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 :)

Leica Lens Resolution test...

Otto Schade found out long ago, that human perception is not influenced that much by the maximum resolution, but by the MFT at about two thirds of the maximum. Humans perceive credibility of texture in that area, so the strength of the Epic sensor in that area might be even more important than the question of the exact maximum.
 
OK, even without totally understanding the units and scales on the charts (and thus what is actually being measured), I'm impressed.

But in the interest of clarity, I'd like to ask the following dumb questions about the charts:

  1. What is being measured on the y-axis? Does this translate to more familiar units such as 18% grey?
  2. Why is the y-axis scale on the first one so different from the other two? (units in the 100's vs units in the 10,000's)
  3. What is being measured on the x-axis? It says distance, but it looks more like lines per picture width. Or is it line pairs, or is it picture height?
  4. How does this info translate to % MTF?

And a few more:

Since this was primarily a sensor resolution test, I assume you used the sharpest lens in the set at it's "sweet spot." Did you also measure resolution across the frame into the corners? And what was the T-stop for this test?

This looks like the gold standard of performance. Will you be publishing charts for your own lenses? I'm especially interested in the Scarlet Fixed zoom. And I bet the competition is terrified of the chart the little Scarlet 2/3 sensor may produce.

Thanks!
 
Or you can also look at the MTF50% point as a good indicator too.

Graeme
 
Steve, I'll give it a go:

1) Y-axis is brightness of the chart at that point. What we're interested in is contrast at a particular resolution, and contrast at each resolution (measured on the x-axis), which is the difference between the brightest and darkest the graph goes at that point, is what we're interested in.

2) Y-axis scale is not particularly relevant - what we look at is what is the maximum contrast at the lowest detail (which the graph is scaled to) and how that diminishes with increasing resolution.

3) X-axis is horizontal resolution in lines.

4) It's a relative MTF test, relative to the maximum contrast at the lowest detail. So the contrast at each resolution (see 1 above) is compared to the contrast at minimum resolution, so when the contrast drops to 50% of that of the reference, you're at MTF50.

Hope that helps. Jarred can fill you in better on the lens details.

Graeme
 
For camera geeks like me the test charts are mind blowing.
As Christopher and Uli both noted, contrast and MTF in the fat part (middle frequencies for Epic, high? frequencies for Alexa) of the image are the "meat".
Charts leave you cold? Go to RSH and see for yourself. If you can't see the difference - I'll cover your gas expense and advise you to spend the coin at the eye doctor ;-)
Just sayin' :coolgleamA:

Cheers - #19
 
Thanks, Graeme. So this:
Uli Plank said:
Otto Schade found out long ago, that human perception is not influenced that much by the maximum resolution, but by the MFT at about two thirds of the maximum. Humans perceive credibility of texture in that area, so the strength of the Epic sensor in that area might be even more important than the question of the exact maximum.
Click to expand...
plus this
Graeme Nattress said:
Or you can also look at the MTF50% point as a good indicator too.

Graeme
Click to expand...
and a comparison of the 50% MTF point - about 1,000 l/ph with Alexa and about 2,500 l/ph with Epic means the Epic image should maintain "credibility" at a much larger image size than the Alexa. Which was Jim's point about finding religion.

Have a great time at the revival next Tuesday or Wednesday!
 
Great test on such a wonderful Royal Wedding Day.

It seems to me that Epic 5K high resolution will have a very happy married life with new Leica SC primes.

Leica lenses were always special because they tried to meet the highest MTF with a unique and gorgeous "Leica Look".

Also dont't forget that you can still use vintage Leica R lenses on Epic 5K with a little help of Canon mount and LR to EOS adaptor.

I've just got back my Leica R Summilux 50mm f/1.4 yesterday.

Leica_Summilux-R_50mf1-4_02.jpg

Leica R Summilux 50mm f/1.4, the last 2nd generation.
Shot on 5DMK2 with Canon 24-70mm f/2.8, CA artifacts visible even after PS CA correction.
Intentionally a film grain (Kodak 5274nc) was added to the image in Shake.

Leica_Summilux-R_50mf1-4_5dmk2.jpg

Titanium RED #694, Shot on 5DMK2 with Leica R Summilux 50mm f/1.4, Lens Made in Germany, metal alloy body, great mechanics, smooth focus,
not so much of lateral CA and other lens artifacts, highest resolution, sharpest, best MTF, amazing natural color rendering that always part of "Leica Look",
price of used lens vary between $800-$1,600 depends on the usage conditions.
Intentionally a film grain (Kodak 5274nc) was added to the image in Shake.
 
Blair S. Paulsen said:
For camera geeks like me the test charts are mind blowing.
As Christopher and Uli both noted, contrast and MTF in the fat part (middle frequencies for Epic, high? frequencies for Alexa) of the image are the "meat".
Charts leave you cold? Go to RSH and see for yourself. If you can't see the difference - I'll cover your gas expense and advise you to spend the coin at the eye doctor ;-)
Just sayin' :coolgleamA:

Cheers - #19
Click to expand...

+1

If I were in LA, I would most definitely be there on tuesday, oh well... OTOH, I'm paying for my M in a couple days. It's so close I can taste it. I'm just plain happy right now....

Oh Leica Summilux-C primes, how do I covet thee....
 
Jannard said:
Actually... you can see a significant difference between a 1080P camera and EPIC on a 40' screen. We have the luxury of demonstrating just that on Stage 4 of RED Studios Hollywood. PLEASE stop by. You will be blown away. And you won't be the 1st.

A 40' screen separates the men from the boys very easily. That's why so many features that have done side by side tests have chosen EPIC.

Jim
Click to expand...

This is why I want to get my EPIC!
Jim, I submitted my inquiry for EPIC-M, and I also emailed Kelly, is there anything else Is should or just just tight?
 
Just posted a new thread with the proper date and time...

Jim
 
Can I please have an idiot's guide (am I the only one?) to what is going on in these charts? What are the XY axes and how to read and understand these charts. I can see the differences, but what exactly do they mean?

Thanks

John
 
John Redfern said:
Can I please have an idiot's guide (am I the only one?) to what is going on in these charts? What are the XY axes and how to read and understand these charts. I can see the differences, but what exactly do they mean?

Thanks

John
Click to expand...

The charts are plots from a zone plate.

X-Axis is resolution, from zero through to 5k. If we had a perfect lens and camera, (which doesn't exist) the height of the modulation on the Y-Axis would be the same at all resolutions. In other words, there'd be no contrast reduction at high details.

The degree at which contrast (the difference between the peaks and troughs on the curve) is reduced or maintained at any particular resolution is governed by the MTF of the system, so this is basically an MTF test.

Always with MTF you have a good figure at low resolutions. So the contrast, is high at low resolutions, and that's why we start with the large swing on the curve at the left. As we increase resolution, the contrast begins to drop off. When it drops off to the noise floor we have no more contrast, and hence no more resolution.

The more contrast you see for more resolution, the more overall detail you'll see in the image.

Graeme
 
Did you use a MegaTrumpet Chart from DSC Labs?
It's kind of hard to find charts good enough for this kind of resolution as it's hard to find a valid chart to test latitude without using their DX1-102dB. Any other suggestion?
I've just met David Corley at NAB and his explications about their new charts were amazing.
 
You must log in or register to reply here.
Back
Top