Sony A7SIII = $3500, 10-bit 4K120p internal, 16-bit 4K60p RAW out

My theory is that the cartel can make an upscale superzoom fixed mount model. Had it wanted to.

Samsung's 108 MPX sensor is 70mm sq. For reference, 1" is ~ 116, 2/3" is 58, the aforementioned Nikon 1000's is 28.5. A typical modern smartphone sensor is ~ 40-50mm sq and 40-64 MPX.
 
DLD said:
My theory is that the cartel can make an upscale superzoom fixed mount model. Had it wanted to.

Samsung's 108 MPX sensor is 70mm sq. For reference, 1" is ~ 116, 2/3" is 58, the aforementioned Nikon 1000's is 28.5. A typical modern smartphone sensor is ~ 40-50mm sq and 40-64 MPX.
Click to expand...
Cellphone sensors are getting pretty good, and such small sensors allow for lenses that are essentially oxymorons: long, compact, fast, sharp, cheap....

If Apple or Samsung partner with Panavision or Arri, they may end up with some nice colour science to go with all that tech. I should re-download filmic pro, I haven't used that since the iPhone 7.
 
I was just thinking - the other cool thing about using stills zooms on the A7SIII is that they might essentially become par focal if the autofocus can keep the subject in focus as you zoom
 
cyvideo said:
Love it!

Not a wedding but Alister in Norway with the FX9 used the Sony 24-240mm. He admits not the best lens but to maximize his shooting time and avoid lens changing in the snow it delivered. All in all the footage looks more than acceptable for 95% of the shoot situations I would cover. Sure not a big fan of variable aperture lenses as you change shot size but then sometimes compromises are imposed on you by conditions you find yourself in. Sometimes bulk, size and weight being some of them.

http://www.xdcam-user.com/2020/02/fx9-footage-from-norway-2020/

Chris Young

Click to expand...

Yeah, nothing about that footage jumps out at me as awful. And the Tamron 28-200 supposedly trounces the Sony 24-240 in every way.

I was shooting beauty shots of an old masters oil painting yesterday - that's the type of thing where the flaws of a lens will really be seen and felt, especially distortion, color rendition, sharpness, and CA. I wouldn't want to use a superzoom for that! (They have in-house kit anyway and try to avoid renting mine)

The one weird thing about the Chapman article you linked is he says:

"For the Aurora I used a Sigma 20mm f1.4 lens with Metabones speed booster and 4K s35 scan. I felt that the extra stop of light gained from the use of the speedbooster was better than the slightly lower noise that would have been present if I had used the 6K FF scan. "

But isn't it the same amount of light as falls on the FF sensor? You're condensing the light, not adding any extra. I fail to see his logic and I assume the noise results would be equivalent or with a slight advantage to the FF capture
 
cyvideo said:
....Alister in Norway with the FX9 used the Sony 24-240mm. He admits not the best lens but to maximize his shooting time and avoid lens changing in the snow it delivered...
Click to expand...

He also used the Sony 85mm f/1.8, Sony 24mm f/2.0 and Sigma 20mm f/1.4. Had the A7SIII been available that would have been a good companion camera for the wide shots which were mostly at night or low light. The A7SIII has a built-in intervalometer so would have been good for the time lapses, and might have avoided swapping lenses on the FX9 so much.
 
James0b57 said:
... If Apple or Samsung partner with Panavision or Arri, they may end up with some nice colour science to go with all that tech...
Click to expand...

The market is too insignificant for either Apple or Samsung to trifle with. The whole photo-video industry - with the exception of the high end enthusiast/pro market - is on the verge of extinction. Nikon projects ~ 8 million units sold total in 2020-21 globally. There may be an opportunity for someone to take over Olympus's production facilities and their engineering teams but they'd still need to contract a sensor manufacturer, as Sony's share of the market is not conducive to an open competition. It is conducive to Sony's dominance.

PS Sony may "only" have 20% of the camera unit sales but they are hitting close to 50% of the sensors sold (themselves + Nikon + Panasonic + Fuji + 1" Canon).
 
Incidentally, here's a corner crop from the tamron 28-200 at 28mm compared to the same from the sony 24-240 at 24mm:

VS-Sony-3.jpg


Read the rest of Dustin Abbott's 28-200mm review here: https://dustinabbott.net/2020/06/tamron-28-200mm-f2-8-5-6-rxd-a071-review/
 
ahalpert said:
I was just thinking - the other cool thing about using stills zooms on the A7SIII is that they might essentially become par focal if the autofocus can keep the subject in focus as you zoom
Click to expand...

Depending on which lens you are using. Have a look at this video and specifically at 6:58 where the Tamron 28-75mm f2.8 is demonstrated as being parfocal. I have one of these lenses and use it on an A7III and can confirm that I find it is parfocal. Using manual focus I can crash into 75mm pull back and reframe and all is in focus right from 75 to 28mm and that's wide open at f2.8. It's small, light, sharp has virtually nil CA and has a very useful close focus range. For the price, it is a no brainer for video in IMHO. If it got trashed for some reason I wouldn't be as upset as having trashed a G-Master at nearly three times the price. As regards the AF, with the A7III I basically find it as fast and as accurate in 99% of cases as any of the Sony f2.8 E-Mount glass as it easily stays in focus as you zoom when using the AF. Sometime in the near future, I must check out the 70-180mm f2.8 Tamron. Again price vs performance it looks to be a decent buy.

Chris Young

28-75 Tamron


70-180 Tamron

 
Yeah, that looks horrible. But I've seen so many A7SIII videos that looked bad. I've come to realize that using the right LUTs and color workflow is crucial to getting a good look out of the camera. It's easy to do, but so many people seem to be using bad LUTs or grading from scratch.
 
ahalpert said:
Yeah, that looks horrible. But I've seen so many A7SIII videos that looked bad. I've come to realize that using the right LUTs and color workflow is crucial to getting a good look out of the camera. It's easy to do, but so many people seem to be using bad LUTs or grading from scratch.
Click to expand...

Many people think that slapping on a LUT IS "grading".
 
mcgeedigital said:
Many people think that slapping on a LUT IS "grading".
Click to expand...

Whatever it was, something went wrong. It only takes a few sec to check skin with the vector scope. The on-screen titles say SLog2 for some shots but if so why are so many highlights blown? Maybe you called it: they *slapped* on a LUT.
 
How do people feel about the potential of the raw out to the Atomos? Is there a chance it could near Redcode quality? I think it's a good sensor with good colour-science so why not? The only footage as far as I know is Bloom's Paddleboarder. I played with the grade on that and was happy with what I could get. I had concerns about the skin-tones in the close-up but then heard Bloom say he did a lot of secondaries on his face to remove a green tint refection from the trees.
 
I can't speak to all the specifics of how the codecs compare, although my gut instinct is that REDCODE is superior.

In any case, people on this forum have theorized that the A7SIII is only doing a 12-bit readout from its sensor and sending it out in a 16-bit container. I think that may be true based on the very fast rolling shutter speeds and the fact that they're the same at every framerate. I think the camera has 14-bit ADCs for stills. So, I would assume that RED has an inherent advantage in its 16-bit ADCs aside from any other considerations about the image processing pipeline or the anti-aliasing filter, etc

That being said, I bet you'll get a damn fine file from the ProRes RAW recording
 
You must log in or register to reply here.
Back
Top