Canon R5 Announced with 8k

It would be cool to have this much DR or more

But that's kind of what I mean. You obviously don't try to hold the sun, shooting into it, but rather expose for sky and buildings.

I don't see a problem why you wouldn't get a similar result with any decent camera?
 
I think it looks better than these examples: https://www.shutterstock.com/video/clip-12400265-sony-pxw-fs7-xdcam--sunset

Are you saying the extra DR is wasted?

I'm not arguing with you that getting your exposure right and distributing your curve to shadows, midtones and highlights in a way that flatters your scene is not important and that you can make smart choices even with limited DR.

Here they made good choices but I still think it's a struggle for ML RAW in these circumstances: https://vimeo.com/70073899

Damn those colors are pretty though
 
Looking at this sample, it looks like the 4K codec does not quite provide as much DR as I would like, but the 5.5K RAW does (granted that it's always impossible to say if it just could have been shot/graded differently to my tastes) - https://youtu.be/MQ7icFkeB74
 
Are you saying the extra DR is wasted?

No, not saying it's wasted. I just wrote more DR is better!

I think of it like this:

HDRstops.png


GradedStops.png


I think the extra shadow and highlight stops lose some significance after grading, which I hope my graphics illustrate.

So why do I say this and still maintain more DR is better? Well. It IS more flexible to have more light to work with. Also, a higher DR camera has a higher SNR. So the higher your DR is when you start, the higher quality you'll get in the end. But I sort of feel that based on the level we are already at (12-13 stops of 6K, 12bit raw, 60fps for $2K), we are more in the 'spit and polish' area, than in the 'make or break'.
 
And in context of actual images. First a mix of sky and then a pan into shadow, wich a look into a restaurant:

Lighthouse_Film.jpg


Lighthouse_grade.jpg


Boathouse_Film.jpg


Boathouse_grade.jpg





And here are some shots into the sun:


BalloonFlat.jpg


BalloonNormal.jpg





A darker, richer version:

BalloonRich.jpg




No effort was made to stop down or rescue huge amounts of highlights here.
 
I think the extra shadow and highlight stops lose some significance after grading, which I hope my graphics illustrate.

So why do I say this and still maintain more DR is better? Well. It IS more flexible to have more light to work with. Also, a higher DR camera has a higher SNR. So the higher your DR is when you start, the higher quality you'll get in the end. But I sort of feel that based on the level we are already at (12-13 stops of 6K, 12bit raw, 60fps for $2K), we are more in the 'spit and polish' area, than in the 'make or break'.

Sure those stops lose significance (probably - unless your actor is in the shadows or something). But isn't the long, smooth tails at the edges part of what make an image pretty? Even if the meat and potatoes is in the midtones where, yes, we want to increase local contrast.

I must say that a lot of Pocket 6K footage looks really good and I would have to agree with you that at $2k it sets a benchmark
 
No benefit at all. And just to be clear: it didn't say lower DR was better in any way. My statement that 12 stops isn't a huge limitation is quite different.
That is why i asked you, “what is the benefit?”, because we are already getting past the 12 stop DR limits, and hopefully moving towards even low end cameras having 15+. Why are we talking about 12stops DR?

I see no benefit for arguing against a few extra stops of DR. Even in an 8bit file the dynamic range can be mapped, and midtones can be preserved.

The problem with 12stops DR is shots that are in motion in natural light will often clip on something at some point in the shot. Having extra DR does not change the image, it just makes the blow out or clipping more organic or analogue feeling. People with a photography background often argue that there are no real limits to 12stops DR, but they don’t often rove in 360 around the scene, and if they do, they likely get some harsh digital clipping. Additional stops would not prevent that “blown out” look, it would just make that blow out look more natural and analogue feeling.

In my experience with 12 stops DR cameras, i’ve never been able to trust my eyes. I have to know what the camera will see from experience. And that is one of the things i hate most: not being able to trust my eyes. The only capture mediums that have allowed me to come close enough to trusting my eyes are Arri Alexa and Film. The Gemini and the Venice are very close, but still not quite. The F55 and F35 were beautiful cameras, but you really had to know how those cameras worked. The F35 was the best 12 stops DR i’ve ever used, but there were still some limits when trying to capture scenes that looked good to the human eye.

Unless you are just trying to help people get over fomo, i see no benefit to arguing for how 12 stops DR is not limiting. It just give manufacturers more an excuse to keep producing mediocre cameras. We make too many arguments for why limits are better, and we are not even close to “too much” yet.
 
Last edited:
And in context of actual images.

This is Pocket 6K? Looks nice. Definitely handles the first couple shots no sweat, but those are not super contrasty.

The last shot I will admit it does a pretty good job, although I think I disagree with the choices you present. I like the feeling of the highlights in the ungraded image. The sun looks a bit harsh once you increased the contrast. (Though maybe I'm not seeing the full quality picture.) Meanwhile, if you were interested in a darker, richer version, why not stop down to begin with and give yourself a little more tone at the top?
 
And in context of actual images. First a mix of sky and then a pan into shadow, wich a look into a restaurant:

Lighthouse_Film.jpg


Lighthouse_grade.jpg


Boathouse_Film.jpg


Boathouse_grade.jpg





And here are some shots into the sun:


BalloonFlat.jpg


BalloonNormal.jpg





A darker, richer version:

BalloonRich.jpg




No effort was made to stop down or rescue huge amounts of highlights here.

If anything, these shots make the case for the need for more DR. The people in the lower part of the composition are under exposed to hold the clouds. And that sunset does not look natural. I don’t stare at sunsets in real life, because it can hurt, but on a TV screen we can look at a sunset all we want, and if it is clipping like that, even though it seems subtle to some, it is very ugly and unnatural at the very end there.

