Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

RED underwater too noisy?

Rudi Herbert

Well-known member
Joined
Jun 1, 2007
Messages
1,433
Reaction score
0
Points
0
I have examined many r3d clips coming from several sources, and no matter where, how and by whom they were shot, I continue to see a lot of bad noise in the blue channel. This concerns me because we're targeting RED for two different projects, both aimed for theatrical projection, where we will do substantial underwater work. Some of it will be done in controlled pool/tank environments, yet a lot of it will be done in really deep water, out in the ocean. A lot of the clips will undergo heavy VFX work, such as a person swimming out of sinking airplane, where we film the person in the water and add the plane afterwards as a background plate. Etc, etc, etc. And I am very concerned because to pull this off, we need the cleanest blues we can get and this seems to be RED's achilles tendon.

Has anybody out there like the folks at Gibcor, Camdiver or Aquavideo done work where the resolution/noise tolerance of the RED has been tested for VFX/big screen projection scenarios? How could this be improved, maybe by adding a few other recording modes in addition to RAW 28 or 36? How about a RAW 98 or RAW 165 (arbitrary numbers) or whatever, something with enough bandwith to preserve as much image fidelity as possible? I, for one, would not mind getting 10 minutes on a RED drive as opposed to 2 hours if it meant I can get the shots I need done. My producers are suggesting film for the underwater scenes, but I've already tested 16mm and it's too noisy, so is HDCAM, even with a fiber cable tethered to an SR recorder on the surface (not to mention VERY expensive), and the Hydroflex or Pace 35mm housings are too big to allow the mobility that some of our scenes require if we are to make these daring shots happen. Besides, I really can't wait to screen dailies to know if we got the shot or not, I need instant feedback to know whether we need to jump in the water again or we can move on. Any ideas, suggestions or words of consolation?
 
All I can say is that Ken Corben from GibCOR will be testing deeper water applications with RED in the very near future. He's the best guy to contact seeing as he is working pretty hard with the units these days. Sorry I couldn't have been more help. Still waiting for my eagle to land.

Best,
Mark.
P.S It may be better to also re-post this topic in the Underwater Thread in the RED User Group section.
 
Blue channel noise is directly related to the color temperature you shoot under, and any optical filtering you might be using.

The Red is daylight balanced, and it looks very clean under daylight, but very noisy under tungsten with no optical white balancing. Assuming that underwater shooting has a pretty high color temperature, you should be fine if its around 5,500K. If its higher, you might wanna use a warming filter, how much will depend on each individual case.

Hope that helps!
 
I have examined many r3d clips coming from several sources, and no matter where, how and by whom they were shot, I continue to see a lot of bad noise in the blue channel.

Have you checked the STS122 R3D clip? It was the first clip i saw with excellent blue channel. I believe that we are going to see better and better blue channels from footage shot with FW15 and FW16.
 
HDCAM too noisy?

HDCAM too noisy?

I shot a feature last summer on a Sony F900 in a variety of very low lit scenarios and I have to say that the noise performance of the camera is exceptional. I screened the footage massively on a Cinealta 4k projector and there was absolutley no issue with noise in any scene. HDCAM does have its issues but I really think noise is at the bottom of the list. I am not sure how red will compete with the F900 but my initial observation is that it does indeed have more noise. BUT this is evaluating it on a LCD screen- something I would reccomend against. I've looked at the grey desktop of avid on some good LCD screens and seen noise there. IMHO never evaluate anything on an LCD screen.

Furthermore- I doubt your delivery is going to be in 4k. Thus you will reduce significantgly reduce your noise in the down res.
 
Agreed,

I have seen significantly more noise from the blue channel of the red camera then any 900 series camera I have shot with. For the "perfectly clean 35mm like image" that the red is touted as having it sure falls apart in the blues.
 
Agreed,

I have seen significantly more noise from the blue channel of the red camera then any 900 series camera I have shot with. For the "perfectly clean 35mm like image" that the red is touted as having it sure falls apart in the blues.

Was this tungsten shot footage?

I honestly haven't been to hot about some of the footage being posted online. A lot of people are doing two things that add up to massive blue channel noise:

1.) Shooting under tungsten with no optical color correction (80A filter)

2.) Under exposing while shooting non filtered under tungsten.

