Cinema EOS Announcement on 19th Jan - EOS R5C

DLD said:
If Canon does come out with this new C5 or whatever it will be called, Sony's going to have to scramble big time.
Click to expand...

It's my perception that Sony really went for it with the a7Siii and FX6 at exactly the same time as Canon undercooked everything. I'm sure the C300 mkiii is lovely but a non-full frame, non-RF mount camera up against the FX6 at half the price?

Given that Sony sold out their own higher-end users by undermining the FX9, perhaps this wasn't the original plan. Maybe Sony got spooked by the prospect of Canon beating them to 8K but then the R5 landed with a squelch (for pro video users).

It would be great if it was Canon's turn to over correct and give people what they want.
 
ahalpert said:
I believe I see what you're saying, but if you log gamma it and keep giving a lot of data to your midtones, you'll be further and further compressing your highlights and shadows. if anything there needs to be recovered or stretched apart, you might have banding...
Click to expand...

yes.


So, the main benefit for cinematographers in wider dynamic range cameras, is to have cleaner richer shadows and more organic highlight roll off. currently most digital get those hot white circles when they clip, and it can look very digital.

if most cameras are around 12.5 and most people shooting out in the wild find themselves deciding whether to preserve the shadows or the highlights, but then sacrificing the midtones, and that is not always a nice compromise in digital. So, if you add a stop in the shadows, and give softer two stops in the highlight roll off you don’t have to sacrifice the midtones anymore. and that is just 15.5 stops of DR.

you get to 17 stops usable (20+ Canon/Red specs equivalent), and then you might start feeling like it is overkill, because then you have 2more stops in the shadows and 3 more stops in the highlights, but even then i can see people utilizing that to good effect.

Personally, i see 20stops usable (23+ Canon/Red specs) as that point where it is all about post control and vfx uses, and the cinematographer begins to lose control over exposure. But let’s get 14.5 to start, before we worry too much. haha.

We need to have a new baseline for digital cinema, and for me it kind of start with where the Arri Alexa left off: 14
 
ahalpert

The human eye is most atuned to what we call the midtones. our eyes and brains discard the shadows and highlight information, but if that information is not there, and we see a white disc with zero texture on a lamp shade, it tends to look weird and unnatural, and that calls attention to itself.

Why is that important?

Ok, so in order to prevent ugly digital clipping, our wonderful engineers build in a highlight roll off, but now out midtones have a contrast curve that feels instinctually unnatural. so, what to do? maybe back off the highlight roll off and avoid highlights? Well, that is what professionals do. haha.

I’d like to be able to throw some spicy light at some things and not have the camera go all digital on me. So having two extra stops in the highlights buys you a more organic look under more normal lighting conditions, but it still doesn’t do much for you when you start throwing harsh light the sensor and you blue channel clips a couple stops before the others, and not you can shoot this or that, or maybe you want to do some bob richardson jfk style shooting and film is dead, so you need a 20+ stop DR camera than can slow down the harsh fast light.
 
You can only really give as much of the 12bit bucket to the midtones as the ADC(or whatver) is effective at capturing in the first place. otherwise you waste the bits.

Even though Red and Arri claim 16bit precision sensor data being converted to 12bit log, we don’t actually know how effective they are capturing the fine gradients.

Most people can’t see grandient steps or banding in 8bit files, and 10bit files are difficult for trained eyes to see any stepping, and 12bit has 4x’s the possible steps, now double that, double it again, and then double it again, and what the heck double it a fourth time and that is 16bit. I highly doubt that a couple extra stops of DR is going to require us to move from 12 bit log to 16 bit log. that is insanity.
 
ahalpert said:
I believe I see what you're saying, but if you log gamma it and keep giving a lot of data to your midtones, you'll be further and further compressing your highlights and shadows. if anything there needs to be recovered or stretched apart, you might have banding...
Click to expand...


I just really despise people thinking that additional DR is for recovering bad exposures.

yes, it can be used for recovering over exposure, but the main purpose is to allow for better image quality on properly expose shots too. it isn’t all about specs for specs sake.


if you get into Hi-Fi at all, and maybe DLD can chime in here, but with amplifiers for speakers, having a bigger power supply isn’t always just about volume, but also about being able to handle transients and peaks with more precision and nuance. oversimplifying it, but in a way that is how more DR works with sensors on cameras, it is about being able to handle the extremes with more grace.
 
