New Epic-W vs. Alexa Mini test

[Nick Vera]

I'm sure a lot of us will, when we get a camera!

Cameras are pretty amazing and cheap now in days. But if you are only looking for REDs and can't afford it, there's an insane cheap market for used Scarlet MX bodies!!!!

 

[Rakesh Malik]

If we're going to be comparing the Helium with the Alexa at 1600 ISO, shouldn't both cameras be at 1600 ISO since one camera is being shifted towards the highlights, while the other one isn't?

No, you should test the both at their "native" ISO if you're comparing them.

 

[Karim D. Ghantous]

No, you should test the both at their "native" ISO if you're comparing them.

It has been pointed out in various ways already, but the solution is simple: test cameras at various ISO values and let the reader be the judge.

 

[Rakesh Malik]

It has been pointed out in various ways already, but the solution is simple: test cameras at various ISO values and let the reader be the judge.

Agreed, that would be a more informative test for how each camera behaves, but also no longer a shootout -- still quite a bit more useful for the readers.

 

[Nick Vera]

No, you should test the both at their "native" ISO if you're comparing them.

I thought it was already said the native ISO was 1250, now it's 1600?

 

[Nick Vera]

It has been pointed out in various ways already, but the solution is simple: test cameras at various ISO values and let the reader be the judge.

Agreed!!!!!!!

 

[Rakesh Malik]

I thought it was already said the native ISO was 1250, now it's 1600?

I'm not sure about that yet, but I do know that I'm getting really good results with mine, so I'm not worrying about it too much right now... though I am going to test it to make sure that I know how to get the most with it.

 

[Aaron Lochert]

I thought it was already said the native ISO was 1250, now it's 1600?

Phil's done some thorough testing and has said "balanced ISO" is in the 1280-2500 range.

"Native sensitivity" is a bit of a misnomer. There's an absolute maximum amount of light that can hit the sensor before clipping. There's also a minimum amount of light before noise takes over detail. Of course, this is what we know as dynamic range.

What we know is that both of those thresholds on Helium are left of where Alexa's are. In other words, Helium can't handle as much light as Alexa, but it also does better with less light. So the solution? ND filter, stopping down, or using the Skin-Tone/Highlight OLPF, which any DP worth his salt would know to do. Or another, cost saving option: smaller lighting package.

It's a good frame of reference to put both cameras in the same setup with the same settings, and I understand why they did it, but it's not an "apples to apples" comparison and I don't believe it ever will be. They're different horses for different courses.

 

[AndreeMarkefors]

The native ISO is set at the factory and not changeable by the user. ISO isn't part of the exposure triangle anymore. At least not as long as everything is gain applied in post.

ISO is good for 'rating' the camera towards the result (stop distribution relative to 18% grey) you're looking for. Highlight protection vs texture. If they wanted to use the Alexa as baseline they could determine what they consider its stops above middle grey. They they could set the Helium ISO to the appropriate ISO that gives the same amount of latitude.

 

[Craig Lees]

Yes every half decent red shooter knows that base ISO massively is effected by which OLPF is in the camera since epic dragon days. Helium 800 for STHL OLPF is a really nice highlight retention combo with regular blackshading though I can't stop shooting standard OLPF at 2500 at the mo. Also full 8k res helps massively... Yawn haven't we all heard this a kerzillion times before?

:beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::beatdeadhorse5::mad2::mad2::mad2:

Craig Lees
UK

 

[Hrvoje Simic]

Helium on a fundamental level clips one stop earlier than Dragon.

And Dragon about one stop earlier than Alev III.
So talk about ISO1600 on Helium to match Alev III on 800 leaves a question mark floating above my head.

 

[Nick Morrison]

And Dragon about one stop earlier than Alev III.
So talk about ISO1600 on Helium to match Alev III on 800 leaves a question mark floating above my head.

Interesting. Can't imagine Helium has same DR as Alev III regardless. Maybe they match better at 2000 ISO?

 

[Hrvoje Simic]

Helium is a fantastic sensor and its performance already is a technological breakthrough. Sensors of this pixel pitch typically above ISO 500 rapidly start turning image into crap and in compact cameras manufacturers get away with it with in-camera NR. This is something else.