Just because you can white balance in post with the camera, doesn't mean you should. Once more people start understanding the above issues, you'll see more clean Red Footage.
 
Another interesting thing to take note of is sharpening. Sharpening and image noise are closely related in that some people sharpen their footage without trying to prevent signal noise from becoming accentuated. A good unsharp mask and careful threshold settings will improve the perceived noise.
 
Chris,

Thanks, but of course, having shot underwater for the last twenty years or so, I understand well the behavoir of both film and video in the medium and what tricks are used to preserve color balance, such as the warm filters, which are good for shallow water where you want to keep skin tones or marine life colors, etc. This, however, does not do much to ameliorate the fact that blue is the MASSIVELY prevalent color underwater and ANYTHING you shoot there will have more of it than anything else. Warming filters do not do much to keep the blues clean.

Oogabooga,

Yes, HDCAM is VERY noisy underwater, and I'm not even talking about 3:1:1 tape recorded signal, but 4:4:4 SR out of the SDI port. I won't be projecting in 4K, but I still need VERY clean images for manipulation in post. I agree that even the F-900 has a better blue channel than the RED currently, but I really want the 4K resolution of the RED, which for FX work, should be invaluable.

Any other opinions from experienced underwater shooters?
 
Ken Corben is the only one that has shot a significant amount of RED underwater (with our PVC prototype housing) and unfortunately most of it was not available for publication/previewing yet.

I have half dozen housings ready to come on line but most are destined for use with Birger mount so we are waiting for that before we get to seriously test the camera underwater.

I have asked Graeme in another thread about filtering but he was unsure, not having any underwater experience. My thought was that we would want to use an underwater (reddish orange) or RED warming filter to keep the blue channel from clipping but I am curious as to what Cardmaverick was saying. Typically underwater is 7000 - 10000 K or higher.

Blue channel noise is directly related to the color temperature you shoot under, and any optical filtering you might be using.

The Red is daylight balanced, and it looks very clean under daylight, but very noisy under tungsten with no optical white balancing. Assuming that underwater shooting has a pretty high color temperature, you should be fine if its around 5,500K. If its higher, you might wanna use a warming filter, how much will depend on each individual case.

Hope that helps!

Does this suggest that we would want to leave the heavy blues to saturate that channel (to presumably reduce noise) and then dial it back in post? That is the opposite of what I had anticipated - as I said before I thought we would want to filter some of the blue out to keep from clipping the blue channel.

Any other opinions from experienced underwater shooters?

Rudi. We expect to be able to get serious about testing later this month/early april. I'm in Ft. Lauderdale, where are you located?
 
The camera should like underwater because there's more blue. If you can send us an R3D of the noisy clip and details we'd like to look. Are you white balancing back to daylight?
 
Try 8000k - 9000k in water with exception clarity: Carribean and Great barrier reef, 7500k in deeper water. For freshwater lakes, you can set your color temperature lower.
 
I shoot in the water, but its usually water sports at the surface (surfing, Jetski, etc.). As mentioned earlier in this thread, Ken Corben ("Sharkguy" on RED User), my production partner at GibCor, has shot a lot of diving underwater footage with RED ONE since last August - Catalina Island kelp forest, manatees in Florida, and as I write this he is in the middle of a week-long shoot deep in the Caribbean, shooting humpback whales for one of our television co-productions, using one of our housings made by Mike Hastings and our ARRI 8R lens.

Some of our previous underwater footage mentioned above is already offered online by our stock footage representative - Mammoth HD RED Library - and MHD had no complaints whatsoever about blue channel noise when they prepped that footage for the library.

Ken will undoubtedly have some more direct input into Rudi's questions when he returns form the Caribbean this coming weekend.
 
Aquavideo,

I'm in Miami, we're practically neighbors :-) and we actually met before years ago.

Anyway, as for my experiences with blue noise underwater, warming up the image with filters or not actually accomplish different goals. Regardless of the filter, you're still intaking a massive amount of blue light in, only tinted, but still most of it "jumps" into the blue channel. Yes, you improve the skin tones, the clown fish spots or the anemone tentacles, but this solution makes the blue channel actually more difficult to work with, since the magenta tinting of it actually gives you a noise explosion the second you move a slider in any direction in post.

