The Gentleman's Underwater Bubble Blower Thread

[Johnny Friday]

Bjorn: is there a reason lights were not used in the open sea portions? Sure would make the colors of fish/reef stand out. The end result of this is quite a thing....very talented group to do this 3d work.

 

[Michael Maes]

sure, on the way...
thanks, i'd hold up the mirror and say the same thing about your shot. compared to what you did to get it, i didn't do anything... seconds of 'work' on a lazy, rainy sunday

i'm going to do an in-depth multi-platform tutorial for you guys, just give me a little bit to find some time, but i promise it will happen. i would do it today, but it's such a lazy, rainy sunday

I'm looking so much forward to your tutorial as I definitely could use it big time ))

 

[Björn Benckert]

Bjorn: is there a reason lights were not used in the open sea portions? Sure would make the colors of fish/reef stand out. The end result of this is quite a thing....very talented group to do this 3d work.

There was two light guys on the reef. But on the wide shots they do not help much. We had some quite impressive light gear with us but as it went down, with stormy weathers and all. It was more about if the dive boat would leave the beach or not than if we had proper lights. Personally I was green as the hulk and puking over the railing for the most part, so my concerns where more towards getting it over with than getting the good light. We had 4 days scheduled at Sodwana bay, the first day was really good, not to much wind, still there where about 3 meter swells, as we just got there we figured we sent the divers out to rekke dive and record as much as possible, as the light was good and we know that there where storms coming in. Even I got to dive that day and to say the least on a sunny day Sodwana has a lot to offer. But sure enough second day was quite an overcast, sun was coming and going and the wind started out not to bad but before noon it looked a bit more like a rescue operation than a advert shoot.. third day the same, 4th day the divers did not want to risk going out and they are quite experience, Eric our DOP (also reducer, has a scarlet) for example has filmed the Sardine run a few times and he did not like the conditions at all. The strong winds, lack of sun and relatively deep reefs in combination with low visibility was not all optimal for what we needed to do but all in all I think we got a lot of good shots. Even though it was pretty far from the normal commercial style of controlled environment, pretty far from a tabletop shoot to say the least

here is a screen dump of whats captured and whats possible to push out of r3d's... http://classic.syndicate.se/Files/~usr/hawk/RAW_MAGIC.png

 

[paulherrin]

bjorn... i always love the work you guys do, always awesommme! i really <3 the dogs, y'all did such a great job with them on set underwater and in the final product... so funny

looks like you added in a bunch of water particles and stuff in the final spot, either cg or captured... one thing that always irks me when going to 8-bit or trying to push contrast is that you get really terrible banding in the wide-open blues, as can be seen in the bts/original captures via vimeo... someone pointed out this tool the other day, claims to do magic on banding, unfortunately it's a dark magic because apparently it starts at $13,000??? will have to come up with some home-brewed textures/simulations to use instead.

 

[Björn Benckert]

bjorn... i always love the work you guys do, always awesommme! i really <3 the dogs, y'all did such a great job with them on set underwater and in the final product... so funny

looks like you added in a bunch of water particles and stuff in the final spot, either cg or captured... one thing that always irks me when going to 8-bit or trying to push contrast is that you get really terrible banding in the wide-open blues, as can be seen in the bts/original captures via vimeo... someone pointed out this tool the other day, claims to do magic on banding, unfortunately it's a dark magic because apparently it starts at $13,000??? will have to come up with some home-brewed textures/simulations to use instead.

I dont find it to be to much of a problem the master file for the 30s is fine and for the BTS there is a lot of gopro3 and partly even edited from proress proxy files from the go pro. Normally we go from 16 bit, work in 12bit and master in 10bit HD and broadcast in 8bit pal... Then there is some banding at times but to me thats more of a problem for the broadcasters to take care of. In sweden commercials are still broadcasted in a quite compressed mpeg2 pal format that is about equal to old DVD standards... as long as thats the case I can only care for the master files we deliver and those are usually far from having banding issues. Hope that vimeo gets h.245 soon...

 

[paulherrin]

right, that's what i mean... it shows up in compression/broadcast, but textures (like bubbles, grain, particles) help to break that up and keep the bands from forming in the final output (i.e. the way the thing is viewed)

Hope that vimeo gets h.245 soon...

agreed, vimeo needs an upgrade. i can throw higher quality stuff (bitrate/resolution) on youtube for free. is it even possible yet to view/encode h265 publicly?