However, expecting on par highlight performance with ALEV III is not realistic and leads to disappointments. As in general does picking technological features and image performance aspects as sweets in a candy shop. Arri does their own magic to pull this off and the effect is unlike any other motion imaging digital camera, and Red has their own magic to get the high end image quality from this approach. Each route has advantages and price. These are both marvelous pieces of engineering and testimony to effort and brilliance of people achieving this. Battles, contests happening between tools are fictional creations inside heads serving just as temporary purchase decision justification and leading to frustration in the long run.
Cameras are designed to shoot stuff, don't accept awards and don't care. Some are better for some things, some for other. None is best in everything and for everything.

Every Red camera owner owns a device enabling him to achieve cinematic high-quality imagery.

To avoid further polarization and "whose camera is better" ( - for what ? ) it is best using tests like these for what they were designed in the first place, before the pissing contests - to see what each camera does in certain conditions and how to use it properly.

Who has a more competitive personality, wastes time in creative domain and sports would be a more adequate choice for self-fulfilment.

 

[Anton Shavlik]

I downloaded and played with those files for a while. I think Helium STD OLPF, as used in that test, needs to be ~2000-2500 to match arri at 800 as far as clipping point. Arri at 800 then would have more shadow latitude, i.e. deeper shadows at 800 than helium at 2500. That's because the Alev III sensor has probably one to two more stops dynamic range than helium, if you match clipping you're losing shadows and vice versa, you can't match both of those points, the range just isn't there for helium. In shots like the lightbulb 1600 is adequite because the filament of the bulb is many stops over, so 1600 or 2500 will blow out the same. In the nice gradient of the window scene it's better to guess at clipping points.

Weapon Dragon with STH OLPF would likely match Arri's clipping point for the same ISO. And Weapon Dragon has such excellent shadows I'm almost certain it would beat the arri in total dynamic range, though the noise might be differently distributed in the lower end of the exposure.

Erich Roland shot excellent xyla charts for dragon and helium and if you look at their waveform you can see how noisy the sensor is at each stop. Four stops below clip for dragon is actually a little less noisy than helium, even though dragon's noise floor is much higher, and that continues at each stop until the exposure reaches the noise floor. The 8th stop under clipping on helium is exactly as noisy as the 9th on dragon. That's essentially the same thing as having less dynamic range, but you can better feel where you lose range. Sampling 8k vs 6k mitigates a lot of that extra noisyness visually but it's there in the waveform.

I also found it interesting that people find dragon to be ~2/3 stop better in the highlights. Dragon's linear data clips at ~449 in 8 bit, and helium at ~288, which is pretty much spot on for 2/3 stop. Is that a coincidence?

 

[Mark Toia]

flawed test. Simple as that.

 

[Akin A]

Let's get the advocates from each side to do a test together. The biased comments are so annoying.
Agree on test parameters and let the results speak. No excuses!

 

[Humberto Rivera]

What this test showed me is that Red and Arri are involved in High Game of who comes on top, the “Old Cinema Cameras”, or the “New Digital Cameras”, the Old Arri Name or the New High Definition Red Cameras THAT have Arri beaten all over the place in all measurable conditions. Yes Arri was very good in the Old Days of “Film”, but Red has them beaten in the “New Digital Cinema” game. It’s a completion for the “Heart and Souls” of Cinematographers and Filmmakers, it’s that simple. This Test represents and attempt to do that, it’s a fraud.

Humberto Rivera

 

[Ben Scott]

Or you know, it was a well meaning person who knows his way around an Alexa a bit better than he knows his way around a Helium sensor camera and who didn't expect his test to gain so much traction.

 

[David Litchfield]

I think the bias claims in this thread are pretty laughable. Whether you set the base iso at 800 or give it that half stop more at 1250, the Alexa has more DR. I have never seen a Red top a camera test. So either all testers are biased or maybe, just maybe, resolution compromises the things DPs really care about.

 

[Phil Holland]

Interesting. Can't imagine Helium has same DR as Alev III regardless. Maybe they match better at 2000 ISO?

That chart I posted is important. ISO 1280 through ISO 2500 on Standard OLPF for reasons.