Dragon Compression Rates and Noise

[Gavin Greenwalt]

No! Just No.

I guess I'm going to have to take some time and make an OLPF post to end all this mis-information.

I wouldn't call it misinformation I would call it "Noise tolerance". It's no more or less valid to rate 5217 at one ISO than another and similarly it's no more or less valid to rate Dragon lower or higher depending on your noise tolerance.

 

[Mark L. Pederson]

I wouldn't call it misinformation I would call it "Noise tolerance". It's no more or less valid to rate 5217 at one ISO than another and similarly it's no more or less valid to rate Dragon lower or higher depending on your noise tolerance.

It's mis-information the way it is presented. I'm VERY snobby about noise - just as snobby as you - and just to say that you get more than 1 stop more sensitivity with the LLO OLPF is very mis-leading. Anyway - next week hopefully I'll get around to an epic OLPF post with lots of examples. I constantly read stuff here that is very, very misleading IMO.

STO OLPF is WAY better is so many ways - but I guess I will need to demonstrate it to get the point across.

 

[Ryan Sheets]

I would love to read an in depth article on this and see examples. Would be huge if you could maybe comment on landscape and wildlife aspects to these two filters as well. Thanks so much!

Ryan

 

[Stacey Spears]

and just to say that you get more than 1 stop more sensitivity with the LLO OLPF is very mis-leading.

Its one of those things where someone posted it once, probably out of context, and it lives on forever. I think the bigger issue of noise is simply the expectation. You have camera X, Y and Z that can produce a clean image at ISO foo regardless of how its achieved in-camera. I am thankful that RED does not perform in-camera sharpening or noise reduction, but sometimes that puts them at a disadvantage to the unknowing.

I wonder if we could live in a world where the RED did not have an ISO control on camera.

 

[Russ Fill]

Its not as simple to say that one OPLF vs another would be better. Thats like saying what ISO is good to shoot with on a Dragon? All the OPLF's have advantages and disadvantages in different shooting conditions. Like golf clubs in your bag. You don't just step up to the T box and grab a putter.. or maybe you do but you have different tools for different needs.
If you look at the Buffalo you can see that we are looking at the shaded side of the Buffalo and that the back ground is full sun. I agree with Bjorn that its probably a little underexposed on the shadow side and that combined with a higher compression brings out the noise more than if you exposed at a hight level.
If you were using the STH OPLF it would handle the brighter background better than say using the LLO OPLF, so maybe that would be a better choice but each setup could cause you to change your choices.
I personally shoot mostly in outdoors in very bright situations, white sandy beach, full sun desert or on the water and I prefer the STH OPLF over the LLO due to the ability to handle the strong hi lights of water, sand, sky, 8 stops over the dark forest floor splashes of light, I deal with all the time so maybe that might be a better choice but I do have both OPLFs in my bag.
An R3D would really help us know what the camera settings were so we all could learn.

 

[Julio Quintana]

It's mis-information the way it is presented. I'm VERY snobby about noise - just as snobby as you - and just to say that you get more than 1 stop more sensitivity with the LLO OLPF is very mis-leading. Anyway - next week hopefully I'll get around to an epic OLPF post with lots of examples. I constantly read stuff here that is very, very misleading IMO.

STO OLPF is WAY better is so many ways - but I guess I will need to demonstrate it to get the point across.

Mark, I also would love to see your examples because if I remember correctly, it was either Jarred or Phil H. who posted early on that the the ST OLPF has a million advantages, but with the trade off of being about 1.3 stops less sensitive than the LL OLPF. If I misunderstood or am remembering incorrectly, I'd be very grateful for you clarification.

 

[Nick Morrison]

Have you guys seen Geoff Boyle's camera comparisons?

For max dynamic range, he rates Dragon Lowlight at 640/daylight.

For max dynamic range, he rates Dragon Skintone at 250-320/daylight.

