LARGE FORMAT LOOK TEST: ALEXA 65 vs. ALEXA MINI

[Chuks Ude]

I shared this over at DVXUSER and figured it would be good to share here. I came across this really well done test of the ALEXA 65 vs ALEXA Mini. It does help to clear up some theories and misconceptions about the effects of sensor size and lens focal lengths, while also proving some of them.

https://vimeo.com/444951736


There is a good write-up on the website of the DP that performed the tests:

https://manuel-luebbers.de/large-for...vs-alexa-mini/

"Both cameras are mounted into a stereoscopic rig, so the cameras are recording the same scene and action with the same perspective center simultaneously. The cameras are recalibrated after every lens change so the nodal points of both systems match perfectly."

[Karim D. Ghantous]

It is very tiring to hear the misconception that longer lenses have more 'compression'. No matter how many tests you do, and no matter what proofs you have, some people will insist on the myths. It really is not helpful to spread misconceptions.

[Joe Cage]

It is very tiring to hear the misconception that longer lenses have more 'compression'. No matter how many tests you do, and no matter what proofs you have, some people will insist on the myths. It really is not helpful to spread misconceptions.

It does look like it compresses space though, or are my eyes deceiving me. The perceived distance between people looks a lot different on a wide or telelens.

[Misha Engel]

It does look like it compresses space though, or are my eyes deceiving me. The perceived distance between people looks a lot different on a wide or telelens.

When you have some spare time, you might want to check this website http://www.yedlin.net/

http://www.yedlin.net/NerdyFilmTechS...hLensBlur.html

[Phil Holland]

Somehow our industry went to "lens compression" when all of this is related to "perspective compression" which is in fact a term used in arts and sciences as well as being around for over a century. Not sure where the confusion came from.

Examine the foreshortening of a character in a medium shot with their hand coming towards the lens with some sort of background framed while framed for the subject, say belly button to top of head, with both a 24mm and 200mm prime tells you everything you need to know. And yes you would have moved the camera. And yes you would have moved the camera differently if it was different a different format size and those exact two focal lengths.

Camera Position, Format Size, Focal Length, T-Stop, and Subject Distance to Camera. All of those things have an impact on your frame.

Manuel's test is great when showing the impact of the format sizes at the same T-Stop and choosing the appropriate focal lengths to match FOV. Beyond that, yes due to the different optical designs you can see subtle different looks on how each of the focal lengths draws a face. As to if that's a "meaningful difference" to you, well yeah, you decide that.

I think some of the confusion itself comes from showing either the "matched frame" or the "difference frame" in some of the comparisons. Wherever it's coming from, it's become so odd for something that is so fundamental.

[Mike P.]

Similar point of confusion, larger formats actually have a *deeper* depth of field than smaller ones, however it doesn't seem that way after you compensate focal length or distance from subject to get similar framing.

I find subjects seem to pop more on larger format (more so than just shallow depth of field) because the focal plane is actually deeper, but the focus fall off (due to focal length and/or distance from subject) still looks shallower (e.g. bokeh being equal, you can get the nose & eyes in focus on LF, while only eyes on S35). So it's the best of both worlds... But I digress.

To be fair, I still dig seeing both cameras side-by-side/matched in any capacity. Also ALEV3 still crushes it... (I watched the 10 year anniversary video the other day [closer to 11 years now], and I'd forgotten how many different models there were with the same sensor.)

[Matt W.]

Does the LF have more highlight dynamic range than Mini/Amira/Classic?

Tests here say yes:

https://vimeo.com/user60380576

But separate set ups for LF and Amira.

Also, does ArriRAW improve highlight detail at all? Does anyone have a comparison?

[Michael Tiemann]

I shared this over at DVXUSER and figured it would be good to share here. I came across this really well done test of the ALEXA 65 vs ALEXA Mini. It does help to clear up some theories and misconceptions about the effects of sensor size and lens focal lengths, while also proving some of them.

https://vimeo.com/444951736


There is a good write-up on the website of the DP that performed the tests:

https://manuel-luebbers.de/large-for...vs-alexa-mini/

"Both cameras are mounted into a stereoscopic rig, so the cameras are recording the same scene and action with the same perspective center simultaneously. The cameras are recalibrated after every lens change so the nodal points of both systems match perfectly."

That's a good test. Bravo! It confirms everything I've come to learn about this subject which is

1. Geometry is geometry, and if you have similar triangles from subject to image space, THAT'S THE PERSPECTIVE, period.
2. Lenses perform differently with respect to themselves: aperture affects not only depth of field, but also contrast and resolution. Field curvature can vary by focus distance. Etc.
3. Lenses perform differently with respect to one another: distortion, bokeh, contrast, resolution--it's all up for grabs.
4. Sensor size brings different lens considerations into play: larger sensors with the same pixel size are going to expose more of how the lens renders away from the center of the image (where MTF often starts to fall off a cliff); larger sensors with the same number of pixels (hence larger pixel size) will prioritize lower frequency MTF, leading to higher contrast and apparent resolution. These two factors trade off on a lens-by-lens basis
5. Pixel quality determines color quality, dynamic range, and noise.

When you put it all together, as this test has done very well, you see that none of the "perspective compression" garbage has anything to do with anything, because the geometry of similar triangles zeroes all that out. What does matter are the intrinsic qualities of the lens--at a given aperture, focus distance, relationship to flare sources, positional relationship to subject and background, etc., the intrinsic qualities of the sensor, and how all of that ultimately affects--as Phil Holland calls it--the rendering of the lens. As can be seen in this test, the aesthetics of noise, sharpness, background blur, bokeh, etc., all come into play. One would be a fool to pick a lens (or a lens crop) solely on achieving the FOV shown in some storyboard sketch. All of the factors have to be considered, and the good cinematographer has a feel for how to get all those elements working together in composition. Which is why we have choices.

[Patrick Tresch]

When you have some spare time, you might want to check this website http://www.yedlin.net/

http://www.yedlin.net/NerdyFilmTechS...hLensBlur.html

I hope he'll never have to shoot 4k+. He made such a good demo about 2k.

[Patrick Tresch]

It's good to see less noise in the Alev3 65mm at 3200 iso the Mini fell apart.