Jump to content


Photo

Megapixels


  • Please log in to reply
25 replies to this topic

#1 hugovillasmythe

hugovillasmythe
  • Basic Members
  • PipPip
  • 24 posts

Posted 11 January 2006 - 03:44 PM

I was and always will be a nikon man but my photo equipment had stayed virtualy untouched for a good 5 years. All of it is compatible with the Nikon D-50 and D-70. I've already checked. I'm no longer making a living out of still photo but I enjoy the odd assignment every now and then and doing my own portraits ans other still work.
How many megapixels does a standard 35mm SLR camera negative produces. I just want to compare.
I know that moving into the digital domain implies printers (and the endless supply of cartridges) and a good handle on photoshop or similar but I was very happy with the results I was getting from my old nikon gear. Should I go with the big boy N-70 or the "barley pro" N-50? Or even worse... should I start thinking about the N-80?
  • 0

#2 John Pytlak RIP

John Pytlak RIP

    (deceased)

  • Sustaining Members
  • 3499 posts
  • Industry Rep
  • Rochester, NY 14650-1922

Posted 11 January 2006 - 04:01 PM

I was and always will be a nikon man but my photo equipment had stayed virtualy untouched for a good 5 years. All of it is compatible with the Nikon D-50 and D-70. I've already checked. I'm no longer making a living out of still photo but I enjoy the odd assignment every now and then and doing my own portraits ans other still work.
How many megapixels does a standard 35mm SLR camera negative produces. I just want to compare.
I know that moving into the digital domain implies printers (and the endless supply of cartridges) and a good handle on photoshop or similar but I was very happy with the results I was getting from my old nikon gear. Should I go with the big boy N-70 or the "barley pro" N-50? Or even worse... should I start thinking about the N-80?


A topic that's been debated here for years. Since film is an analog media, with randomly dispersed and sized silver halide grains as "sensors", "megapixel" is normally not considered a film parameter. In actuality, if each grain is considered a "pixel", you are probably talking "gigapixels" per frame. ;)

A few years ago, some Kodak motion picture film ads suggested a 35mm movie frame had the equivalent of 12 megapixels --- this simplification was based on an assumption of needing to scan a 4:3 aspect ratio Super-35 full frame at 4K to get most of the important visual information from that frame. So a still frame 35mm SLR image would be 24 megapixels as a conservative approximation.
  • 0

#3 Dickson Sorensen

Dickson Sorensen
  • Basic Members
  • PipPipPip
  • 131 posts
  • Cinematographer

Posted 11 January 2006 - 05:19 PM

A few years ago, some Kodak motion picture film ads suggested a 35mm movie frame had the equivalent of 12 megapixels


Having worked for years in advertizing I am always a little suspicious of something I read in an ad, even from the venerable Kodak. One thing to also remember in comparing apples to oranges is that in addition to the number of pixels there is bit depth. That is the description of how many shades there are in each color of each pixel. In film as (I understand the difference) the bit depth would be in theory infinite or am I now comparing apples to grapefruit?

Edited by Dickson Sorensen, 11 January 2006 - 05:21 PM.

  • 0

#4 Jon Allen

Jon Allen
  • Basic Members
  • PipPipPip
  • 227 posts
  • Other
  • USA

Posted 11 January 2006 - 05:55 PM

A topic that's been debated here for years. Since film is an analog media, with randomly dispersed and sized silver halide grains as "sensors", "megapixel" is normally not considered a film parameter. In actuality, if each grain is considered a "pixel", you are probably talking "gigapixels" per frame. ;)

A few years ago, some Kodak motion picture film ads suggested a 35mm movie frame had the equivalent of 12 megapixels --- this simplification was based on an assumption of needing to scan a 4:3 aspect ratio Super-35 full frame at 4K to get most of the important visual information from that frame. So a still frame 35mm SLR image would be 24 megapixels as a conservative approximation.


Hard to compare digital sensors with actual pixels to photo-chemical emulsion. I read in Popular Mechanics/Science the other month that there was some guy using 10+ in film stock and an old spy plane camera and was taking pictures which he claimed were the equivalent of 4gigapixies. I was extremely sceptical, I mean, you can always scan a print at a higher res, it becomes a point of how much people will notice vs cost involved in producing material of such high magnification.