You take nice images, and i don’t discredit your abilities, but the camera does have it’s issues, but for now, that is part of our reality, cameras just kind of suck. I should have done film. Is it too late? Yeah, it is too late, i stopped liking grain 10years ago ;)
 
If the the middle portion of the sun is very strong then it's going to clip for most cameras. One needs 16+ stops to retain any information when it's strong as heck.

There's more room to work with this one though.

yY77G7B.jpg
 
If the the middle portion of the sun is very strong then it's going to clip for most cameras. One needs 16+ stops to retain any information when it's strong as heck.

There's more room to work with this one though.

yY77G7B.jpg

Did you colour in the sun, or was there another file down load somewhere? Because even the log inage that Andree posted, the area surronding the sun was a hard white clip.
 
I recovered some highlights. He said he didn't "rescue huge amounts", so I imagine he didn't at all for the LOG image either, or not much.

The overall grade may be more deceiving though than something more normal and plain.
 
I recovered the highlights. He said he didn't, so I imagine he didn't for the LOG image either.
I’m sorry, i don’t know what you are referencing to “he didn’t”. Could you add a little more context?

There is a strong orange cast to all the highlights in your grade, so it looks more like it was coloured in, than recovered. Like you added an Orange/Teal lut.
 
To evaluate dynamic range, you're probably right. But 2 things -

The R5 will probably have even worse rolling shutter than the 1DXmk3 in 8K mode. In 4K 120p it will be fine. Who knows about 4K 30p and below.

And the other thing is that if you pitch your DSLM as a matching B cam for your cinema camera (be that a C300III, C300II, C500II, FS7 or FX9), it would be good if you could match the gamma of the A camera. Sure, the camera won't match that closely anyway. But it's good to say that the settings are the same, which was always nice about Sony DSLMs offering every flavor of s-log even though it was 8-bit.

But 8-bit s-log out of the Sony mirrorless cams sucked and most guys actually stopped using s-log with them, because it graded so poorly, and went to just a burned-in look and they could match it to bigger cams easier/better.
 
That is why i asked you, “what is the benefit?”, because we are already getting past the 12 stop DR limits, and hopefully moving towards even low end cameras having 15+. Why are we talking about 12stops DR?

When I say "12" there's always some manufacturer that will say "18.5+".

I wish Cinema5D had a more complete library of tested cameras. They use ImaTest to spit out numbers and take noise (where all the digging is taking place) into account. They end up with conservative, repeatable, unemotional numbers.

P6K: 11.9
C300 mk3 (new one with DGO): 12.8
C500 mk2: 12.9
Arri Alexa: 14

So...


EDIT:

The people in the lower part of the composition are under exposed to hold the clouds.

I don't agree here. They sit where they need to be and they are not lifted in post.
 
Last edited:
The C series offers log 2 and 3 in compressed 10-bit.

I disagree with you re: 12 stops of dynamic range. True, the matter is subjective. But I believe one of the reasons people like the Alexa is the high dynamic range. And especially doc filmmakers I've heard from love it because they run into high contrast situations they can't balance by lighting. Or with the C200 - I've heard from people who shoot longform interviews with it in compressed 8-bit. Then they bust out RAW for the b-roll since the contrast will be less controlled.

Personally, I look for a camera that can retain blue color in the sky at midday with a gradual rolloff into the sun, but keep adequate detail on the unlit people in the foreground. Similarly, at sunset. These are situations I run into all the time that I only sometimes want to silhouette.

Looking at Pocket 4K footage, it looks like it can not quite handle that degree of contrast. Pocket 6K is maybe barely there. And the UMP G2 can handle it, no problem.

With the Arri’s, it’s not so much the absolute DR, it’s the roll-off of the highlights and desaturation of the colors instead of them clipping.
 
In photography we have a saying: every photo is a close-up. You can have a photo that's a close-up of an eye's iris, or you can have a photo that is a close-up of an Italian village nestled in the foggy mountains. It just means, be intentional with what your frame contains. Same goes for exposure...

I like this.
 
I wish Cinema5D had a more complete library of tested cameras. They use ImaTest to spit out numbers and take noise (where all the digging is taking place) into account. They end up with conservative, repeatable, unemotional numbers.

P6K: 11.9
C300 mk3 (new one with DGO): 12.8
C500 mk2: 12.9
Arri Alexa: 14
I did a similar test when eating food to test the log size what would later come out of me. I found with repeatable results that if I ate a burrito, the log curve would have a lot of fall off and a very soft texture. On the other hand, when I ate a steak, the log curve was very long and hard. Cinema5D may be able to get repeatable results with their sh*t tests, but they're still just that, sh*t tests.
 
But isn't the long, smooth tails at the edges part of what make an image pretty? Even if the meat and potatoes is in the midtones where, yes, we want to increase local contrast.

I agree! It adds to the flexibility and makes everything smoother. Remember: I'm not in a 'anti DR camp'.

The last shot I will admit it does a pretty good job, although I think I disagree with the choices you present.

The richer version was more a 'cover all bases' and referencing previous posts in the thread. And you are absolutely right about the exposure: if the darker version was my goal, I would have stopped down. I don't ETTR. I place exposure where I expect to pivot contrast in post and the the image fall where it falls*.

I've also made a point to not use secondary color correction or masks. These are global corrections. I've also not beautified the hotspot. There are tricks for that—and some are pretty effective. But this discussion was more about the camera and some context, not the final frame. People post to few pictures when making statements. Isn't this more fun? =)


*= This is what I meant by not trying to rescue highlights in exposure. I will try to not clips normal practical lights indoors, but I don't fight the sun when shooting into it.
 
Back
Top