lack of understanding here I may appear backward in this regard I dont even know the lingo its all NEW to me.. Sorry... I accept though that if I want to be a DOP This is an area I will have to learn at some point..

 

SO HOW does a fourth generation 16mm film exposed using a standard gate and framed for 16:9 stand up to THE HD standard of 1080i or 720p would I lose anything? Compared to super 16 The public are facing a digital revolution where all TVs are going to become HDTV As well as DVDS being released in this format so DV Camcorders will be to..

 

HOW does my Camera (standard 16 framed for 16:9) after four generations stand up against This new format..

 

Why would you go through four generations today? In almost all cases your camera original is the element transferred/scanned. No generational loss. At worst, you might cut your original as A/B rolls and make a timed master positive for transfer and as a preservation element.

[Mark Williams]

Why would you go through four generations today? In almost all cases your camera original is the element transferred/scanned. No generational loss. At worst, you might cut your original as A/B rolls and make a timed master positive for transfer and as a preservation element.

Well I guess if you were finishing on film then you could.. Although most would have it transferred to 35mm for this process and would undoubtably have filmed on Super 16mm

 

I wanted to understand what My camera can do is capable off and my limitations.. If possible to perhaps make ADVERTS Or Music videos originating on film.. I am not ready yet to commit to a big investment although I have spent quite a lot already I want to see what I can do and to learn as much as possible before this.. :) Obviously anything I do I want it to be good enough for broadcast..

[Filip Plesha]

OK I dont mean to frustrate through my lack of understanding here I may appear backward in this regard I dont even know the lingo its all NEW to me.. Sorry... I accept though that if I want to be a DOP This is an area I will have to learn at some point..

 

SO HOW does a fourth generation 16mm film exposed using a standard gate and framed for 16:9 stand up to THE HD standard of 1080i or 720p would I lose anything? Compared to super 16 The public are facing a digital revolution where all TVs are going to become HDTV As well as DVDS being released in this format so DV Camcorders will be to..

 

HOW does my Camera (standard 16 framed for 16:9) after four generations stand up against This new format..

 

Why do you even ask how would a film print compare to HD footage on a monitor?

Are you worried about the impression people are going to have seeing your print after comming to cinema from their livingrooms with HD TV's?

Isn't that a concern for cinema owners more than for a filmmaker?

 

It makes sense to compare two image capturing methods in the same medium (print or TV for example)

but comparing one method in one medium vs. another method in a nother medium is trully apples and oranges.

 

How can one even begin to answer a question like: what's better HD on a monitor or super16 on a print?

It's like, what is better, eating apples when you feel like eating oranges, or eating potatoes when you feel like eating sandwitches?

 

 

 

If you'd like to know which would prefer better on HD, then it makes sense. In that case

you wouldn't be using 4 generations of copying, but either the original negative or an IP, and in that

case you would have over 3K woth of image information to work with with a slow speed film, and

then you could start comparing HD footage with a HD scan of super16.

[Filip Plesha]

Here is how 5245 in 35mm looks at 2K, which is like 1K scanning of super16:

 

 

 

It is obvious that this frame would benefit from AT LEAST 4 times the resolution scanning (4K)

because it appears pixel to pixel sharp at 2K.

 

I suspect 5248 is not far away from this quality. And new vision2 films are probably even better.

[Leo Anthony Vale]

John I looked at some old 8mm footage through an old eumig projector viewscreen which is about the size of a small TV.. I couldnt believe the Quality compared to DV... SO SUPERIOR.. I also think its on par or better than HDV... I dont understand? because 16mm Must be 4 times better? So why do some say in order to get a good HDV TRANSFER the old 16mm cameras converted to super are ideal? SURELY all 16mm cameras must surpass HDV BY a wide margin..

 

The new HD TVs coming out in this country and soon to be broadcast HD.. ARE still no match for standard 16mm Even if it has gone through 4 generations? Of course thanks to new info here I now know there is differences in film stocks performance and also in lens ability etc..

 

Would I be right in thinking my standard 16mm Camera still qualifies me to make broadcast quality HD footage even when its masked for 16:9 or even 2:35:1 and been through 4 generations?

 

---Once your film is transfered to a video format, the picture will never look better than the video format it's been transfered to. Think of all those 16mm segments in old BBC shows like Monty Python. & those are transfered from a print from the OCN, not a 4th generation.

 

Sure stocks have improved since then, but there is a generational loss even for sharpness. David was saying that the less isn't much in 35mm. But the smaller the format the greater the loss. Super8 prints are hopeless. & transfering from a print give contrast problems and blocked shadow detail.

 

The larger the original format and the fewer generations the better the transfer.