But yes, it becomes a point of comparing apples to acorns after a while. This whole debate is rather old, obviously film is better as a standard, yet the cost of digital video provides an acceptable medium at a significantly lower cost.
  • 0

#5 dd3stp233

dd3stp233
  • Basic Members
  • PipPipPip
  • 102 posts
  • Other

Posted 11 January 2006 - 09:18 PM

35mm films have a higher resolving power then any of the best lenses every made. The highest resolution films can resolve down to the atomic level. Grains are groupings of atoms, which tend to have crystral clusters of about 6 atoms of silver halide. Until digital starts using nanotechnology it will not have a resolving power close to film, I don't think anyone has ever actually calculated the amount of megapixels that would be equivellent to. As for the aesthetic values of color redition and contrast differences between film and digital, that will depend on the look that you want to acheive, not considering expense.
  • 0

#6 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 11 January 2006 - 11:56 PM

4000 to 6000 pixels horizontally is considered necessary to resolve all the grain and fine detail on a 35mm movie frame. So if the 35mm frame is 4x3, or 4000 x 3000, that's 12 MP. Kodak is not really exaggerating anything; this is generally accepted by anyone working with digital scans of movie film frames (except Phil).

Since a 35mm SLR frame is the same as two 35mm movie frames side-by-side, 24 MP is not an outrageous claim. However, the lack of grain in a digital image allows greater flexibility in image resizing, so many people would say that a 12 MP digital still camera produces frames that are competitive with 35mm SLR frames for fine detail, enlargement capability, etc.

Take a look at this article to see examples of 4K vs. 2K scanning of a 35mm movie frame:

http://millimeter.co...cial/index.html
  • 0

#7 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 01:32 AM

It is difficult if not impossible to say how many megapixels you need to match a frame of a 35mm film, it is only possible to say how many pixels you need to CAPTURE a 35mm frame in scanning, but that is NOT the same thing as matching that resolution with a digital camera.
Sometimes you need 6K to properly sample the shorter side (24mm) of the 35mm frame, but that does not mean that it resolves 6000 lines, because if it did, you'd have to scan it at an even higher resolution to sample it properly. If the film does resolve 6000 lines (some films can) at decent contrast, then you should scan it at 8000dpi at least, on a drum scanner on imacon to sample it properly.



Since a 35mm SLR frame is the same as two 35mm movie frames side-by-side, 24 MP is not an outrageous claim. However, the lack of grain in a digital image allows greater flexibility in image resizing, so many people would say that a 12 MP digital still camera produces frames that are competitive with 35mm SLR frames for fine detail, enlargement capability, etc.


This is the best practical advice you can get
  • 0

#8 dd3stp233

dd3stp233
  • Basic Members
  • PipPipPip
  • 102 posts
  • Other

Posted 12 January 2006 - 01:42 AM

From my person experience, scanning film at 4K never seemed to look good to me. It always looks like it is losing so much information from the original film. It really needs to be scanned at 40K to 400K to really capture the details. I know this isn't cost effective at the present but in order to maintain a certain level of quality, I think this is goal that should be strived for.
  • 0

#9 Landon D. Parks

Landon D. Parks
  • Basic Members
  • PipPipPipPip
  • 1727 posts
  • Producer
  • Cincinnati, Ohio

Posted 12 January 2006 - 02:08 AM

dd3stp233: What planet are you living on? We Don't have 400k scans yet, and may never have them. by the time a CCD is capable of 400k, digital cameras will be the norm... not 35mm.

A 4k print is about 4 Megapixels, since a 2.2MP HDW-f900 is generally considered what a 2k scan in film would be... 4 Megapixels may not seem like a lot, but think about this, most 35mm release print film is only capable of 2k resolutions anyway, so really your scanning at 4k, only to get a 2k again for it to be seen.

The real advantage to 4k scans is in the contrast, ect area. more pixels capture more info from the film frame, better contrast, and so on.

Until the film you release it on can go to 400k as well, don't put the cart before the horse. You can only see what the print your seeing it on can display, so if you scan it at 400k, and then downrez it to 2k again so people can actually see it, you just waisted a HUGE amount of money and time for nothing. Currently, nothing exist's (To my knowlege) to even display a 4k imagine, no TV's, No Projectors, ect.

