C300: The New Canon C300 Mark III and New 25 - 250mm Zoom Lens

robotfist said:
The standard Alexa Mini isn't even a 4k sensor. It's a 3.2k sensor and Arri achieves 4k by upscaling it in-camera.
Click to expand...
It only uses 3.2 in anamorphic mode. When shooting spherical, you're likely only using 2.8K crop of the sensor.

I think the whole down sampling thing is over hyped a bit and the Alexa is the proof. Yes, it can help with sharpness to some degree, but there are many ways to achieve sharpness in a camera. That footage of the C300 III looks sharp as a tack to me. In fact, I probably wouldn't want anymore sharpness than what I see in that sample video.
Click to expand...
Agreed.
"the perfect downsample" was important for HD and 2K. But as we go up in resolution, down sample is relative.
 
robotfist said:
The standard Alexa Mini isn't even a 4k sensor. It's a 3.2k sensor and Arri achieves 4k by upscaling it in-camera. I think the whole down sampling thing is over hyped a bit and the Alexa is the proof. Yes, it can help with sharpness to some degree, but there are many ways to achieve sharpness in a camera. That footage of the C300 III looks sharp as a tack to me. In fact, I probably wouldn't want anymore sharpness than what I see in that sample video.
Click to expand...

obviously beside many other things ARRI did something right :) and now Canon too
I would not call it a hype. It was just a matter of fact that the C300ii and the FS7 were a bit
soft and cameras with an oversampled sensor like FX9, C500ii and EVA1 had a crispier image
 
Anyone own a stills camera and shoot 14bit raw?

Well it is very clear that if you own such a camera and and FS7 the softening in the video camera is the huge compression.

A typical stills camera is 20mb per frame.
 
morgan_moore said:
Anyone own a stills camera and shoot 14bit raw?

Well it is very clear that if you own such a camera and and FS7 the softening in the video camera is the huge compression.

A typical stills camera is 20mb per frame.
Click to expand...

then why is fx9 so much sharper with the same compression?
 
I dunno to be fair - twice the sensor size I guess.

Of course sharpening is an odd thing as some camras (stills and motion) need sharpening in post as the dont sharpen in camera to mitigate the AA filter.

Personlly Ive always sharpened the FS7 and would only evaluate its image after sharpening. Slog is AFAIK designed to be sharpenied in post.


--

this is a vid I did back in the day - designed to show 8bit banding. Really the vid failed to do that and the only banding I was able to produce was driven by compression :)

https://vimeo.com/163942387
 
robotfist said:
The standard Alexa Mini isn't even a 4k sensor. It's a 3.2k sensor and Arri achieves 4k by upscaling it in-camera. I think the whole down sampling thing is over hyped a bit and the Alexa is the proof. Yes, it can help with sharpness to some degree, but there are many ways to achieve sharpness in a camera. That footage of the C300 III looks sharp as a tack to me. In fact, I probably wouldn't want anymore sharpness than what I see in that sample video.
Click to expand...

Proper oversampling & pipeline processing is one of the most important features a camera can have. But what's overhyped, what's arguable is how much resolution is necessary for the brain to look "good enough".

High-quality 2.5K, 3K-ish motion pictures have proven to be more than enough...at least for current brains. That may change since we see a difference in footage from the past decades and future humans might too (between 4K-6K-8K).

[Also, extreme dynamic range might be so normal in 30 years that blocks of overexposed white from history might not make sense for many.]
 
James0b57 said:
It only uses 3.2 in anamorphic mode. When shooting spherical, you're likely only using 2.8K crop of the sensor.
Click to expand...
Alexa Mini is 3.4K in open gate ArriRAW spherical and Amira is 3.2K in ProRes sperical (2.8K in ArriRAW). For anamorphic, I believe the Mini is up to 2.8K.



NorBro said:
High-quality 2.5K, 3K-ish motion pictures have proven to be more than enough...at least for current brains. That may change since we see a difference in footage from the past decades and future humans might too (between 4K-6K-8K).
Click to expand...
Over time Darwin will murder through natural selection those with eye sights that cannot truly appreciate the latest Academy Award winning movie in it's full 16K glory.


________


I'm in agreement here that this whole over-sampling thing seems to be exaggerated. The way I recall it was we had the 5D Mark II revolution and one of the main problems with DSLRs were their sensors had TOO much resolution such that down-scaling 20ish megapixels to HD was causing various issues such as aliasing and moire. And now on the flip side people are saying you need more resolution than what you're recording for the sharpest, cleanest image. So which is it? There's some perfect number for oversampling? Or the old DSLR processors just weren't designed to handle aliasing and moire properly?

https://nofilmschool.com/2011/03/sony-f3-times-light-gathering-ability-red
 
Last edited:
Eric Coughlin said:
The way I recall it was we had the 5D Mark II revolution and one of the main problems with DSLRs were their sensors had TOO much resolution such that down-scaling 20ish megapixels to HD was causing various issues such as aliasing and moire.
Click to expand...

lol. Those cameras were limited on purpose. Video was an afterthought, an accident.

