That was an interesting test. Practical is always reasonable. In this case, it does show where the cameras fall and why people may/may not have issue with it. I’d say somewhere in the neighborhood of 85% use cases, the R5C is perfectly fine. When you push the limits of DR, you push slightly further into the use cases. The A9III gets to about 90% use cases in my opinion. And then the RX hits probably 97% use cases where it can just do about everything. The A9III is a great camera but when it counts, that 7% is the world. Same with going from R5C to A9III, that 5% is the world. As long as people know their cameras and control the light as they should, almost any camera will produce a great image (today) So, nice test. It showed what happens in laymen’s terms and views.
Love it. ISO's below 2000 are useful on the A93 when you don't have an ND handy, and yiou will actually see a little bit of improvement in noise performance over 2000 if pushing in post. You do sacrifice and need to be careful with the highlights in those instances. I think that you will find that the A93 stands up very well to the V raptor if used ISO 2000 and then exposed down with ND. I can tell you for sure that the A93 stands up to the FX6, A7S3 and the A1 when tested at base ISO's and adjusted with ND. Real world, of course you are probably not going to blow it by more than a stop and if that stop is over, I consider all of those cameras equal in terms of the image that they will produce.
So glad I watched this video. I've been testing the R5C and felt it had too much noise in the shadows. I thought it was just me but you're getting similar results. BTW... I always thought your wife was your sister, both of you look a lot alike....LOL. Thanks for the hard work man.
Hi Nick. Is really great hearing narrative filmmakers speaking their knowledge on RUclips. Congrats with the F-Zero Cam. You know in most RAW footage (if not in all) noise is easier to get rid because the debayerig is within the NLE, so the noise is more like a grain with no blocks. Blocks of noise are created in baked-in video extensions when the image processor reads low voltage luminescence signal below the first and second noise floor threshold. Depending on the type of sensor electronics the image processor grab that low voltage signal and mathematically fill the gaps for the almost non existent minimum information to generate detailed RGB values in those zones of the frame, resulting in different patterns of limiting integrals like noise blocks. It's funny that those patterns obey the Brownian movement law. Instead with RAW will always do this mathematical process in the NLE (Davindi for example) and result in a more grain noise pattern rather than block noise pattern . The thing is that with that grain noise pattern you can use like 1% to 9% of temporal threshold, unlink chroma from luma spatial threshold and used no more than 20% chroma ( in Davinci). In a 1080p timeline (can be with 4k raw under the hood) this setting will make magic on RAW footage (yes, even in Prores Raw converted to cinemaDNG: workflow used in The Creator), but with a Long GOP or baked-in video extensions the NLE or Plugin will struggle with more calculations. Years has passed since Neat Video sensor profiler for specific cameras, has evolved the science of eliminating noise superposing sensor noise samples at different ISOs (because the temperature rises when more voltage is feed into the sensor by the operational amp) with many fun tools (something like a sound de-noiser). Sony it's been experimenting with this and other data samples (aside from Giro data), to work around a similar solution like Neat video but inside their camera's image processors, so the noise reduction depend on the ISO noise floor and all the surrounding anomalies that can be mathematically cancelled. Sony has many Japanese years on developing image wise technologies that they re study to resolve domestic and professional problems, and try to get into a sweet stop between the 2: they are the oldest company on the subject. Different companies have different priorities. That's why I find very interesting what Nikon is doing now (even more buying Red Cameras). All that and more, from the in-camera processing world, is very intriguing but I feel that RAW is freedom and timeless solutions for anybody that want to know this truth behind different qualities. I also think that Cined could make more Imatest tests with raw (Prores Raw converted to CinemaDNG) instead on only long gops, I assure that they will deliver very interesting results that WILL change many people mindsets in all the internet, because one start to understand the different raws from stills. Example the raw NEF from Nikon, the raw ARW from Sony, the raw CR3 from Canon, the raw RW2 from Panasonic ( RAW video use something like this still frames per second), and they are all very different. We people need more tests that debug this to make more informed decisions that can impact on cinema gear technologies ahead. Greetings
Loved it. Imo the Red and Sony footage feel similar in many ways. Both initially noisy, but clean up nicely in post. It’s clear the Red has an edge… but that A9iii is not an actively cooled sensor. We can only imagine what this sensor would perform like in an FX6Mk2 recorded as internal BRaw (which we know is coming to the Sony FX line). Imo, this is the sensor that should have been used in Burano. Keep it up!
