Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Lens resolution and 4k

Status
Not open for further replies.
M

Mike N

Well-known member
Joined
Nov 4, 2008
Messages
232
Reaction score
0
Points
0
Hello,

I was having a discussion about digital cinema with a guy who really knew his tech stuff and he told me that no lens currently on the market (except ones designed for special applications) gives more than 60lines/mm and that 4k requires more than this.
So 4k or any other higher resolution is from this point of view useless.

Is this true ?
 
another urban legend

another urban legend

Lenses have been resolving 200lpmm Nyquist for 15-20 years now. Cine lenses that is. Maybe your friend has been hididng somwhere all this time......
Jacek Zakowicz, optitek-dot-org:thumbsup:
 
No, it's not true.

Even the 18-50 RED zoom can out-resolve the sensor. The RRPs are greater than 100 line pairs / mm at the center. You need 92lp/mm to out-resolve the sensor.

Perhaps he doesn't really know his tech stuff after all?

Graeme
 
Mike N said:
Hello,

I was having a discussion about digital cinema with a guy who really knew his tech stuff and he told me that no lens currently on the market (except ones designed for special applications) gives more than 60lines/mm and that 4k requires more than this.
So 4k or any other higher resolution is from this point of view useless.

Is this true ?
Click to expand...

No, it is not true.:icon_bs:

MP lenses resolve 50lp/mm with 75% MTF. You can easily see 120 lp/mm on a projector and all MPs have MTF higher than the 4k S35 sensor.

What about still picture cameras? They often have more than 4k resolution and still lenses are generally not as sharp as cine lenses. You can also mount still lens on Red One, if you want to.

BTW, what is a lens for "special applications"?

edit: hehehe Jacek and Graeme beat me to it :)
 
We might be talking about two different measuring systems; your eyes and a MTF set-up.
Your eyes are are very forgiving, everyone is slightly different, and might stretch lp/mm down to as low as 2%. There is a reason that Clairmont Camera bought (for many thousands of dollars) a Zeiss MTF machine, to take the human equation out of the process.
There are cine lenses that I have seen go 200 lp/mm (6mm Century for 16mm for example, designed by Leitz in Canada) on a projector, but put that same lens on a MTF machine, I do not think that number (at 50% MTF) would stay at 200.
 
What is the theoretical max number of line pairs per mm for 4K? How do you calculate that? Do you divide 4096 by 22mm (the width of the RED sensor area used), which is 186 lines, which would be 93 line pairs per mm?

So if the RED is more like 3.4K, doesn't that work out to be around 77 lp/mm?

Sometimes people throw the 80 lp/mm as the resolving power of the RED sensor, where does that come from?
 
These are the numbers

These are the numbers

David Mullen ASC said:
What is the theoretical max number of line pairs per mm for 4K? How do you calculate that? Do you divide 4096 by 22mm (the width of the RED sensor area used), which is 186 lines, which would be 93 line pairs per mm?

So if the RED is more like 3.4K, doesn't that work out to be around 77 lp/mm?

Sometimes people throw the 80 lp/mm as the resolving power of the RED sensor, where does that come from?
Click to expand...
Hi David,
You are correct, these are the theoretical numbers to go by- it's been discussed here a while back- or was it CML- I don't recall.
I don't know if you noticed with all that was going on at the SALT but the Optimator has up to 2000 LPH(lines per height) target in the middle of the reticle and that could give you the real life actual resolution for any camera tested with it. That is one of the many purposes of the Optimator- to generate "reference" frame that is always identical and constant no matter what camera it is put on. Also since the optics are built in they are a "constant" factor so only the sensor/camera is tested.
BTW I'll be showing the Optimator in the Band Pro booth at Cinegear today and tomorrow.

Jacek Zakowicz, optitek-dot-org:thumbsup:
 
Standard reference for TV lines is vertical resolution, i.e number of horizontal line pairs visible over the vertical image height divided by senser height. Maximum vertical res for the sensor at 100% mtf would be vertical pixels/2/sensor height.

For 2/3" Scarlet for example 1620/2/5.35= 151.4lpmm
 
David Rasberry said:
Standard reference for TV lines is vertical resolution, i.e number of horizontal line pairs visible over the vertical image height divided by senser height. Maximum vertical res for the sensor at 100% mtf would be vertical pixels/2/sensor height.

For 2/3" Scarlet for example 1620/2/5.35= 151.4lpmm
Click to expand...

But isn't it line pairs per millimeter, not per whatever the entire height of the sensor is? So wouldn't you start out using the number of pixels that existed within a millimeter section of the sensor? (Of course, all of this ignores what the OLPF is doing...)
 
100% mtf?!!!

100% mtf?!!!

David (Rasberry to be clear),
With all due respect- Ive been testing lenses fo 20+ years, hundreds of them on MTF machine and have never seen 100% mtf- this is a theoretical number.
Maybe if you look at 5, 10 lpmm....
Jacek Zakowicz, optitek-dot-org
 
A line pair is alternating white and black lines. Divide the vertical pixel count by 2 gives maximum line pairs the sensor can record at 100%mtf over its full height. This would be the maximum vertical resolution in TV lines for the whole image. Divide by the sensor height in mm and you get LPmm. I'm not sure if a sensor can record higher than this resolution at lower MTF since each pixel is recording a discrete value.
The resolving power of lenses is generally specified over a range of MTF values, with 50% being the most common reference. As MTF falls below 100% you get reduced contrast between the black and white lines of a pair until at some point you just see a grey field.
You are right Jacek. A lens that can hit close to 90% Mtf in the 30 lpmm range is doing really well.
 
how MTF testing works

how MTF testing works

David,
I know how MTF works, I have tested hundreds of lenses with it. I'm just trying to make people aware that they will not see 100% mtf ever in the res range they need so they don't expect it or reject lenses if they don't get 100%. You have to apply realistic standards....
BTW this applies to lens MTF only(although I doubt that with the noise accounted for any sensor gets 100% either)
Jacek Zakowicz, optitek-dot-org:)
 
Agreed. Not meaning to belittle your knowledge. Just wanted to define it clearly for others that may not know.
 
square pixels

square pixels

David Mullen ASC said:
I'm just asking how you calculate the theoretical maximum lp/mm of a sensor. If it's a millimeter section, does it matter if it's a vertical or horizontal section?
Click to expand...

No it doesn't. Pixels are square. The LPH standard is accepted for still and video- that's all. However you skin the cat-so to speak.
Jacek Zakowicz, optitek-dot-org
 
I suppose you could calculate using horizontal resolution with vertical line pairs. Same calculation would apply for the theoretical maximum, but in a practical sense vertical resolution is the limiting factor.

So for 2/3" Scarlet 3072/2/10.1= 152.x LPmm. This is of course an approximation.

Another approach would be (pixel count/2) / ( pixel count * pixel size)
Example for RED 1 horizontal: (4096/2)/(4096*.0054)=2048/22.12= 92.58 LPmm

Jacek is right vertical res is the accepted standard.
 
I found this link interesting from Cooke.

http://www.cookeoptics.co.uk/cooke.nsf/0/04F15CD71BDD0E5C852576340056ECBA/$FILE/Are%20your%20lenses%20good%20enough%20for%20that%204k%20high%20definition%20camera%20--%20Cooke%20Optics.pdf
 
Status
Not open for further replies.
Back
Top