Last edited:
the issue isn't seeing banding in the file that rolls off the camera, though. the issue is if you stretch apart the tones, especially in the highlights and shadows, and then you end up with a gradient of 10 values in your export being occupied by just 2 values from your original capture

so I feel like you could record 20 stops of DR, but if you want to darken those highlights and create greater gradient up there, you need to have enough bit-depth in that part of the curve to do it
 
ah, sorry, i deleted before i saw your reply. i was being hyperbolic for humour, but figured it was twisting your meaning too much.


I think you can do the math though. I just don’t know why you are posing it as a question, so it feels like a hypothetical argument about something i personally am not worried about using at this time.
 
Last edited:
well, sometimes I like to close my eyes and spin my aperture wheel and just take the shot wherever it lands. it would be nice to recover that exposure in case it doesn't look good
 
ahalpert said:
well, sometimes I like to close my eyes and spin my aperture wheel and just take the shot wherever it lands. it would be nice to recover that exposure in case it doesn't look good
Click to expand...

Ah, there it is! haha! you got my humor before, glad for that! :)
 
ahalpert said:
well, sometimes I like to close my eyes and spin my aperture wheel and just take the shot wherever it lands. it would be nice to recover that exposure in case it doesn't look good
Click to expand...

Did you find you answer yet?
 
I think if you're just looking for overhead and you're not going to print down those highlights, then you're right that you won't need more than 12-bit log. If you're recording linear, I think you need more. If you're burning highlight detail, I think the result from more stops in the same curve will be thinner tone at the top and bottom

In a few years, we'll all be gassing out in a new thread about how we need to shoot 32-bit float
 
Andy9 said:
It's my perception that Sony really went for it with the a7Siii and FX6 at exactly the same time as Canon undercooked everything. I'm sure the C300 mkiii is lovely but a non-full frame, non-RF mount camera up against the FX6 at half the price?

Given that Sony sold out their own higher-end users by undermining the FX9, perhaps this wasn't the original plan. Maybe Sony got spooked by the prospect of Canon beating them to 8K but then the R5 landed with a squelch (for pro video users).

It would be great if it was Canon's turn to over correct and give people what they want.
Click to expand...

From all the rumors, Sony is likely to respond with A7RV - an 8K update on A7RIV (with the same or updated 61 MPX sensor) and something in the A9III range, which will generally compete against R3. There are currently no rumors about A1II or FX-6II. This doesn't mean that the low end cinema market will be ignored, however. It's just that Sony can sneak its models under a radar if it so wishes. FX-6 was first brought up almost two years before its reveal, then all the gossip disappeared until it was ready.

But it looks like Sony will go for the hybrid market first and only then after the low end cinema. That'd make more business sense,

PS. Global shipments for the ILC's were roughly the same in 2021 as in 2020, ~ 5.3 million units. The DSLR's were slightly down, the mirrorless were slightly up. The P&S/fixed mount was down to 3.5M or so. That market has sailed.
 
James0b57 said:
if you get into Hi-Fi at all, and maybe DLD can chime in here, but with amplifiers for speakers, having a bigger power supply isn’t always just about volume, but also about being able to handle transients and peaks with more precision and nuance. oversimplifying it, but in a way that is how more DR works with sensors on cameras, it is about being able to handle the extremes with more grace.
Click to expand...

That is correct. Nelson Pass, one of the greatest American amplifier designers ever, has a (relatively speaking) new company called First Watt Its power supplies are huge, as are its heat sinks, but it's all about sound quality at lower-medium volumes.

https://img.canuckaudiomart.com/uploads/large/890982-aa73e125-pass-labs-first-watt-m2-amplifier.jpg

The famed speaker designer Paul Klipsch and his protege Bob Heil had the same ideas. (well, Klipsch had it first and then communicated them to a very young Heil).

PS. Pass was most famous for his Threshold designs back in the 80's and 90's. Back in 1990, I stumbled into the long defunct Beverly Audio on Olympic Blvd. The owner was selling a floor model Threshold for $7,000. He turned it on for me, playing some Rush. It was nothing like I heard before or since. But $7K was far too steep for my budget. In today's money,. it'd be ~ $20,000.

PPS. Paul Klipsh founded Klipsch speaker company, known for its very efficient folded horn designs. Those sounded great off tube amps, once again at low volume. Bob Heil incorporated them in what became a first modern PA system back in 1971. The system was designed in St. Louis for Grateful Dead, which had its previous equipment seized by the Canadian police after arrests on drug charges.
 
