AJA Cion vs F65

Ebrahim

I will certainly do more testing given the opportunity to do so, which I hope will come soon.

I understand what you are saying, but in the environment I had the camera (a harsh one certainly, but not unrealistic) and the settings I had the camera on (also possibly wrong) it was performing in an underwhelming fashion in both the highlights and the shadows.

I did look at the camera exposed for both the highlights and the shadows, in either case the image suffered terribly. Exposing for the window so that the highlights looked the same as the F65 resulted in the left half of the room being nearly totally black.

When I know more, I'll be sure to publish it here because I consider it of great importance that I do not unfairly judge a product. I would really like the Cion to perform as advertised.

One thing I will say in their favour, the global shutter certainly appears to work.

James
 
Thank you, James.

So you did actually try to expose for the highlights and the shadows were not usable when brought up? I would really appreciate a simple Tiff or DNG of the exposed-to-highlight Cion so we'd try pushing the shadows ourselves, because many Cion testers went on about that specific quality, and if it doesn't have it then the camera has very limited DR vased on you test (and what we saw so fat from others)
 
The way I set up the camera I set my exposure using the F65 (which I know how to expose by eye), then I set a matching level of exposure on the Cion, adjusting for the ISO difference with the aperture.

If the camera was genuinely a 12-stop camera then I would expect to probably see the highlights clipping a bit sooner than the F65, but the shadows to remain similar. As you can see this was not the case.

Apparently I was using the wrong gamma setting, however having now re-read the manual for the camera I put this down to some very confusing wording which I would really like to see AJA change. It does not make any sense from a cine perspective, at all.

However, the issue for me isn't really how many stops are available as much as how does the camera handle over-exposure. If AJA are nice to me, I might see about pairing the camera up with some filters to see if I can help it at all.
 
I haven't read a single review where the user was pleased with the DR of the Cion. All have said the DR is not 12 stops so either AJA has a usability issue if no one can seem to get the settings correct or the camera is in fact limited below 12 stops.

A $10K sub 12stop ISO400 camera is a tough sell in 2015. I was looking at renting the Cion for shoots I would want the great data and codec options with global shutter. Now I'm resigned to the fact that what I would gain in post wouldn't be worth the extra cost & time of lighting for a limited DR, low ISO sensor over just using my FS7 in 10bit XAVC or rent an URSA with the new sensor when its released.
 
Of course you're assuming the URSA 4.6k is going to be better. I don't see it getting 15 proper stops.
 
JMtheDP said:
The way I set up the camera I set my exposure using the F65 (which I know how to expose by eye), then I set a matching level of exposure on the Cion, adjusting for the ISO difference with the aperture.

If the camera was genuinely a 12-stop camera then I would expect to probably see the highlights clipping a bit sooner than the F65, but the shadows to remain similar. As you can see this was not the case.
Click to expand...

Well I think this is where the problem ''might'' come from, the F65 and Cion have different approchesto set exposure where you get the most DR. So matching the exposire to the f65 might be a bit unjust for it.

Why I believe this is the sensor is the same one in the Blackmagic 4K cameras, Production and URSA, both of which I know have a great amount of DR in LOG, definitely not limited or 10 stops, 12 seems accurate and agreed by all. So it's would be strange if the Cion has significantly less DR, unless they did something horrible with the processing pipeline like clipping two stops with a gamma curve, which could very well be true. Let's just get the exposure variable out of the way before gamma, they say under expose for the highlights and push up, wjere the shadows have a great range of information.
 
I would like to think after sixteen feature films and over fifty shorts, including work on 35mm, I would know how to expose an image. I'm not saying you don't of course, but with the setup I had there was no "happy compromise". Exposing for the highlights left the dark side of the room so far gone it was unusable - and no way recoverable in post. Until I get my hands on a Cion again I will not draw a more definitive conclusion but I know a great many situations where "exposing for my highlights" simply wouldn't be viable, especially not if the camera comes in at a measured 8 stops.

Still, further testing will shed more light on this. I never put my F35 and F65 up against each other, they both looked great in isolation. The F3 was also pretty good too. Time will tell but in any case I think AJA need to relabel their menu, I still can't tell if there is a proper log like option without shooting raw.
 
Last edited:
James, there is no one way to expose a camera that comes from past experience of lonf shooting with different cameras, each new camera in the digital era has an optimum way of setting exposure that can vary greatly from another camera, and most certainly from exposing 35mm.

I am solely trying to give genuine advice to help getting the most out of this comparison slash test.

Thanks you anyway for volunteering time to do this.
 
I know that of course, my main aim wasn't so much to create a picture as to check out the DR claim made for the camera. It was based on my previous experience (also limited) with the Black magic 4k.

In the real world if I was shooting this as a scene I would have turned lights on inside the room, closed the curtains and tried to narrow down the exposure range. But I like to see what a camera does outside of its comfort zone.

With my F900 I would expose for the highlights a bit, but with the F35/F65 and Alexa I like to expose for skin tones - the cameras have such reach into both the highlights and shadows I was rarely worried about how the highlights looked.

