TOIA - RED MX Latitude test 2 - HI CONTRAST

[Mark Toia]

Hi ya all...
Since I'm big on sharing...

I'm in London shooting at the moment and once again all I here about is ARRI ARRI ARRI, The big hire joints are just pushing it hard... why? I don't know why.
98% of all Arri Alexia shoots over here are shot using PRORES files only, so none of them will ever get the true grading latitude that camera is actually capable of. They are all using it like a glorified Video camera. It's a real pity.

Anyway, We did a shot yesterday that everyone on set said couldn't be saved. A super hi contrast shot with a 8 stop difference between ends.
The monitor image looked terrible, the quicktime proxies looked just as bad, But all of my previous testing has proven that the shot was not a concern and no lights were needed to fill the scene to lesson the contrast. I was hoping for a cloudy day.

Worried clients, concerned crew... all non believers all said I was mad, that I should light and flood the whole scene with soft light to lesson the contrast.... blah blah blah..
So it was time to convert them to the real world of RED RAW and get them off those silly PRO RES video cameras. (if they are not going to use them properly)
Here is the result.

http://www.zoomfilmtv.com.au/ftp/Dynamic_range/RED_hi_contrast_test2.mov

Another one for Crucial Ordnance I think.

 

[Christopher Barrett]

Sweet Jesus, Toia....

Suddenly I love my R1 even more.

 

[Sidney L. Plaut]

wow!

 

[Sidney L. Plaut]

so you stay in relogfilm even when you export for the TV-master - or do you apply an output-LUT (rec709 compliant?)

 

[Richard Foster]

Bravo. No windows? Looks great. Besides Redlogfilm and CameraRGB what did you adjust, Mark?

 

[Mark Toia]

Bravo. No windows? Looks great. Besides Redlogfilm and CameraRGB what did you adjust, Mark?

I put a small Vignette over the top in FCP... thats it. nothing more.

so you stay in relogfilm even when you export for the TV-master - or do you apply an output-LUT (rec709 compliant?)

No... I grade in REDLOGFILM and CAMERA RGB in REDCINE X. Export the footage out as a simple PRORES 422(hq) drop it in my FCP time line and thats about that... Broadcast levels are within standards. no rocket science going on.
Just send it out like that on a tape. done and dusted.

 

[M Most]

I would say that they had a point in that if lights had been used to lessen the contrast, the result wouldn't have included the rather telltale clipping in both the clouds and the background. What you got is pretty good, and certainly a good demonstration of "saving" a shot, but the image would likely have been better and richer if that had been done.

As for Alexa grading on ProRes vs. Arri Raw, I have not found that to be the case. While you usually get a bit more sharpness and clarity from ArriRaw due to the higher resolution, the grading latitude is essentially identical if LogC is used for the debayer. In fact, I've graded dual recording tests (ProRes and ArriRaw recorded simultaneously) and applied the exact same grade to both with the same color result.

 

[Mark Toia]

I would say that they had a point in that if lights had been used to lessen the contrast, the result wouldn't have included the rather telltale clipping in both the clouds and the background. What you got is pretty good, and certainly a good demonstration of "saving" a shot, but the image would likely have been better and richer if that had been done.

As for Alexa grading on ProRes vs. Arri Raw, I have not found that to be the case. While you usually get a bit more sharpness and clarity from ArriRaw due to the higher resolution, the grading latitude is essentially identical if LogC is used for the debayer. In fact, I've graded dual recording tests (ProRes and ArriRaw recorded simultaneously) and applied the exact same grade to both with the same color result.

Thats really unfortunate for ARRI RAW then, Because all our testing with the ARRI, there was no way in hell the PRO RES files would ever grade out as far as a 16bit RAW ARRI file. I think your dreaming. Real mathematics gets in the way of your argument sorry.
Regarding clipping on the shot supplied , I think thats probably more my dodgy 45 second grade... and the demonstration was not about lighting, but about what you can save from a hi contrast shot using RED RAW. If I punched this out a PRO RES 4444 file tha same as the Alexia, there was no way I was going to get the same result. 10bit vs RAW 16bit... I think your testing could be pushed to the limit more.

Also ARRI RAW is awesome.

 

[Sidney L. Plaut]

I put a small Vignette over the top in FCP... thats it. nothing more.



No... I grade in REDLOGFILM and CAMERA RGB in REDCINE X. Export the footage out as a simple PRORES 422(hq) drop it in my FCP time line and thats about that... Broadcast levels are within standards. no rocket science going on.
Just send it out like that on a tape. done and dusted.

what if you export to file and not tape? (more and more tvstations here want a digital master)

 

[Mark Toia]

what if you export to file and not tape? (more and more tvstations here want a digital master)

Same... no difference, all you have to do is make sure your blacks and whites are not over TV broadcast safe and supply a file of there wishes. simple stuff.
FCP and AVID have a simple drop on plugin filter that will do this for you automatically.

 

[Sidney L. Plaut]

ok - pretty simple

One thing I wonder - I shot some films for a major jewlery company here, and I noticed that with half-good transcodes the jewlery was "edgy" (like dv-blocks) and there was alot more noise, than when I did a full debayer- but you seem to do a half-good - did you test it compared to a full debayer?

 

[M Most]

Thats really unfortunate for ARRI RAW then, Because all our testing with the ARRI, there was no way in hell the PRO RES files would ever grade out as far as a 16bit RAW ARRI file. I think your dreaming. Real mathematics gets in the way of your argument sorry.
Regarding clipping on the shot supplied , I think thats probably more my dodgy 45 second grade... and the demonstration was not about lighting, but about what you can save from a hi contrast shot using RED RAW. If I punched this out a PRO RES 4444 file tha same as the Alexia, there was no way I was going to get the same result. 10bit vs RAW 16bit... I think your testing could be pushed to the limit more.