PS) Just think about what your saying... a 400k scan is 400,000 scan lines. Thats overkill, major overkill.

Edited by Landon D. Parks, 12 January 2006 - 02:09 AM.

  • 0

#10 Max Jacoby

Max Jacoby
  • Sustaining Members
  • 2955 posts
  • Other

Posted 12 January 2006 - 04:32 AM

Landon

You need to pay more attention in math class. We are talking about a 2 dimensional surfaces here, so 4K resolution is 12 Million Pixel (4Kx3K).

It seems to me that once again you are just spitting forth information that you haven't understood yourself and therefore getting it all wrong again.

A release print is more than just 2K, otherwise there would be no point for people to start making 4K scans, would there? This 2K number just comes from the Lucas corner in an attempt to justify that their HD projection is 'good enough' which of course it isn't.
  • 0

#11 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 04:34 AM

From my person experience, scanning film at 4K never seemed to look good to me. It always looks like it is losing so much information from the original film. It really needs to be scanned at 40K to 400K to really capture the details. I know this isn't cost effective at the present but in order to maintain a certain level of quality, I think this is goal that should be strived for.



You are going to far. 4K may not be enough to compleatly recreated the frame, but 8K or 10K would be more than archival.
  • 0

#12 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 12 January 2006 - 04:39 AM

Thats overkill, major overkill.


Landon,

I am sure you will find the HVX 200 camera is the perfrect tool for you. Some people prefer 65mm over 35mm, I am pleased we still have a choice!

Stephen
  • 0

#13 Mike Williamson

Mike Williamson
  • Basic Members
  • PipPipPipPip
  • 534 posts
  • Cinematographer
  • Los Angeles

Posted 12 January 2006 - 04:39 AM

35mm films have a higher resolving power then any of the best lenses every made.


I've actually heard that the exact opposite is true, which is to say that current motion pictures lenses (don't know about still lenses) can out resolve any current capture media including film. This was told to me by a professional DP who owned a rental house.

Where did you hear that film can outresolve lenses? I'm very curious, as this seems like an interesting topic in terms of figuring out where the limitations of these imaging systems are.
  • 0

#14 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 04:41 AM

dd3stp233: What planet are you living on? We Don't have 400k scans yet, and may never have them. by the time a CCD is capable of 400k, digital cameras will be the norm... not 35mm.

A 4k print is about 4 Megapixels, since a 2.2MP HDW-f900 is generally considered what a 2k scan in film would be... 4 Megapixels may not seem like a lot, but think about this, most 35mm release print film is only capable of 2k resolutions anyway, so really your scanning at 4k, only to get a 2k again for it to be seen.

The real advantage to 4k scans is in the contrast, ect area. more pixels capture more info from the film frame, better contrast, and so on.

Until the film you release it on can go to 400k as well, don't put the cart before the horse. You can only see what the print your seeing it on can display, so if you scan it at 400k, and then downrez it to 2k again so people can actually see it, you just waisted a HUGE amount of money and time for nothing. Currently, nothing exist's (To my knowlege) to even display a 4k imagine, no TV's, No Projectors, ect.

PS) Just think about what your saying... a 400k scan is 400,000 scan lines. Thats overkill, major overkill.



now that's a lot of misleading information in one post


Like auridis says, 4Kx3K would be 12 megapixels.

And image contrast has nothing to do with the number of pixels. In digital world, contrast is just a push of a button. The appearance of image tones might seem a bit different from 2K to 4K because of the way grain is rendered, and because there are more gradations in a 4K scan, but that's just theory, and it has nothing to do with contrast.




I've actually heard that the exact opposite is true, which is to say that current motion pictures lenses (don't know about still lenses) can out resolve any current capture media including film. This was told to me by a professional DP who owned a rental house.

Where did you hear that film can outresolve lenses? I'm very curious, as this seems like an interesting topic in terms of figuring out where the limitations of these imaging systems are.



well of course.

Lenses have always been outresolving capture media.