[John Pytlak RIP]

Prints are rarely used for making transfers anymore, if the original or a pre-print element is available. If you are shooting 16mm or Super-16 for HD transfer, the camera original is almost always the scanned element.

[Mark Williams]

---Once your film is transfered to a video format, the picture will never look better than the video format it's been transfered to. Think of all those 16mm segments in old BBC shows like Monty Python. & those are transfered from a print from the OCN, not a 4th generation.

 

Sure stocks have improved since then, but there is a generational loss even for sharpness. David was saying that the less isn't much in 35mm. But the smaller the format the greater the loss. Super8 prints are hopeless. & transfering from a print give contrast problems and blocked shadow detail.

 

The larger the original format and the fewer generations the better the transfer.

Thanks leo.. I agree with you!

[Mark Williams]

Why do you even ask how would a film print compare to HD footage on a monitor?

Are you worried about the impression people are going to have seeing your print after comming to cinema from their livingrooms with HD TV's?

Isn't that a concern for cinema owners more than for a filmmaker?

What I actually said was

Im thinking that HDTV AND HDV uses 1080 lines? Super 16 only resolves 1k to 1.5k when 3 generations are used? so where does that leave standard 16mm?

 

That was why I asked I simply wanted to know HOW my standard 16mm footage with a 16:9 mask stacks up against HD Although generation loss as explained by John doesnt matter because footage destined for telecine would be done at the neg stage I can go on get my film timed colur corrected printer lights all at the lab and when Im happy with the result I can go do the whole thing all over again in post on video..

 

QUOTE

It makes sense to compare two image capturing methods in the same medium (print or TV for example)

but comparing one method in one medium vs. another method in a nother medium is trully apples and oranges.

--------------------------------------------------------------------------------

wELL no not really not if your converting one medium to another you need to know what your likely to get

-------------------------------------------------------------------------------

QUOTE

How can one even begin to answer a question like: what's better HD on a monitor or super16 on a print?

It's like, what is better, eating apples when you feel like eating oranges, or eating potatoes when you feel like eating sandwitches?

--------------------------------------------------------------------------------

yES But and as you yourself have tried to do you have to come up with a test a plan to try to understand because you have to get things right for the end product

-----------------------------------------------------------------------------

QUOTE

 

If you'd like to know which would prefer better on HD, then it makes sense. In that case

you wouldn't be using 4 generations of copying, but either the original negative or an IP, and in that

case you would have over 3K woth of image information to work with with a slow speed film, and

then you could start comparing HD footage with a HD scan of super16.

--------------------------------------------------------------------

My Plan and it seems was wrong was to do the colour correction timing and lights on film because that is the medium I want to learn and work with when My film was done I could have it telecined BUT also to have the telecine done at the neg stage too because I can colour correct and mess about with it on digital I wanted both I want to learn both Experiment and I have a suspicion that film correction is probably a whole lot better quality wise than video correction..

[Filip Plesha]

As far as I understand you want to finish the film on film and then have two options: telecine and print

 

Well in that case your "end product" is the master positive (or IP)

But still, even with a color corrected IP you have to do some color correcting when scanning film, at least to match the scan to the element that was being scanned.

 

 

But anyway, doing the same kind of math, with the same film stocks used, your IP

would have about 75 lp/mm maximum, which is about 62% of the original negative resolution.

For a 12.5mm image of super16, that is around 1875 individual lines resolved.

For regular 16mm (10.26mm) it is about 1540 lines resolved

Since digital has problems with aliasing and moirre due to its pixelated nature, and digital images are interpolated, it takes more than 2000 pure full RGB pixels to match such resolving power of super16 (around 1900 lines) and full RGB HD resolution to match the resolving power of regular 16mm (more than 1500 lines resolved)

 

So there you have it, at best, your 16mm camera can match a really good HD camera performing to the limits of its format.

 

But that's what you get if you are a densiometer, In case you are a human viewer, it may look

better or worse to you than HD footage depending on what your preferences are.

 

Even though with sharpening you could fill up the HD frame with adequatly sharp detail with your IP scanned, the grain would still make it look different than HD footage. Now the question is, do your viewers like smooth images of HD or grainy images of super16? That's up to them.

 

Since you would be operating on the edge by using IP as your source, it would be better to

use the original negative, the image would be sharper and would better compete against HD cameras.

You would start with about 120 lp/mm, which is quite an advantage.

 

And as for projection, well that's another story. HD footage would be degraded too by copying from one piece of film to another. In fact to some degree it has to be degraded otherwise you'd see pixel edges in the print, the pixels have to be smoothened out by analog copy degradation.

