F35: Arri Alexa vs Sony F35 skintone

eddavid said:
I'd fire you buddy just as quickly. :)
I did the match to chart in resolve - and I was more curious to push the image and see what happens. If you notice, I left the log c and slog video on there - feel free to do it yourself and post the results!

Also notice the Amira has more noise than the F35 when pushed in 200% - that surprised me. Both shot at 400 ASA.
Click to expand...

I might.

F35 is 1080 right? Arri 3.5k - no suprise that the former is less noisy?

It seems you cant win with tests as no one knows what you are trying to do.

I have a specific test for testing th 'bending images in post' - shoot under tungsten with the camera at 5500k and see how bad the file breaks when you bring it back to grey - it is repeatable and simple.


S
 
please go for your test - I would love to see it.
F35 is a 8k photosite sensor. Amira is a 2.8k photosite sensor.

Resolution is just marketing. It's all about the sensor and the debayer.
 
Here I did have a quick fiddle. My comment would be that I cant be bothered to see any difference AKA neither was bad. The beef I have always had with Slog(1) is that I cant find a good lut as sony AFAIK never issued any. With Slog1 you are swimming about not knowing where to go. Arri did good by handing out a functioning LUT.

Id give Arri the win because the better LUT means less time in Resolve.. but I dont know if it is a vimeo thing but the Arri did band in the shadows above the card and the sony did not.. so Id give the win to the Sony.

Anyway using Arri LUT and MgGragor F35Lut I tried to make 'normal' images and then versions that have a green murky vibe of 'prison'

Untitled_1.11.1.jpg
Untitled_1.11.2.jpg
Untitled_1.12.1.jpg
Untitled_1.12.2.jpg
 
morgan_moore said:
Here I did have a quick fiddle. My comment would be that I cant be bothered to see any difference AKA neither was bad. The beef I have always had with Slog(1) is that I cant find a good lut as sony AFAIK never issued any. With Slog1 you are swimming about not knowing where to go. Arri did good by handing out a functioning LUT.
Click to expand...

A valid point. I believe there were S-Log 1 LUTs about, however Sony never really made them public because they didn't envision "consumers" having an F35 at home.

How they thought people would deal with F3 S-Log, I do not know.

This is one area they listened - S-Log 2 and 3 are much better served by them.

One could equally ask why they did not use Panalog, as that was designed explicitly for the same sensor...
 
Jim
I would not have bought the FS7 without the LUTs being in place. I used to shoot on a Sinar/Hasselblad 2004(?) (digiback) and while it created a superb file making one was a right mission - Phase One pulled whipped Sinar/Hassy by creating Capture One that had 'luts' for their digibacks and made 'normalisation' a one click process. Ever since then I have seen the camera as only the first step in the chain. I really sold my F3 as the second step in the chain was missing..

Ed.
I got the vid and popped in Resolve, cut it in half, for the Arri section I added the Arri Lut and did a very minor tweak in curves. For the Sony I added the F35Macgregor LUT and spend a little longer tweaking the the curve, mainly pulling a bit of red out of the shadows which I seem to have to do for all Sony cameras. The F35Maggregor LUT has been on my system for a year so may not be his latest/greatest effort.

I now had the 'normalised images' - the Arri took 30s and the Sony 1min, so Arri 'won'

I then added a 'look' node (the look being 'prison') to above each normalised image and pulled the curves to add some blue to the shadows and green to the mids/highs and globally desaturated a little.

Both the 'normalised' node and the 'look' node were in front of the LUT node.

S

flow.jpg
 
Last edited:
Goes to show the crazy marketing behind cameras. With all the new releases sony gave us the last couple years still none beat the f35. A couple exceptions but none in the price range you can pick up an f35 for. Thanks for the test.
 
Sony's problem is they never sold an F35 for $10,000! We will never see cameras quite like it again. The only thing I think is comparable to that era of F35/D-21 is maybe the Alexa65?
 
morgan_moore said:
Here I did have a quick fiddle. My comment would be that I cant be bothered to see any difference AKA neither was bad. The beef I have always had with Slog(1) is that I cant find a good lut as sony AFAIK never issued any. With Slog1 you are swimming about not knowing where to go. Arri did good by handing out a functioning LUT.

Id give Arri the win because the better LUT means less time in Resolve.. but I dont know if it is a vimeo thing but the Arri did band in the shadows above the card and the sony did not.. so Id give the win to the Sony.

Anyway using Arri LUT and MgGragor F35Lut I tried to make 'normal' images and then versions that have a green murky vibe of 'prison'

View attachment 101673
View attachment 101674
View attachment 101675
View attachment 101676
Click to expand...


I think these are better results. The grading in the clip has the highlights going out of control.

As for Amira vs F35, without considering the form-factor because there is not much sense in comparing that since they were made for two totally different shooting scenarios, one could say it's a toss between the two in terms of looks. At least if you are looking at stills you can match them pretty seamlessly. But I still prefer the F35 in motion.

