Dragon noise - do i smth wrong ?!

[Jan Reiff]

stop - breath -
this is not what you think - BUT: still i don´t get it - we shoot with my MX and Dragon today - before that, this small test here in a available light, right dark, end of office floor - kid´s stuff - red car.
yesterday again i heard someone complaining i wanna haveback the "clean" MX

curious again. both cameras in best condition, BS, open lense 1.8 - Dragon with latest FW ...

still, for me, there is NOT more noise in Dragon ... it´s a simply overall better image ...

so - is smth wrong with my MX, or with my Dragon ... ?

Dragon image is so much brighter at same ISO btw

don´t know. what you think ?

no ADD at Dragon. OLPF 2

bigger size here:
https://www.dropbox.com/sh/ylx381yk7k3t0nq/AACX3_iKPZsDts6EMV5sVMRya?dl=0

 

[David Battistella]

Jan,

i think it's all about managing expectations. Your samples show that side by side, same conditions dragon has more. It's how much more that people debate. Some expect improvements, some expect miracles.

Your thread title will will "add to the noise" though.

i still ill think it's like having a whole new camera, especially when you really dig into it, but others will and are completely entitled to disagree.

Battistella

 

[Jan Reiff]

you are right - you always see the massive difference if you have both side by side -

with the option of the OLPF switch this image will be even cleaner if really needed (as you see at histogram, it was not best of condition of light). actually still i see less noise in my dragon images than sometimes in others. don´t know why. not my business, but still curious what´s the reason.

really glad to be on the Dragon track ... with all the options soon - prepared for everything.

 

[Santiago Marti]

Jan, you are right. I have my Dragon with OLPF V2 and couldn't be happier. I think this thing blew out of proportions really quick. The camera is great, you should be happy.

Now Amira will have 3.8 upsampled from 3.4, but nobody will shout IT IS NOT 4K, as everybody shouted back then, but well... it is like that with Red.

 

[David Battistella]

I think this thing blew out of proportions really quick. The camera is great, you should be happy.

understatement of the year.

Battistella

 

[Björn Benckert]

Jan, you are right. I have my Dragon with OLPF V2 and couldn't be happier. I think this thing blew out of proportions really quick. The camera is great, you should be happy.

The dragon new OLPF is great. Dragon old OLPF was fantastic, but yes there where the red orbs...

You need to have shot with both of them understand the difference. I do not know how many more stops the old olpf has in the bottom but it for sure kept the blacks clean really far down and you really have to push it to show the clip to noise floor. I think the frustration comes from when you seen the lowlights of the olpf1 you know thats what you want but without the orbs.

 

[David Battistella]

I think the frustration comes from when you seen the lowlights of the olpf1 you know thats what you want but without the orbs.

Yeah. I was really of the school to keep the blacks clean and I will manage the highlights (like we have for so long on MX). Now that I have those highlights, man, I'd rather manage the blacks a bit more.

There are tradeoffs and each will pick based on what their needs are and because those needs can change well, soon, the OLPF can change as well.

Battistella

 

[Jan Reiff]

i think a good tool especially for rental Dragons may be a information in the display showing "please wait for temperature to achieve best result" or smth like that
i saw, again, some Dragon Shoots and the cameras recorded after the start up without having the best temperature and cal map. i fear a lot of user will not know the cal map infos as they are quite small.

 

[David Battistella]

i think a good tool especially for rental Dragons may be a information in the display showing "please wait for temperature to achieve best result" or smth like that
i saw, again, some Dragon Shoots and the cameras recorded after the start up without having the best temperature and cal map. i fear a lot of user will not know the cal map infos as they are quite small.

What do you keep your temp set at on the Dragon Jan?

No sure about this "child warning labels" on professional gear. It's the same reason actors shouldn't test drive ferrari's

 

[Nawaf Alsabhan]

i think a good tool especially for rental Dragons may be a information in the display showing "please wait for temperature to achieve best result" or smth like that
i saw, again, some Dragon Shoots and the cameras recorded after the start up without having the best temperature and cal map. i fear a lot of user will not know the cal map infos as they are quite small.

I disagree. I've rented a lot of cameras before, and I learned each one before I rented it. Blackshading isn't something new or specific to red cameras.
If someone doesn't learn to use a tool the proper way, they'll get bad results.
Thankfully, manuals these days are so easy to get online, and forums like reduser help users and renters know more.

 

[Jan Reiff]

I disagree. I've rented a lot of cameras before, and I learned each one before I rented it. Blackshading isn't something new or specific to red cameras.
If someone doesn't learn to use a tool the proper way, they'll get bad results.
Thankfully, manuals these days are so easy to get online, and forums like reduser help users and renters know more.

you are right, but i saw and see many mistakes on sets. still.

 

[Matt Ryan]

I honestly don't think the problem is v2 olpf is sooo noisy. The problem is a lot of us shot with v1 and played with v1 samples, so that is what we expected out of the cam in terms of low light noise (which was much cleaner than MX and better with highlights than MX). Then when the new olpf came and changed the calibration and created a noisier low end, we all freaked out a bit. You can't deny it's a huge difference, but then again the guys who wanted better highlights got it. I think if the old olpf didn't have the artifacts we'd all be happily shooting with v1.

This is why we freaked out a bit...(unfortunately shot at different times of day but you can see even with the extra window light in v2 the v1 olpf blows it away)

 

[Scott C]

I honestly don't think the problem is v2 olpf is sooo noisy. The problem is a lot of us shot with v1 and played with v1 samples, so that is what we expected out of the cam in terms of low light noise (which was much cleaner than MX and better with highlights than MX). Then when the new olpf came and changed the calibration and created a noisier low end, we all freaked out a bit. You can't deny it's a huge difference, but then again the guys who wanted better highlights got it. I think if the old olpf didn't have the artifacts we'd all be happily shooting with v1.