Good test, thanks. 7:28 Tell me, is there an extended sensor backlight mode in most RED cameras? I would like to use this by buying a RED DRAGON 6k camera. Is it possible?
This sensor in something like an fx6 body would be amazing but doesn't seem like that's happening anytime soon. Sony's line up is so strange were they'll sell you a 30 grand cinema camera with an arguably worse sensor than a 2k vlogging camera
I'd say the Burano has also been unfairly panned. It's a very decent camera with very respectable latitude rivaling RED and ARRI's full-frame cameras (again, thanks to CineD for providing such great standardized testing). Plus it has autofocus, and even a stabilized sensor. It's a very capable camera. However for me, I strongly prefer the V-Raptor X as my main camera, plus a smaller autofocus one for more nimble work. But the Burano is definitely a defensible choice for many. However, put the A9 iii sensor into the Burano or FX6 body and ... wow, that would be pretty compelling.
I mean yeah, if you throw away highlight information, you gain in the shadows. No news here. If you align the clipping points the footage would fall apart
Keep in mind that for every group of shots, the shutter speed and aperture were held constant, with no ND. So for underexposure, nothing can be made cleaner simply by changing a setting. This accurately shows what kind of shadow information is available for each camera. For overexposure, yes we can protect the highlights by shooting differently, but again that would require changing shutter, iris, or ND which was kept constant.
@nick_salazar just switch the Canon into second gain, if all you care about is shadow information. Because that is basically what you are doing on the Sony. Nobody claimed the Sony has terrible high ISO performance (which is basically what you are testing here). The claim is you have bad dynamic range. Which is true. Because it has insufficient full well capacity. The A9 iii has Basically all the same shadow information every Sony had for the past decade. It lacks 1.5 stops in the highlights. This is the issue
@@AcastroAXlmao the canon OOC footage actually looks worse Its weird how canon went from better colors to worse OOC colors than sony 😂. Not sure who is in charge for color calibration there. Would’ve never expected that 5 years ago
@ ? I’m agnostic to camera brands lmao. Ex fuji and ex canon. Whatever works best for fashion editorial work…but OOC doesn’t matter for my needs and I still think its funny the lengths people go to justify a camera brand. OOC has always been subjective with some parts objective. Sony has always leaned towards green but more recently magenta. Canon was more magenta now more green. Don’t care much for brands but I think its hilarious that as soon as it switches Canon people are like “yup canon colors are still better”. Makes no sense to my why people are so brand loyal when OOC standards changes over time and OOC doesn’t matter for real work. Does canon pay you? Lol
That was an interesting test. Practical is always reasonable. In this case, it does show where the cameras fall and why people may/may not have issue with it. I’d say somewhere in the neighborhood of 85% use cases, the R5C is perfectly fine. When you push the limits of DR, you push slightly further into the use cases. The A9III gets to about 90% use cases in my opinion. And then the RX hits probably 97% use cases where it can just do about everything. The A9III is a great camera but when it counts, that 7% is the world. Same with going from R5C to A9III, that 5% is the world. As long as people know their cameras and control the light as they should, almost any camera will produce a great image (today)
So, nice test. It showed what happens in laymen’s terms and views.
Have you tried the Sony yet?
Awesome. The V-Raptor X and A9III kicking ass 💪 Thanks for the test.
Yeah the Sony results were surprisingly good. And really this test just reminded me how unbelievably good the V-Raptor X is.