DLD said:
That is correct. Nelson Pass, one of the greatest American amplifier designers ever, has a (relatively speaking) new company called First Watt Its power supplies are huge, as are its heat sinks, but it's all about sound quality at lower-medium volumes.

https://img.canuckaudiomart.com/uplo...-amplifier.jpg

The famed speaker designer Paul Klipsch and his protege Bob Heil had the same ideas. (well, Klipsch had it first and then communicated them to a very young Heil).

PS. Pass was most famous for his Threshold designs back in the 80's and 90's. Back in 1990, I stumbled into the long defunct Beverly Audio on Olympic Blvd. The owner was selling a floor model Threshold for $7,000. He turned it on for me, playing some Rush. It was nothing like I heard before or since. But $7K was far too steep for my budget. In today's money,. it'd be ~ $20,000.

PPS. Paul Klipsh founded Klipsch speaker company, known for its very efficient folded horn designs. Those sounded great off tube amps, once again at low volume. Bob Heil incorporated them in what became a first modern PA system back in 1971. The system was designed in St. Louis for Grateful Dead, which had its previous equipment seized by the Canadian police after arrests on drug charges.
Click to expand...

Living in big cities, i find a good HiFi system has grown in appeal to me over the years. i still spend expendable income more on camera support equipment, but as that all levels off, i find myself growing more interested in Hi-Fi speakers and sound to just decompress when i can’t go camping or what not. In the meantime i have some smaller setup that does the trick. I listen at lower volumes anyway. I found HiFi could get really expensive for me, as my preference for movie sound and music sound requires two speak and amp setups. lol. So, i just have forced my self to use one. And i have down scaled considerably and gave my tower and sub system to a friend that has a big space, and kept a smaller system for myself. What a wonderful rabbit hole H-Fi is though. It is possibly my favorite tech/art hobby. As my study in camera tech is more to understand them so well that when i use them i don’t have to think about the tech.

I love the bit about Pass and Klipsch.
 
ahalpert said:
I think if you're just looking for overhead and you're not going to print down those highlights, then you're right that you won't need more than 12-bit log. If you're recording linear, I think you need more. If you're burning highlight detail, I think the result from more stops in the same curve will be thinner tone at the top and bottom

In a few years, we'll all be gassing out in a new thread about how we need to shoot 32-bit float
Click to expand...

i’ve always had a specific bar of quality in mind.

The base for cinema camera, in my mind, is a camera that is good enough to take the place of film.

The ARRI Mini LF is the first camera to be a viable start.

The whole idea of digital killing film, is for it to eventually be far superior. Much like the audio world, where digital has gotten to the point that it no longer has to remain in a purely analog world for quality, there is enough bit depth and nuance in digital files that some of the more aspirational duties can be tasked in digital. That doesn’t replace an 80yo Neumann u47, or a Neve pre, just like there is not replacement for a Kowa Anamorphic, but it does allow for some applications to get good quality without large format tapes, and even allows for fairly convincing tape emulation if needed. But in the early days the difference between digital and tape was much bigger, sort of like the way we are with digital cameras and film.

So hopefully, cinema cameras will start at the Arri Mini LF baseline, and go up from there. Not that all the components have to be of the same quality, like the bayer filter color doesn’t have to be as high grade etc, but one would hope that a cinema camera base entry level has some attributes that rival the celluloid it killed.

4K
14stops DR usable
iso1600 clean
23.98, 24, 25, 30, 48, 50, and 60fps
12bit codec with better than 9:1 compression and a good 10bit SOOC profile at around 400Mbps.

That as a baseline earns a digital cine badge in my mind. anything less is typically marketing. Film might get away with 12stops, but it is an analog process and it can handle extremes with more grace, digital needs more head room to carve off and prevent hard clipping or digital noise.
 
Anyways, to answer your question Abe, yes, 12bit log codec is good enough for a 20stops DR camera. It would be the analog converter or whatever device that is converting the sensor signals into digital form that would need to have higher quality, or as you said if the files were recorded linearly. I think you have it figured out.

i was watching ‘Baracka’ on DVD the other day, and I could still tell it was shot on film, and felt it handled scenes better than the R5C or Red Raptor will. ‘Baracka was shot around 1990, probably older than Norbro, and it is better visual capture medium than the very best digital camera we have today in some regards. I was never a fan of film until digital decided to not improve in 25years other than resolution. And now i find myself potentially abandoning digital and going to film. I’m not in broadcast, and i’d rather spend $100K on images captured in light than a life time of stupid upgrades and like $100K in lenses and filters to make digital easier on the eyes, plus about $50K in cameras and camera upgrades.
 
Last edited:
You must log in or register to reply here.
Back
Top