This is why we freaked out a bit...(unfortunately shot at different times of day but you can see even with the extra window light in v2 the v1 olpf blows it away)

Hey Matt, 2 questions:

1) In your experience, is the low light of the V2 OLPF comparable/better/worse to what you get with an MX Epic? If not, where would it be in the pecking order? For instance:

V1---------MX--V2 vs V1---------V2/MX vs V1---------MX--V2 (if those diagrams make sense)

2) Someone in the other thread (maybe you) mentioned the easy grading fix for the green tint in the shadows on that last V2 image. What would your procedure be in this instance? It almost seems to me like the tint is more apparent in the bag pockets than elsewhere, which might necessitate power windows etc. to get rid of. Unless I'm missing something there. Maybe just a simple adjustment on the lift is all it needs.

My last concern which has been brought up previously is regarding external recorders.

I actually have another question on that now that I'm thinking about it--

As has been noted elsewhere previously, the image coming out of the SDI port on MX is generally cleaner at any given ISO than the .r3d files you're recording to the card. In the past I've used this to my advantage on projects that are more suited to a 1080p prores workflow anyway, and been able to comfortably go to the 3200 ISO range in some situations when simply recording externally.

Has anyone taken a look at this phenomenon with the Dragon sensor? I'm curious if it behaves the same way, and how pronounced the differences are between V1 and V2 there.

 

[Matt Ryan]

Hey Matt, 2 questions:

1) In your experience, is the low light of the V2 OLPF comparable/better/worse to what you get with an MX Epic? If not, where would it be in the pecking order? For instance:

V1---------MX--V2 vs V1---------V2/MX vs V1---------MX--V2 (if those diagrams make sense)

2) Someone in the other thread (maybe you) mentioned the easy grading fix for the green tint in the shadows on that last V2 image. What would your procedure be in this instance? It almost seems to me like the tint is more apparent in the bag pockets than elsewhere, which might necessitate power windows etc. to get rid of. Unless I'm missing something there. Maybe just a simple adjustment on the lift is all it needs.

My last concern which has been brought up previously is just what to do re: external recorders. I actually have another question on that now that I'm thinking about it--

As has been noted elsewhere previously, the image coming out of the SDI port on MX is generally cleaner at any given ISO than the .r3d files you're recording to the card. In the past I've used this to my advantage on projects that are more suited to a 1080p prores workflow anyway, and been able to comfortably go to the 3200 ISO range in some situations when simply recording externally.

Has anyone taken a look at this phenomenon with the Dragon sensor? I'm curious if it behaves the same way, and how pronounced the differences are between V1 and V2 there.

V1 was at least a stop cleaner than the MX on the low end. So it was far superior to MX. It was also about a stop or so better on the high end over MX. So it was superior in every way. The issue with it was the olpf orb flares. Other than that it was stunning and a huge step up from MX. 2000ISO was the new 800ISO with v1 of the olpf. Like Jan's samples the Dragon is brighter than the MX at any exposure setting.

("weakest" to best) (MX is still excellent, this is just for the sake of this comparison) MX and Dragon v2 are neck and neck in low light.
If you shoot low light then the order would be MX(last)---Dragon v2(2nd best)---Dragon v1(best)

If you shoot daylight and like to have better roloff and DR up above MX(last)---Dragon v1(2nd best)---Dragon v2(best)

That was David who said the tint was easily correctable. I prefer not to have to correct it in the first place v1 OLPF didn't need to be corrected in the low end as you can see above.

It will always look cleaner out of the SDI due to downsampling and bypassing of compression. You are getting a "cleaner" image in doing this and retaining more detail. That is how it will be with any iteration of RED camera.

 

[David Battistella]

"Blows it away" is a bit overstated. These examples are less than exemplary to be calling it a "test".

By the time you NR'd and pushed a "not intended for any kind of real use" shot. They do actually come out about equal.

Plus, an overall ADD kind of (general Ruser, not you) fixation on shooting in absolute darkness has also pushed the issue. I am pleasantly surprised when images that have ZERO light being registered on a light meter actually are passable. This kinda blows me away.

Meter reading ZERO (as in, not enough light to make images) I don't expect a ton more out of my RED camera. V2 OLPF.

Processing in RCX at 2000ISO

Battistella

 

[Matt Ryan]

But you're taking a reading at 250ISO? At 800ISO it would have picked up light being registered.

If you think that shot is cool, try it with v1 lol. Again, I'm not saying v2 is bad, it's just for low light there isn't even a comparison really. Can you get good images in lowlight out of v2, yes. But you can get better results from v1... Just my $.02

 

[David Battistella]

But you're taking a reading at 250ISO. At 800ISO it would have picked up light being registered.

If you think that shot is cool, try it with v1 lol. Again, I'm not saying v2 is bad, it's just for low light there isn't even a comparison really. Can you get good images in lowlight out of v2, yes. But you can get better results from v1... Just my $.02

Right and the choice will be there soon too. I think the meter was set to 2000 there though and the flashlight might have had also had a nice big red ball around it.

Battistella

 

[Matt Ryan]

Right and the choice will be there soon too. I think the meter was set to 2000 there though and the flashlight might have had also had a nice big red ball around it.

Battistella

hahaha that orb killed it! The only reason not to use it.

 

[David Battistella]

hahaha that orb killed it! The only reason not to use it.

The orb's could be a bokeh killer.

Battistella