The resolution of SD video on the other hand is so low that you can't even lose enough sharpness through generations

to lose the pixelated edges, so they still appear in print.

 

I've seen a documentary in cinema which was a combination of DV and archive super8 and some 16mm.

The ironic part is that super8 footage looked smoother and more natural than DV , exept for the

grain which was huge.

Edited March 3, 2006 by Filip Plesha

[Mark Williams]

As far as I understand you want to finish the film on film and then have two options: telecine and print

 

Well in that case your "end product" is the master positive (or IP)

But still, even with a color corrected IP you have to do some color correcting when scanning film, at least to match the scan to the element that was being scanned.

But anyway, doing the same kind of math, with the same film stocks used, your IP

would have about 75 lp/mm maximum, which is about 62% of the original negative resolution.

For a 12.5mm image of super16, that is around 1875 individual lines resolved.

For regular 16mm (10.26mm) it is about 1540 lines resolved

Since digital has problems with aliasing and moirre due to its pixelated nature, and digital images are interpolated, it takes more than 2000 pure full RGB pixels to match such resolving power of super16 (around 1900 lines) and full RGB HD resolution to match the resolving power of regular 16mm (more than 1500 lines resolved)

 

So there you have it, at best, your 16mm camera can match a really good HD camera performing to the limits of its format.

 

But that's what you get if you are a densiometer, In case you are a human viewer, it may look

better or worse to you than HD footage depending on what your preferences are.

 

Even though with sharpening you could fill up the HD frame with adequatly sharp detail with your IP scanned, the grain would still make it look different than HD footage. Now the question is, do your viewers like smooth images of HD or grainy images of super16? That's up to them.

 

Since you would be operating on the edge by using IP as your source, it would be better to

use the original negative, the image would be sharper and be a better compete against HD cameras.

You would start with about 120 lp/mm, which is quite an advantage.

 

And as for projection, well that's another story. HD footage would be degraded too by copying from one piece of film to another. In fact to some degree it has to be degraded otherwise you'd see pixel edges in the print, the pixels have to be smoothened out by analog copy degradation.

 

AT LAST AT long last I have the answer I was looking for. That is definative David Mullen told me and you provided the technical bit

 

Thank you

 

Its been quite rewarding if a little hard going

 

I will type out your answer Filip and I will put it in my film folder.. :D

[Scot McPhie]

I think the thing is Mark if people are watching your film and worrying about if it was colour corrected in a digital intermediate or purely in film then you've got other major issues you've overlooked.

 

I've sat on the fence here but now I've decided I don't understand what your point is?? There's alot of different post production routes you can take, done properly they will all be fine - splitting hairs over them ultimately becomes unproductive and a distraction from actual film making.

 

 

Scot

[Tim Carroll]

I simply wanted to know HOW my standard 16mm footage with a 16:9 mask stacks up against HD Although generation loss as explained by John doesnt matter because footage destined for telecine would be done at the neg stage

 

Just finished a test and here are a couple of frame grabs of what Black & White regular 16mm footage looks like when scanned 16:9 anamorphic SD and stored uncompressed, and when it is scanned HD and stored uncompressed. There is also a video clip of the SD footage.

 

Frame grab Regular 16mm scanned SD Anamorphic 16:9, stored uncompressed.

 

Frame grab Regular 16mm scanned HD, stored uncompressed.

 

Short clip of Regular 16mm scanned SD Anamorphic 16:9, stored uncompressed.

 

The two Frame grab jpgs were saved at Maximum quality in Photoshop (so they are compressed now), and the clip was compressed for the internet with the Sorenson Video 3 Pro codec with a data rate of 2500 kbits/sec.

 

-Tim

[Stephen Williams]

AT LAST AT long last I have the answer I was looking for. That is definative David Mullen told me and you provided the technical bit

 

Thank you

 

Its been quite rewarding if a little hard going

 

I will type out your answer Filip and I will put it in my film folder.. :D

 

Mark,

 

So you see there is no need for Ultra16! LOL

 

Cheers

 

Stephen

[Filip Plesha]

AT LAST AT long last I have the answer I was looking for. That is definative David Mullen told me and you provided the technical bit

 

Thank you

 

Its been quite rewarding if a little hard going

 

I will type out your answer Filip and I will put it in my film folder.. biggrin.gif

 

well it has been a pleasure

 

 

Mind one more thing. All of those numbers are reffering to high contrast targets (1:1000 contrast)

Which means regular contrast subjects would appear less sharp.

To pull out such resolution from your 16mm system, your scenes would have to have some

features that would enhance percieved sharpness, like higher contrast lighting.