When I speak to AJA next I will know a lot more - I am not expecting miracles but I'll always admit when I'm wrong.
 
FrameFarmMedia said:
I don't think it will have 15 proper but adjusted for marketing stat inflation of BMC4K/Ursa4K it should come in at 13ish.
Click to expand...

BMD hasn't released a single camera with 'inflated' DR numbers, at least not to that degree. Shooting RAW on any of the cameras you're pretty much at spec'd DR.

URSA being a solid 12, the 4K camera 11-11.5, Pocket straight-up 13 and the 2.5K hovering at 12-12.5

I expect URSA 4.6K and Mini to be dead on 15, as it's actually UNDER Spec'd from their design.
 
Let's see how accurate AJA is on rating DR. It's their first ever so it's hard to get any kind of impression on the camera. From their recorders back in the day though, they were excellent in customer feedback, and their products were extra reliable and TOUGH, which would be all great if they transfer to their cameras.
 
Dynamic range is a tricky beast indeed. Certainly there is always argument about whether we mean "absolute" or "usable" dynamic range. In some cameras these are very close, in others not.

Critically, some cameras have nearly or actually identical DR in both raw and and video/log (ARRI state that is the case with the ALEXA). On other cameras, you need raw to get the best from it. I do not like being forced to shoot raw to get better DR (I don't like having to do it for the resolution of the F65 but there is a good reason behind it).

I really must stress that I will be testing this camera again in a more detailed fashion, however for those curious here is what the shot looks like if one exposes "for the highlights". The window was set so I had the same exposure as the F65 saw outside, with no horrible burning on the sofa arm.

wula52.jpg


As you can see that looks miles better.

But, now the rest of the room is badly under.

xfzwj8.jpg


Doing some research the BM4K really benefits from shooting raw and it doesnt even really come at a space penalty (I believe ProRes 4K is a waste of MB for these cameras). I wish AJA had made this do raw internally.
 
Jim's informal findings appear very close to CML's tests: http://cinematography.net/cml-uwe-2015.html

It's AJA's first outing, and by most accounts they've missed the mark- as compared to the competition in 2015.

I've not shot with it yet, but as a potential customer I'm disappointed in their decision to use their proprietory media when cheaper and faster off the shelf solutions are currently available. As a small boutique manufacturer their calling card should be speed and flexibility of design- how else are they going to compete?

Hoping BMD's new sensor is a big step forward- the price and feature set of the Ursa mini is very interesting. Also hoping Adobe incorporates ACR/Lightroom in Premiere because their debayer is much kinder to the image than Resolve's.

A camera's workflow should really be designed to help us, the user. If their workflow, menu, setups are cryptic and not implemented well, it really doesn't help anyone so yes, I think they deserve the harsh words knocking about on the internet. It is a lot to overcome but that is the price of subpar work these days.
 
JMtheDP said:
Doing some research the BM4K really benefits from shooting raw and it doesnt even really come at a space penalty (I believe ProRes 4K is a waste of MB for these cameras). I wish AJA had made this do raw internally.
Click to expand...

Agreed with these statements. In the case of BMD cameras, it reads like:

- 2.5K doesn't benefit that much, color does though.
- Pocket Camera absolutely hits its peak performance shooting RAW.
- 4K Production Camera's Lossy RAW will allow it to hit 11.5, ProRes 4K 444 not so much.
- URSA 3:1 or Lossless is a straight-up 12 stops. Shooting Lossless and 3:1 on URSA often (weekly), it's actually still a bit surprising how sturdy the 3:1 is in terms of overall IQ.

I haven't had a chance to mess with a Cion yet, but I've been with that sensor for over a year now in two different iterations, it's not the worst thing out there... talking about you, GH4.
 
I agree Karl - I do not see why they could not have used either standard 2.5" SSDs, or even Cfast cards (which seem to be becoming popular).

From the sales persons' point of view, AJA creating proprietary media is just additional cost they don't want to have to ingest. Especially when there is no other camera that will use said media.
 
JMtheDP said:
But, now the rest of the room is badly under.
Click to expand...

Of course it is straight out of the camera, but the question is, how much information/data rely in these shadows. If the Cion really has the most DR in the shadows as few testers claim, then there would be tons of usable data there once brought up.

In fact, even the atrociously compressed 720p Jpegs posted here show a great deal of information without noise/grain (Jpeg/Compression square artefacts visible here)

P3Mr6LB.jpg

AfJpn7Z.jpg


So how does the original ProRes 444 4K one look when brought up with a simple curve? I am just still trying to hold my hope in the Cion. Maybe the DR really is in the shadows (you can recover shadows) as I've seen with many photographic cameras vs Canons which had all their range above (you can recover highlights), some cameras have really rich shadows.

Just thinking out loud.
 
I will do some testing myself but yeah that's the size of it. There is a big difference between "detail in the shadows" and "crushed", these are crushed blacks. The chairs themselves are brown leather but lose detail very fast.

What is odd is some of the BM4K footage I have seen looks very nice, but I believe was all raw. More info when I have it!
 
You must log in or register to reply here.
Back
Top