Besides...different world, different mindset. Back then companies had the upper hand and you still had to pay for quality until Blackmagic started making cameras. And then shortly after, Sony had a 3-year market changing impact.
 
"There's a limit to how much resolution you can capture with any pixel-based sensor. Sampling theory dictates that a system can only perfectly reproduce signals at half the sampling frequency (a limit known as the Nyquist Frequency). If you think about trying to represent a single pixel-width black line, you need at least two pixels to be sure of representing it properly: one to capture the line and another to capture the not-line.

Just to make things more tricky, this assumes your pixels are aligned perfectly with the line. If they're slightly misaligned, you may get two grey pixels instead. This is taking into consideration of the Kell factor, which says that you'll actually only reliably capture resolution around 0.7x your Nyquist frequency."

And then there's the problem of color resolution being diminished by the alternating colors of the Bayer array...

https://m.dpreview.com/articles/356...ght-loss-why-we-love-bryce-bayers-baby-anyway

But who cares? Billions of tickets were sold to movies shot in HD that were probably filtered in camera to below 4K resolution anyway, right? As far as I'm concerned, shooting 4K or above is only about VFX, nature films, or marketing...
 
ahalpert said:
But who cares? Billions of tickets were sold to movies shot in HD that were probably filtered in camera to below 4K resolution anyway, right? As far as I'm concerned, shooting 4K or above is only about VFX, nature films, or marketing...
Click to expand...
I dunno, I always bring a telescope when I go to a movie theater and I've noticed a severe lack of resolution in the images. All I saw when I watched Parasite was a single pixel that was often changing color and brightness. I'd hardly call that film-making. The audio was good, though.
 
Eric Coughlin said:
I'm in agreement here that this whole over-sampling thing seems to be exaggerated. The way I recall it was we had the 5D Mark II revolution and one of the main problems with DSLRs were their sensors had TOO much resolution such that down-scaling 20ish megapixels to HD was causing various issues such as aliasing and moire. And now on the flip side people are saying you need more resolution than what you're recording for the sharpest, cleanest image. So which is it? There's some perfect number for oversampling? Or the old DSLR processors just weren't designed to handle aliasing and moire properly?

https://nofilmschool.com/2011/03/sony-f3-times-light-gathering-ability-red
Click to expand...


The problem was that they weren't "super sampling"/"downscaling", they were line-skipping and/or pixel binning to create the video image, which led to the negative artifacts. Very different things.
 
Eric Coughlin said:
I dunno, I always bring a telescope when I go to a movie theater and I've noticed a severe lack of resolution in the images. All I saw when I watched Parasite was a single pixel that was often changing color and brightness. I'd hardly call that film-making. The audio was good, though.
Click to expand...

Pixel peeping tom!

If the movie is good enough, you should be able to focus on a single pixel position and still appreciate the cinematography and story. Watch a different pixel position on each viewing to get the deeper meaning. If you have to watch the whole frame at once to appreciate the film, they didn't do their job right.
 
ahalpert said:
Pixel peeping tom!

If the movie is good enough, you should be able to focus on a single pixel position and still appreciate the cinematography and story. Watch a different pixel position on each viewing to get the deeper meaning. If you have to watch the whole frame at once to appreciate the film, they didn't do their job right.
Click to expand...
I like where you're going with this. If I watch one pixel per viewing, and Parasite was mastered at 2:39:1 at 3996 x 2160, 8,631,360 pixels...I'd need to the view the movie 8.6 million times to see the whole movie. As I was a math major in college before switching to film, I could watch the 132 minute movie eight times per day for the next 60 years, 175,000 times before I died of old age. Of course I'll do workouts while watching the movie to maximize my life expectancy. But we're still short of the 8.6 million viewings needed to complete the film. I wonder if perhaps there was a way to view more than one pixel at a time, perhaps with a mirror system which could reflect other pixels onto the same image plane that my telescope is zoomed into, kind of like pixel shift technology for projectors. If we could setup about 50 mirror reflections to hit the telescope plane simultaneously then I could view 50 pixels at once and finish the movie at the age of 93. I don't know what could possibly go wrong.
 
Last edited:
Wasn’t the C300iii supposed to be delivered in June? It’s June, Canon! I am on 1st place at my dealer pre-orders.
 
B&H says end of June...so maybe July for most people if Canon doesn't start shipping until then (and it's not delayed).
 
You must log in or register to reply here.
Back
Top