Now, letting the blue channel saturate massively does indeed give you better noise management and the ability to dial it down somewhat in post, but of course, you'll have no skin tones or reds of any kind to speak of. Also, you'll get an effect curiously similar to blue screen spill, where the edges of any shapes in the foreground are rimmed in blue all over. I've gotten around this issue with fixed objects (corals) by doing two passes, one with filters and one without and then compositing them in post, but all this becomes a moot point if you have moving objects. In my case, to complicate things, I'll be doing a lot of shooting tilting the camera up and down a lot, and as you know, the second you go from clear shallow water to deep, dark water, your noise just goes crazy. So, I need the 4Kresolution of the RED but with better blue noise management. How can this be accomplished? I'm thinking uncompressed RAW port, but since that ain't gonna happen no more, then perhaps higher modes?
 
With such a high amount of blue light, I would think red and green channel noise are more of an issue. Red would hurt more than green though, at least thats what I've seen when shooting RAW under very high color temperatures unfiltered, and my cameras (Andromeda) natural balance is around 6,000 K more or less.

Have you considered using a heavy density yellow filter combined with an 85B or 85C filter? That might be your magic ticket. I find that when you get into extreme color temperatures that are not in the photographic norm, you need to start employing certain amounts of secondary color corrections (cyan, magenta, yellow).
 
Does this suggest that we would want to leave the heavy blues to saturate that channel (to presumably reduce noise) and then dial it back in post? That is the opposite of what I had anticipated - as I said before I thought we would want to filter some of the blue out to keep from clipping the blue channel.

You want to optically filter the camera into its natural balance, 5,500K in this case for the Red One. Like I said in my prior post, sometimes you need to apply secondary color correction filters in order to get that perfect balance. Shooting unfiltered is just going to cause problems, like the unbalanced highlight clipping, noise, etc..
 
Chris,

Most professional, actually make that all, of the underwater filters are multicoated for several colors, and in the case of the ones I use, they have three different layers of red, two of yello, one orange, one magenta and a very light purple one. You're right on the money, but as I said, this balances the image and helps with the blue noise some, bit it doesn't really solve the problem if the sensor is particularly noisy in the blues.
 
Rudi,

Have you looked at this yet?

http://ftp.datausa.com/imageshoppe/outgoing/RED_232/blue_channel_compare.png

It comes from this thread:

http://www.reduser.net/forum/showthread.php?t=7454

I certainly hear you on lots of posted footage having blue channel noise in it, but I think thats mostly due to to people shooting under tungsten, or just plain old underexposing way to much and shooting under tungsten, which is a deadly combo for this camera.

Also, I highly recommend looking at 3rd party sharpening options as well. For my own camera, my RAW conversion program has a way to simplistic sharpening filter, so I now render out with it turned off and sharpen in AE, which is where I do my color correction currently. A global sharpening setting that doesn't allow you to determine where the noise floor is just hurts no matter how clean the original image is. I imagine a more complex unsharp mask type filter is down the road somewhere for Red Cine.

Another thing worth looking at is the cameras current ISO rating, Red is saying its 320, but I noticed that in Red Alert, you can select lower ISO ratings like 100 and 50. I can't say too much about it though, as I'm still waiting to get my hands on a camera myself and test that out...

Hope that helps!
 
Aquavideo,

I'm in Miami, we're practically neighbors :-) and we actually met before years ago. So, I need the 4Kresolution of the RED but with better blue noise management. How can this be accomplished? I'm thinking uncompressed RAW port, but since that ain't gonna happen no more, then perhaps higher modes?

We hope build 16 will have some magical improvements.

Rudi: do you have a RED yet? I have #206 but have my hands full keeping up with orders both RED and others. It might be helpful for all of us to let you use my camera and a housing for some testing as I simply can't do it all right now.
 
Rudi: do you have a RED yet? I have #206 but have my hands full keeping up with orders both RED and others. It might be helpful for all of us to let you use my camera and a housing for some testing as I simply can't do it all right now.

No, not yet. We have three of them on order but I figure we might be getting the first one by late summer. I appreciate the offer and will try to take you up on that at some point, but would need time to design a comprehensive set of tests so that we all actually get something out of it, instead of me just playing with your rig for free :-) I'll keep in contact with you and would like to pass by your place in the meantime and take a look at the housing at some point, to get an idea about its ergonomics and the like, so let's go from there.

Thanks again for the offer,
 
Back
Top