 

[Michael Hastings]

thats some great thoughts Rudi..my problem..other then living in a frozen north country and running a horse ranch is I already have a working red one mx housing..should I just buy the new sony z100 4k Gates or Amphibico housing and say thats it ..we have 2 gates ex1r units they all work great and nobody ever complains..even a small canon xf105 amphibico housing we have also looks great on tv..tired of chasing the upgrade tail........even the blackmagic 4k toy footage looks great other then no slow mo stuff....ALSO CHECK OUT BACKSCATTER..they introduced a cannister 22,000 lumens video light for $3500.00.....on my list or the spring I think..I shoot deep wrecks..need tons of light

Jim: You have raised some good points. IMHO I would make two points:

1) Although there are some nice features of the Gates housing, I doubt if they would make much difference in the actual images you acquire, particularly when the likely lenses you are going to use are a rectilinear wide zoom like the Sigma 8-16, the tokina 11-16 or Canon 10-22; a fisheye like the Canon 8-15 or Tokina 10-17; and a macro lens like the Canon 60 or 100 macro which are easily handled by your current housing. I agree with Rudi that the Gates housing is an amazing piece of gear, and the port support for all of those Hollywood style PL zooms and such is great as well, but it seems that most documentary and wildlife shooters use the lenses I mentioned or something similar. In the end you are shooting with a raw recording camera where not much that you do makes a difference in the end - it boils down to the lens and the port and there isn't much difference there - our (and most other) acrylic domes and glass domes are very similar if not virtually identical. And there is even less of a significant difference when it comes to flat ports for macro use. And it doesn't mean that you can't use real PL cinema lenses - your housing easily accommodates Ultraprimes for example, we just never made the various ports for the variety of larger zooms and such that get called for in Hollywood feature work. In short, it doesn't seem to make a lot of sense to get a second housing for a six year old camera given all of the new models out there right now.

2) Although I hate to say it, as reflected in the post above about Peter Jackson using a GoPro for an underwater shot in the Hobbit and your own experience using and selling footage from "lesser" cameras, it seems to me it is getting harder and harder to monetize the high dollar super systems like Epics/F55s etc. given the market is somewhat saturated with cameras and housings for rent or shooters with their own packages that might be even cheaper!

You mentioned the Blackmagic camera. Well the current 2.5K camera for $1995 LIST INCLUDING A FULL VERSION OF RESOLVE! has to be the probably the best bang for the buck ever in a video/cinema camera.

13 stops DR from large pixels - basically same size pixels as RED MX sensor
Shoots open source RAW
Shoots ProRes with no add-ons
Shoots DNXHD with no add-ons
Canon Smart mount - easy electronic iris control
Uses off-the-shelf SSDs - Raw is not as compressed as RED, but you can buy 12-15 256 SSDs for the price of one proprietary 256 gig REDMAG
Internal lithium battery 90 mins and also backs up if you use external battery or AC power.
Easy external DC (extra AC adapters for $20 on amazon) Common dc plug allows easy use of almost any 12-30volt battery.
Has nice 1/4-20 hard points and solid body for easy mounting in housing or addition of accessories for regular land shooting.
LANC control if desired
Large touchscreen monitor

all totally integrated in compact, very housing-friendly form factor (easy controls for power, iris up/down, auto iris, menu and important menu functions)

4K camera is same body as 2.5K so just pops into the housing when available.

Housing/camera package for a little over $4K - half the size of an epic housing with lots of lens and battery flexibility, even smaller with certain limitations.

Some of this gets into philosophy, and given some of the recent requests from some of the bubble blowers without housings, when I get some time I'd like to do a long post - if it wouldn't be take as too much of a sales pitch - regarding some of my philosophy regarding the various tradeoffs and how I would go through a decision tree to figure out the best use of someones resources to get the best overall system - meaning the whole package, since for example I have seen a number of people use up their whole budget getting a $10K - $20K housing that might have a sexier look, and more cool features and controls, but really doesn't get a better image - and then have nothing left over to get a good set of high-powered lights that would make a dramatic difference in the quality of images they get.

 

[Jeff Whitehurst]

Michael, I don't think the water color in your source image is all that bad, though it looks murkier than I remember in the Caymans. How was the viz on this day?

 

[paulherrin]

jim/michael, i'd buy a used/bt r1 before i grabbed a bmc... personally. i've seen them in action and i've seen the images - they're alright. at the end of the day you want to feel good about whatever you're going out and capturing with. and then the next day you want to be happing when you're looking at them until the project is finished and after when someone else sees it. if you're happy, i'm happy. personally, i'd be happier with the red.