Image quality, with the ARRI being a bayer sensor running at it's resolution limit and the F35 being a full RGB CCD with enough oversampling to spare, it's hard to say the ARRIs are in the same level of image quality, at least technically. But in reality it will be a toss up. With the ARRI having a slight edge in dynamic range but loosing in color fidelity and depth, having the upper hand in low light but losing in motion and so on. It's a give and take. Pick the set which suits you better.

What makes it interesting is exactly the point Ed brings up. The F35 can be had for way cheaper these days, which of course has nothing to do with how capable the F35 is.
 
What surprised me in my crappy test was that the Amira had more overall noise than the f35 at 400 ASA - which is what I rate the Alexa sensor at. I don't like it at 800 ASA - I find it too noisy - and so does the DP of Mad Max. So it doesn't really have the low light edge.
 
I often wonder how closely related the ALEV-I (as used in D-20/D-21) and ALEV-III are (as used in ALEXA). They shared the same basic pixel count I think, both CMOS. Apparently the original sensor was pinched from something else, or certainly not designed by ARRI. The D-21 was very noisy even at 200 ISO.
 
Weird since just about every DP in every AC article I have seen shoots at 800 or even 1600 when needed on the Alexa. I was just reading an article today stating that actually. I guess guys who hang out on DVX know best though. Keep hating that amazing Amira Ed... maybe you need to call Arri and tell them how bad that noise (and camera) is. They may show you a list of movies shot on that sensor that would make you cry, so be prepared. They may start with this one and keep going, for DAYS. http://www.theasc.com/ac_magazine/June2012/TheAvengers/page1.php
Not the exact sme camera, I know, but close enough. Lets just admit its the best thing going and we can rest. :beer:

-McGarvey typically maintained a T4 or T5.6 for day exteriors, and T2.8½ in other situations. “I shot everything at [the Alexa’s base ISO of] 800. When I tried to rate it lower, like at 400, it seemed to build up in the shadows, and I didn’t feel it had the same range. So I simply used IR neutral-density filters to bring down the stop for exteriors.

From Roger Deakins forums: by Roger » Wed Oct 01, 2014 2:23 pm
You don't see noise in the image until you rate the Alexa above 1280ASA. I have rated it that high and even at 1600ASA for a night time shot.

And Im sorry dude but saying the Amira doesnt have the low light edge is just.. well, you are full of BS this time.
 
Last edited:
Mad max fury road dp rates at 400 asa. I rate lower than others. But yes its a personal choice. Im not hating on amira. I am just showing my results. I also thought the alexa was a low light beast but more i test it the more issues i find with it. Its not a miracle camera. It has great hightlight rolloff but i rather shoot on film than alexa.
 
Let's not forget DoPs used to rate their film stocks differently - no-one is "right" or "wrong"... they just have different tastes and needs.

Panavision's official rating for the Genesis was "between 400 and 640" - it was up the DoP to decide what they felt comfortable with.
 
Updated grading per Morgan's example:
https://vimeo.com/129045883

Both cameras have a different style of seeing the world - in the end I think if I did a more thorough test, motion of the f35 feels more natural - the alexa motion has its own feel - it's a new style so in the end my tastes which always hold the motion of film the best helps, but of course its so much more complicated than all that - many factors in it all that dictate using one camera's preference over another's - and again the power lies in the camera, lens, lighting, and colorist - nothing can be taken independently - it's all a collective effort.

And further proof to me to not chase shooting on the most expensive camera on projects when I can or shooting on the alexa when a f35 can do - when the money can be spent towards a better lens system or hiring a colorist or more crew members to make it look better.

Of course if cost is not an issue I would shoot film. But we all live in a world of costs, and various costs per project vary - and I mostly dwell in the indie filmmaking world in the lowest cost world. And the f35 for me is the king of this lowest cost world. As long as the cameras keep firing up.
 
JMtheDP said:
I often wonder how closely related the ALEV-I (as used in D-20/D-21) and ALEV-III are (as used in ALEXA). They shared the same basic pixel count I think, both CMOS. Apparently the original sensor was pinched from something else, or certainly not designed by ARRI. The D-21 was very noisy even at 200 ISO.
Click to expand...

I'm not sure how close the sensors are. But the D21 is miles worse than the Alexa in image quality. To be honest, unlike the F35 which transcended it's time and hangs out with new cameras without any problem and still beats 95-100% of them, the D21 is a true product of it's time and doesn't hold well at all today. Not only because it is very cumbersome to work with but the image quality of even a F3 with S-log 444 will leave the D21 Gobsmacked wondering what happened. I would also say that rating the D21 ISO200 would be like rating the F35 ISO900. The D21 is basically ISO100.
 
You must log in or register to reply here.
Back
Top