Love it. ISO's below 2000 are useful on the A93 when you don't have an ND handy, and yiou will actually see a little bit of improvement in noise performance over 2000 if pushing in post. You do sacrifice and need to be careful with the highlights in those instances. I think that you will find that the A93 stands up very well to the V raptor if used ISO 2000 and then exposed down with ND. I can tell you for sure that the A93 stands up to the FX6, A7S3 and the A1 when tested at base ISO's and adjusted with ND. Real world, of course you are probably not going to blow it by more than a stop and if that stop is over, I consider all of those cameras equal in terms of the image that they will produce.
Love this test! Thanks man. I never even considered the A9iii before your testing, this is very impressive.
So glad I watched this video. I've been testing the R5C and felt it had too much noise in the shadows. I thought it was just me but you're getting similar results. BTW... I always thought your wife was your sister, both of you look a lot alike....LOL. Thanks for the hard work man.
Yeah, it struggles in the shadows but is really good in the highlights. I always expose the r5c 1.7 + stops above middle grey in log and raw
Great info. Really selling me on the Sony.
Thanks Mark! It’s not perfect, but pretty nice as a B cam for me.
Dang that a9III held up way better than expected. I wonder how it'd do in the overexposure test at iso 2000. Thanks for posting this info!
Hi Nick. Is really great hearing narrative filmmakers speaking their knowledge on RUclips. Congrats with the F-Zero Cam.
You know in most RAW footage (if not in all) noise is easier to get rid because the debayerig is within the NLE, so the noise is more like a grain with no blocks. Blocks of noise are created in baked-in video extensions when the image processor reads low voltage luminescence signal below the first and second noise floor threshold. Depending on the type of sensor electronics the image processor grab that low voltage signal and mathematically fill the gaps for the almost non existent minimum information to generate detailed RGB values in those zones of the frame, resulting in different patterns of limiting integrals like noise blocks. It's funny that those patterns obey the Brownian movement law. Instead with RAW will always do this mathematical process in the NLE (Davindi for example) and result in a more grain noise pattern rather than block noise pattern . The thing is that with that grain noise pattern you can use like 1% to 9% of temporal threshold, unlink chroma from luma spatial threshold and used no more than 20% chroma ( in Davinci). In a 1080p timeline (can be with 4k raw under the hood) this setting will make magic on RAW footage (yes, even in Prores Raw converted to cinemaDNG: workflow used in The Creator), but with a Long GOP or baked-in video extensions the NLE or Plugin will struggle with more calculations. Years has passed since Neat Video sensor profiler for specific cameras, has evolved the science of eliminating noise superposing sensor noise samples at different ISOs (because the temperature rises when more voltage is feed into the sensor by the operational amp) with many fun tools (something like a sound de-noiser). Sony it's been experimenting with this and other data samples (aside from Giro data), to work around a similar solution like Neat video but inside their camera's image processors, so the noise reduction depend on the ISO noise floor and all the surrounding anomalies that can be mathematically cancelled. Sony has many Japanese years on developing image wise technologies that they re study to resolve domestic and professional problems, and try to get into a sweet stop between the 2: they are the oldest company on the subject. Different companies have different priorities. That's why I find very interesting what Nikon is doing now (even more buying Red Cameras). All that and more, from the in-camera processing world, is very intriguing but I feel that RAW is freedom and timeless solutions for anybody that want to know this truth behind different qualities. I also think that Cined could make more Imatest tests with raw (Prores Raw converted to CinemaDNG) instead on only long gops, I assure that they will deliver very interesting results that WILL change many people mindsets in all the internet, because one start to understand the different raws from stills. Example the raw NEF from Nikon, the raw ARW from Sony, the raw CR3 from Canon, the raw RW2 from Panasonic ( RAW video use something like this still frames per second), and they are all very different. We people need more tests that debug this to make more informed decisions that can impact on cinema gear technologies ahead. Greetings
Loved it. Imo the Red and Sony footage feel similar in many ways. Both initially noisy, but clean up nicely in post. It’s clear the Red has an edge… but that A9iii is not an actively cooled sensor. We can only imagine what this sensor would perform like in an FX6Mk2 recorded as internal BRaw (which we know is coming to the Sony FX line). Imo, this is the sensor that should have been used in Burano. Keep it up!