 

 

 

 

 

 

 

Just finished a test and here are a couple of frame grabs of what Black & White regular 16mm footage looks like when scanned 16:9 anamorphic SD and stored uncompressed, and when it is scanned HD and stored uncompressed. There is also a video clip of the SD footage.

 

Frame grab Regular 16mm scanned SD Anamorphic 16:9, stored uncompressed.

 

Frame grab Regular 16mm scanned HD, stored uncompressed.

 

Short clip of Regular 16mm scanned SD Anamorphic 16:9, stored uncompressed.

 

The two Frame grab jpgs were saved at Maximum quality in Photoshop (so they are compressed now), and the clip was compressed for the internet with the Sorenson Video 3 Pro codec with a data rate of 2500 kbits/sec.

 

-Tim

 

great examples

Like incarnation of the numbers above.

 

5248 would probably appear finer in grain

[Mark Williams]

I think the thing is Mark if people are watching your film and worrying about if it was colour corrected in a digital intermediate or purely in film then you've got other major issues you've overlooked.

 

I've sat on the fence here but now I've decided I don't understand what your point is?? There's alot of different post production routes you can take, done properly they will all be fine - splitting hairs over them ultimately becomes unproductive and a distraction from actual film making.

Scot

Well the first thing Im going to do is chill because I know my camera is good for any format I choose even 35mm Then Im going to make a film... NOW which route should I go? I have an educated choice! I know for fun I will try them all! and put into practice what I have read my fav directors do.. oF COURSE I will try and be cost effective and of course I will make mistakes.. I know one thing though I will have fun learning and If I get stuck... Well..... ;)

 

Big THANKS Everyone :D

Edited March 3, 2006 by Mark Williams

[Filip Plesha]

"Well the first thing Im going to do is chill because I know my camera is good for any format I choose"

 

 

You didn't need to ask anyone to know that. Wheather some film format is good enough or not for some exibition conditions is totally up to you as the author and the esthetical preferences of the audience.

It's not something technical in nature, but psychological.

It also depends on trends. Perhapse some day in the future when HD is filled with clean clinical images of HD, it will be a trend to use super8 footage for music videos commercial or even films, who knows.

Or maybe tomorow everyone starts asking for grainless images, in which case even fast 35mm will be

considered inadequate for HD. Like I said it all depends on what people prefer.

As you can see from those screenshots, technically the detail is there in 16mm, with some sharpening it would match output from an HD camera easily, not to mention using a pin registered film scanner, which would extract even more detail from 16mm, but the grain is a thing of preference. Some see it as unwanted noise, some see it as artistic texture.

Now if your idea is to make a kind of a gritty rough looking film, you can use 16mm even in 35mm print, but

if your idea is to get clean images, then it won't work.

"Good enough" depends on your definition of "good"

 

For me as a viewer, 16mm footage is very interesting to look at because the colors look different than from 35mm film due to higher enlargement. They seem somehow more rugged than in larger formats, like rock vs. smooth subtle classical music.

But then again I also like super8 because it looks even grittier, exept that I wouldn't always like to watch a 2 hour movie in super8, that could get somewhat distracting. Super8 would be better for short high-visual-impact stuff like ads or music videos.

 

But I know how you feel, you'd like to have options even though you may not need them.

Same reason I'd like to go to medium format photography I may not need the extra resolution, but It's nice to know it's there and I can use it if I need it, plus it makes you feel more professional.

But it's a weakness, what's important is to make images as good as you can with best tools that you can affoard.

[David Mullen ASC]

Film would be transferred from the original negative in most cases for a conversion to HDTV, or at worst, a timed IP made off of the original negative. HD photography, conversely, may have to go through the same IP/IN steps to make release prints as something shot on film. So comparing a 16mm blow-up to 35mm through an IP/IN process to an HD image shown on an HDTV monitor is completely apples & oranges.

 

You either have to compare both seen on an HDTV monitor or both seen in a 35mm print projected.

 

Assuming you use good lenses on your 16mm camera, the problem isn't resolution compared to HD photography, it's GRAIN. Some future HD broadcasters may object to the amount of grain in your image, especially if it's 16mm cropped to 1.78 (16x9). Super-16 has the advantage of being slightly wider and better suited to 16x9 transfer, so there is less grain (and more resolution over cropped 16mm, but still, cropped 16mm should rival HD for resolution still assuming it is shot with sharp lenses.)

 

So while I can't guarantee that some HDTV market will always buy your cropped 16mm material, because some are quite prejudiced against grain, the issue isn't really resolution per se unless you are using old lenses or are shooting wide-open. But it may mean restricting yourself to slower-speed film stocks if you're worried.

 

But in terms of resolution loss, just transfer to HD from the original negative.