A lens can reach as high as 400 lp/mm at some apertures, and film can reach from 100 to 200 lp/mm
  • 0

#15 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 04:58 AM

A release print is more than just 2K, otherwise there would be no point for people to start making 4K scans, would there? This 2K number just comes from the Lucas corner in an attempt to justify that their HD projection is 'good enough' which of course it isn't.



Actually in order to get a print that has resolving power of 2K you have to start with an original of higher resolution.

Why?

Because every optical copying process follows the same rules of loss of resolving power.

Here is an example of this, I'm using imaginary resolutions just to illustrate (but they are not unrealistic aproximations):

You shoot with a film that can resolve maximum 120 lp/mm with high contrast targets,
and you use a lens that can resolve some 300lp/mm at a certain aperture.

The maximum resolving power of the exposed negative will be:

1/x = 1/120 + 1/300

x= 85.71

The maximum resolved frequency on your exposed negative is 85 lp/mm

Now you want to contact print the negative. Since there is no lens, you only need to consider
the resolution of the print stock.
Let's say it's resolution of your print film is 200lp/mm

So, maximum resolution on your negative is 85lp/mm and the print stock can resolve up to 200lp/mm
then the actual resolution of your finished print is:

1/x = 1/85 + 1/200

x= about 60 lp/mm

There you have it. You started with a negative that can capture almost 6K of data, and a direct PERFECT print ends up having a little less than 3K (60lp/mm)
Now this is the first generation print, and imagine the loss till you get to a release print.
And imagine the loss if you started with a negative that was recorded from 2K files.

The point is, if you want 2K release prints, you have to start with a lot more information than 2K
  • 0

#16 dd3stp233

dd3stp233
  • Basic Members
  • PipPipPip
  • 102 posts
  • Other

Posted 12 January 2006 - 07:58 AM

How many pixels you can fit into a frame of 35mm (still) film depends on how
* many line pairs per millimetre you can achieve with your photography.
*
* As a reference for his 54 megapixel number,
* a paper by William Oliver. Now some parts of the paper are hard to
* to understand, but the 54 megapixel number isn't. On a high contrast
* subject, a medium speed film like Ektachrome can resolve 125 lp/mm.
* Now, it doesn't matter whether your test chart has 125 black lines
* on a white background, or 125 white lines on a black background:
* you need 250 pixels per millimetre to record 125 line pairs.
*
* So a frame of Ektachrome film can hold 36mm x 250 pixels/mm one way,
* and 24mm x 250 pixels/mm the other way. That's 9000 pixels by 6000
* pixels, or 54 megapixels in all.


And scan should have a much higher resolution then the original (for the reasons Filip mentions). Slow speed, high resolution film would be even have a higher number of megapixels.

Something else that isn't mentioned is that film is still resolving the image beyond recognizable line pairs. It may be fuzzy but it is still detail so that the actual amount of megapixels would still need to be higher to capture the additional details.
  • 0

#17 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 12 January 2006 - 08:02 AM

I've actually heard that the exact opposite is true, which is to say that current motion pictures lenses (don't know about still lenses) can out resolve any current capture media including film. This was told to me by a professional DP who owned a rental house.

Where did you hear that film can outresolve lenses? I'm very curious, as this seems like an interesting topic in terms of figuring out where the limitations of these imaging systems are.


Hi,

Its very easy to test, put a lens on a lens projector and see for yourself!

Stephen
  • 0

#18 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 09:56 AM

How many pixels you can fit into a frame of 35mm (still) film depends on how
* many line pairs per millimetre you can achieve with your photography.
*
* As a reference for his 54 megapixel number,
* a paper by William Oliver. Now some parts of the paper are hard to
* to understand, but the 54 megapixel number isn't. On a high contrast
* subject, a medium speed film like Ektachrome can resolve 125 lp/mm.
* Now, it doesn't matter whether your test chart has 125 black lines
* on a white background, or 125 white lines on a black background:
* you need 250 pixels per millimetre to record 125 line pairs.
*
* So a frame of Ektachrome film can hold 36mm x 250 pixels/mm one way,
* and 24mm x 250 pixels/mm the other way. That's 9000 pixels by 6000
* pixels, or 54 megapixels in all.
And scan should have a much higher resolution then the original (for the reasons Filip mentions). Slow speed, high resolution film would be even have a higher number of megapixels.