Sony slightly denoises the footage in camera. I don't mind it, but thought I would put it out there. The Raptor is super impressive.
Thanks for this. Super interesting.
Good test. I’m not about to switch to Sony (too much RF) but it confirmed my decision to sell the R5c and get a Red K-X or C400.
This was awesome! I have always been super interested in comparing the Raptor to Sony stuff. Thanks for sharing!
Cheers for this review.
Good test, thanks. 7:28 Tell me, is there an extended sensor backlight mode in most RED cameras? I would like to use this by buying a RED DRAGON 6k camera. Is it possible?
EH mode is specific to the Raptor-X. Some older RED cameras had a similar mode called HDRx.
This sensor in something like an fx6 body would be amazing but doesn't seem like that's happening anytime soon. Sony's line up is so strange were they'll sell you a 30 grand cinema camera with an arguably worse sensor than a 2k vlogging camera
I'd say the Burano has also been unfairly panned. It's a very decent camera with very respectable latitude rivaling RED and ARRI's full-frame cameras (again, thanks to CineD for providing such great standardized testing). Plus it has autofocus, and even a stabilized sensor. It's a very capable camera. However for me, I strongly prefer the V-Raptor X as my main camera, plus a smaller autofocus one for more nimble work. But the Burano is definitely a defensible choice for many. However, put the A9 iii sensor into the Burano or FX6 body and ... wow, that would be pretty compelling.
I mean yeah, if you throw away highlight information, you gain in the shadows. No news here. If you align the clipping points the footage would fall apart
Keep in mind that for every group of shots, the shutter speed and aperture were held constant, with no ND. So for underexposure, nothing can be made cleaner simply by changing a setting. This accurately shows what kind of shadow information is available for each camera. For overexposure, yes we can protect the highlights by shooting differently, but again that would require changing shutter, iris, or ND which was kept constant.
@nick_salazar just switch the Canon into second gain, if all you care about is shadow information. Because that is basically what you are doing on the Sony.
Nobody claimed the Sony has terrible high ISO performance (which is basically what you are testing here). The claim is you have bad dynamic range. Which is true. Because it has insufficient full well capacity. The A9 iii has Basically all the same shadow information every Sony had for the past decade. It lacks 1.5 stops in the highlights. This is the issue
Z 8
2:43 out of the gate the color on the Sony 😂 leaving the video here. Sad that Sony has never fix there colors
The colors are fine. You just have to know how to color grade. The PTC here was shot on the A9 iii and I’ll bet it didn’t bother you.
@ why downgrade? You get good colors from the R5C
@@AcastroAXlmao the canon OOC footage actually looks worse
Its weird how canon went from better colors to worse OOC colors than sony 😂. Not sure who is in charge for color calibration there. Would’ve never expected that 5 years ago
@ worst than what? Sony lol. I now you are a Sony shooter when you think Sony colors are great. 🤡
@ ? I’m agnostic to camera brands lmao. Ex fuji and ex canon. Whatever works best for fashion editorial work…but OOC doesn’t matter for my needs and I still think its funny the lengths people go to justify a camera brand.
OOC has always been subjective with some parts objective. Sony has always leaned towards green but more recently magenta. Canon was more magenta now more green. Don’t care much for brands but I think its hilarious that as soon as it switches Canon people are like “yup canon colors are still better”. Makes no sense to my why people are so brand loyal when OOC standards changes over time and OOC doesn’t matter for real work. Does canon pay you? Lol
The Red Raptor X. Still the best digital cinema camera available today.
Arri Alexa and Sony Venice both better.
You start by saying it's not Apples to Apples? In a test of Apples to Apples.