I'm not "dreaming," I'm speaking from experience. Perhaps if I used absurdly overexposed or underexposed images and tried to turn them into something they're not, I would see a difference in terms of noise and/or compression artifacts. But in my world, such tests are not really useful because what I want to know is what can be done with a sensibly exposed image on one format vs. the other, not what can be done with badly shot footage that is the exception rather than the rule. Such tests can be interesting, but ultimately don't represent real world photography done by professionals. Saving a bad shot is usually only going to give you a "less bad" shot, not a great one. But each to their own....

The clipped areas on your image are the typical telltale magenta that is the result of the sensor reaching the clip point. That's why I mentioned it.

 

[Mark Toia]

ok - pretty simple

One thing I wonder - I shot some films for a major jewlery company here, and I noticed that with half-good transcodes the jewlery was "edgy" (like dv-blocks) and there was alot more noise, than when I did a full debayer- but you seem to do a half-good - did you test it compared to a full debayer?

Yes correct, Half debayer is blocky, I just used it for this test for a quick export, Sometimes I use it for jobs when I want a super sharp finish, and you only really see those JPG blocks if you zoom up 400%... but FULL debayer is the best for sure.
Since most of our AD's are only going to air compressed to buggery by the stations and in PAL most times, Half debayer is all you'll ever need for them.

 

[Mark Toia]

I'm not "dreaming," I'm speaking from experience. Perhaps if I used absurdly overexposed or underexposed images and tried to turn them into something they're not, I would see a difference in terms of noise and/or compression artifacts. But in my world, such tests are not really useful because what I want to know is what can be done with a sensibly exposed image on one format vs. the other, not what can be done with badly shot footage that is the exception rather than the rule. Such tests can be interesting, but ultimately don't represent real world photography done by professionals. Saving a bad shot is usually only going to give you a "less bad" shot, not a great one. But each to their own....

The clipped areas on your image are the typical telltale magenta that is the result of the sensor reaching the clip point. That's why I mentioned it.

Looking at your previous posts on this forum, you seem to be a very argumentative over opinionated soul. So I'll not continue this debate with you as I think it's non productive and you just really enjoy an augment than a realistic chit chat.

Have a nice day.

 

[Zach Gray]

Mark,

Thanks for sharing the test. I think some of the projects may like shooting Alexa because what they see on the monitors is closer to the final result and seeing rich images on the monitors while you shoot builds confidence. I think the client would be more confident as well, to know from seeing, rather than having to trust the dp. This is a different approach than shooting with film and not seeing anything, but hey, it's a new world.

I remember an early test Rodney Charters did with the Alexa vs. Red that he shot on his iPhone from the viewfinders. He was raving about how much more dynamic range the Alexa was keeping in the windows -- and he was buying into that based on what he saw in the viewfinders on set, not a grading suite.

I know the Red is capable of loading monitoring LUTs, but I think there is some real wisdom in what Arri and Canon have done with their rich default look. It makes the right thing the easy thing. If your Red on set monitor had showed the range, there would have been no drama on set.

Again, thanks for the clip.

 

[Mark Toia]

100% agree...

Mark,

Thanks for sharing the test. I think some of the projects may like shooting Alexa because what they see on the monitors is closer to the final result and seeing rich images on the monitors while you shoot builds confidence. I think the client would be more confident as well, to know from seeing, rather than having to trust the dp. This is a different approach than shooting with film and not seeing anything, but hey, it's a new world.

I remember an early test Rodney Charters did with the Alexa vs. Red that he shot on his iPhone from the viewfinders. He was raving about how much more dynamic range the Alexa was keeping in the windows -- and he was buying into that based on what he saw in the viewfinders on set, not a grading suite.

I know the Red is capable of loading monitoring LUTs, but I think there is some real wisdom in what Arri and Canon have done with their rich default look. It makes the right thing the easy thing. If your Red on set monitor had showed the range, there would have been no drama on set.

Again, thanks for the clip.

 

[Matt Steinauer]

Great test Mark of what's really in the RAW. I know everyone likes to see rich images on the monitor but film has been shot for clients for many years with only a less than optimal video tap and clients seem just fine with it.

 

[D Fuller]

Very nice demo, Mark.

 

[Paul Hazlett]

Thanks for sharing very nice demonstration. Could I ask how you set the shot in camera? did you view raw in camera? What metering in camera did you use, histogram, exposure check?

Thanks again,

Paul

 

[Deanan]

Mark,

Thanks for sharing the test. I think some of the projects may like shooting Alexa because what they see on the monitors is closer to the final result and seeing rich images on the monitors while you shoot builds confidence. I think the client would be more confident as well, to know from seeing, rather than having to trust the dp. This is a different approach than shooting with film and not seeing anything, but hey, it's a new world.

I remember an early test Rodney Charters did with the Alexa vs. Red that he shot on his iPhone from the viewfinders. He was raving about how much more dynamic range the Alexa was keeping in the windows -- and he was buying into that based on what he saw in the viewfinders on set, not a grading suite.

I know the Red is capable of loading monitoring LUTs, but I think there is some real wisdom in what Arri and Canon have done with their rich default look. It makes the right thing the easy thing. If your Red on set monitor had showed the range, there would have been no drama on set.

Again, thanks for the clip.

This is contradictory... Rodney was comparing a log image (not a pretty picture image) to a 709 image on red.
The same 709 image on alexa would look very contrasty and clippy also. Red one is already showing the
final image but not the log image (epic has both and red one is getting log soon). The image as is on
red one or epic has a rich default look and you can enhance it more or less very easily.