I have a few questions about his RED workflow - he shot 5K (not 6K), don't believe we know @ what compression (5:1?, 8:1?) and not clear if he blackshaded: all variables which effect "noise", the threshold by which Geoff appears to subjectively be judging dynamic range...so I take his RED DR numbers with a pinch of salt...(ie I worry he's effectively shooting at Scarlet Dragon specs)

BUT...I thought his base ISO numbers were interesting, and useful...

Bear in mind picking your ISO/exposure on each OLPF isn't just about noise...it's also about determining your CMOS smear threshold.

We found lowlight OLPF to be quite smear controlled if we stayed at 680-800.


PS - to Mark's point, I just saw a PERFECTLY shot Dragon Skintone job this week, all at 320...and it may be the most draw-dropping shoot I've seen in a while. Was DUMBFOUNDED by how amazing it looked. So yes...SKINTONE is amazing. All depends on what you need.

For our promo work, we're often at 680-800...and then shoot at 96/120fps...effectively putting us at 3200. So we've so far been keeping it safe with LLO. But looking forward to Skintone!

 

[Stacey Spears]

We found lowlight OLPF to be quite smear controlled if we stayed at 680-800.

Hi Nick, are you saying above ore below that range smearing test worse? Mostly curious with your below the range findings.

Comparisons are difficult to do well. I think Geoff did a great job overall and I appreciate him doing the work. The comparison went smoother this time around.

 

[Nick Morrison]

Hi Nick, are you saying above ore below that range smearing test worse? Mostly curious with your below the range findings.

Comparisons are difficult to do well. I think Geoff did a great job overall and I appreciate him doing the work. The comparison went smoother this time around.

We found that more light, less chance of smear. Smear happened generally due to underexposure. LLO at 640/800 seemed fine in daylight. At night oddly enough, we were pushing higher 1280/2000 with good results.

In general more light you bathe sensor, less chance. Look at Probst's Super Bowl Lexus spot - he shot 320 on LLO in a scenario normally prone to smear (very high contrast, dark scene, many bright practicals....and it looked PERFECT). His workflow seemed telling of how to handle that kind of scenario carefully on a project demanding flawless results.

 

[Matt Bass]

BLACK BALANCE for each dramatic SHUTTER SPEED change!

ALWAYS WARM up the camera and then BLACK BALANCE. I use adaptive fan and wait til the camera is 48/65 and then BLACK BALANCE. Take advantage of the FOUR Sensor Calibrations, for example make one for 48th, 96th, 144th, 192th

Each time you change your shutter then you MUST change the sensor calibration. I have found this method to eliminate noise all the way up to 1600-2000 ISO with the LLOLPH.

 

[geoffboyle]

I have a few questions about his RED workflow - he shot 5K (not 6K), don't believe we know @ what compression (5:1?, 8:1?) and not clear if he blackshaded: all variables which effect "noise", the threshold by which Geoff appears to subjectively be judging dynamic range...so I take his RED DR numbers with a pinch of salt...(ie I worry he's effectively shooting at Scarlet Dragon specs)

Shot at 5K because I was trying to maintain image size across all the cameras and use a single lens for all the tests.
Compression was 5:1, I always use the lowest that I can or what the manufacturer recommends.
Black shading was carried out as recommended.
We worked with AC's and owners who are very experienced in all the cameras.
Post workflow was the same for all cameras.
The idea is to remove as many variables as possible, there are R3D files of the tests, and RAW of every other camera, available to download so you can check for yourselves.

 

[Nick Morrison]

Shot at 5K because I was trying to maintain image size across all the cameras and use a single lens for all the tests.
Compression was 5:1, I always use the lowest that I can or what the manufacturer recommends.
Black shading was carried out as recommended.
We worked with AC's and owners who are very experienced in all the cameras.
Post workflow was the same for all cameras.
The idea is to remove as many variables as possible, there are R3D files of the tests, and RAW of every other camera, available to download so you can check for yourselves.

Thanks Geoff. And no worries - others had claried the workflow too. Thanks for the tests!