Something else that isn't mentioned is that film is still resolving the image beyond recognizable line pairs. It may be fuzzy but it is still detail so that the actual amount of megapixels would still need to be higher to capture the additional details.



what many people fail to understand is that a film that has resolving power of 125 lp/mm
doesn't record as much in actual photography because of the lens. The lens limits it.

The ONLY way you could record 125 lines per mm is if you didn't use a lens (like in case of contact printing) , or if your lens has INFINITE resolving power, which is impossible.

So even a lens that can resolve 800 lp/mm limits the resolution of such film UNDER 125 lp/mm
  • 0

#19 Chris Fernando

Chris Fernando
  • Basic Members
  • PipPipPip
  • 148 posts
  • 1st Assistant Camera

Posted 12 January 2006 - 03:00 PM

Filip,
Aren't you talking about lost rez in an optical environ, though? Wouldn't you be scanning that 120 lp/mm neg or something near to that if you are going into a digital environ (for a DI, say), so that say a 6K out to a (conceivably) 120 lp/mm intermediate gets you back (somewhat) to where you were with your o-neg?
  • 0

#20 Filip Plesha

Filip Plesha
  • Basic Members
  • PipPipPipPip
  • 1267 posts
  • Other
  • Croatia

Posted 12 January 2006 - 03:33 PM

Filip,
Aren't you talking about lost rez in an optical environ, though? Wouldn't you be scanning that 120 lp/mm neg or something near to that if you are going into a digital environ (for a DI, say), so that say a 6K out to a (conceivably) 120 lp/mm intermediate gets you back (somewhat) to where you were with your o-neg?



I'm talking about optical copying from a digitally recorded negative to a print or to a master positive.
I'm saying that if you want a 2K print, your digitally recorded negative has to contain a 4K output image.

Even if you skip every kind of optical copying that is possible to skip, you still have to make contact prints out of your digital negatives, and to make them, your digital negative has to be recorded at 4K or greater.


And as for sampling film. Well lp/mm can NOT be converted to pixels for this obvious reason:

Let's say the film can record 80 lp/mm in camera. Times two (line PAIRS) and times 24mm equals 3840 lines across 24 millimeters.
So some reckless people might say, great let's sample it at 4000 pixels and capture ALL the lines

yea right!

Film is an analog medium, and direction does not matter. In digital format, there are only two directions, horisontal and vertical, the rest is an illusion of tiny squares.
So this film can record 80 straight REAL and smooth lines per mm with any angle, 30 degrees, or diagonal, or horisontal, or 43.9996 degress, it doesn't matter, the resolving power is still 80 lines per mm.

Try sampling 80 lp/mm diagonal lines with 4K. Impossible, you'd either get a checker board of pixels, or some soft gray background without any detail.

The only way you could sample 80 lp/mm diagonal lines is to use a higher resolution.
With 6K the image might start showing diagonal detail. But to trully sample diagonal lines at 80lp/mm, one should go for 8K or more.

Of course little would be gained my using such resolution for making cinema prints, but
if people are planing to use digital format for archiving film frames in the future, then they must use at least two times more resolution (8K or more) to be able to sample diagonal lines of modern stocks that can resolve 120lp/mm to 200 lp/mm
  • 0


Aerial Filmworks

Paralinx LLC

Media Blackout - Custom Cables and AKS

The Slider

Metropolis Post

Visual Products

Willys Widgets

Gamma Ray Digital Inc

CineTape

Abel Cine

Rig Wheels Passport

CineLab

Opal

Glidecam

FJS International, LLC

Technodolly

Wooden Camera

rebotnix Technologies

Ritter Battery

Tai Audio

Broadcast Solutions Inc

CineLab

Rig Wheels Passport

Glidecam

Technodolly

Willys Widgets

rebotnix Technologies

CineTape

The Slider

Ritter Battery

Gamma Ray Digital Inc

Opal

Media Blackout - Custom Cables and AKS

Wooden Camera

Broadcast Solutions Inc

FJS International, LLC

Abel Cine

Aerial Filmworks

Tai Audio

Paralinx LLC

Metropolis Post

Visual Products