 

[paulherrin]

Michael, I don't think the water color in your source image is all that bad

i agree, but it's obvious to me that the color science is not designed for this particular situation - you can tell when you slide around the white-balance for debayering. i highly doubt that was the color of the water. compare it to the water and sand in your profile picture and tell me there isn't a magenta tint... this also brings some of the brilliance/saturation out of the image and gives it a kind of murky feeling because that tint essentially neutralizes the blue-green.

although, the reason there is a magenta tint is because the r3d was literally set to -16.684 magenta tint as a way of trying to compensate for the white balance of the water. it's totally possible to get all kinds of water-colors by adjusting the raw white-balance point (which is the most important point), but then the rest of your colors will be off - some of which probably has to do with daylight sources that effectively have less filtration than the sunlight shining through a whole bunch of saltwater at various levels of filtration itself - so you're dealing always with mixed sources in a very foreign and bizarre environment for color science testing. personally, i think the adjustment here lies primarily in post-processing (debayering) and post-correction. but it's super tricky. just remember we're always lightyears away from breakthroughs, until that one second where an idea comes together, the best thing at this point is to build a database and have an awareness of everything going on underwater, and it sounds like that's exactly what's happening. be aware that you're asking for a lot with a 'fix'... the simplest fix at this point, although imperfect, is in post-processing as i've tried to suggest and demonstrate, and will continue to do so in more detail until it sinks (or floats) or until a breakthrough occurs.

 

[Dave Blackham]

the thing is, far blue and UV usually get fixed with a UV filter and even better if a red filter is used in conjunction. This does not seem to yield any appreciable differences in the RED, while it does on other cameras. So, yes, while the trigger factor is indeed the blue/UV contamination, even more dramatically so on certain locations, the MX sensor can't still recover after all those fixes have been applied. So there has to be something in the RED image path that contributes to the poor performance and which we can't still figure out.

BTW, having seen lots of stuff shot my Michael in Grand Cayman before, I can tell you that without the 20,000 lumens, that image would look absolutely horrendous, without any color saturation or separation whatsoever.

I agree re Red (filter) filtration and a suitable UV filter appears to help. We are not seeing the issue with artificial light or in green water so the issue appears to be common to sunlight. We still need a fix for this problem on MX cameras. When shooting on other more recent cameras to market in RAW, Magenta casts are also a concern. Im guessing the issue is more generic than we first thought. Im not wanting to target any camera system critically other than saying this issue needs fixing to help make better images.

 

[Michael Maes]

Michael, I don't think the water color in your source image is all that bad, though it looks murkier than I remember in the Caymans. How was the viz on this day?

Jeff and Paul, you guys remind me of Will Smith in Man in black: "I make this look good!".... where the "I" are you guys )
Well, while this looks good it is definitely not how it truly was that day.
The viz was not super, to Cayman standards, meaning.... 80-100ft ;-)
Granted it was a bit "cloudy" but certainly not "murky"!
Check my wife's shot of a reefie to get an idea.
Same spot but, say, 55-ish feet in stead of 72 feet.

She also has a snapshot of the same grouper at the same spot.
Bad picture, just for color purposes.

Shot on a D800.

So, baseline, ... you gurus are clearly able to "make this look good" but it is not how it should look like.
Anyway it is a bezillion times better than what I can "cook" in RCX ;-)))

 

[Michael Maes]

i agree, but it's obvious to me that the color science is not designed for this particular situation - you can tell when you slide around the white-balance for debayering. i highly doubt that was the color of the water. compare it to the water and sand in your profile picture and tell me there isn't a magenta tint... this also brings some of the brilliance/saturation out of the image and gives it a kind of murky feeling because that tint essentially neutralizes the blue-green.

although, the reason there is a magenta tint is because the r3d was literally set to -16.684 magenta tint as a way of trying to compensate for the white balance of the water. it's totally possible to get all kinds of water-colors by adjusting the raw white-balance point (which is the most important point), but then the rest of your colors will be off - some of which probably has to do with daylight sources that effectively have less filtration than the sunlight shining through a whole bunch of saltwater at various levels of filtration itself - so you're dealing always with mixed sources in a very foreign and bizarre environment for color science testing. personally, i think the adjustment here lies primarily in post-processing (debayering) and post-correction. but it's super tricky. just remember we're always lightyears away from breakthroughs, until that one second where an idea comes together, the best thing at this point is to build a database and have an awareness of everything going on underwater, and it sounds like that's exactly what's happening. be aware that you're asking for a lot with a 'fix'... the simplest fix at this point, although imperfect, is in post-processing as i've tried to suggest and demonstrate, and will continue to do so in more detail until it sinks (or floats) or until a breakthrough occurs.

Hey Paul,

Having tint set to -16 is, for Cayman u/w footage, VERY moderate.
Normally I have to go from -25 to -45 !!
In other waters that is totally NOT the case.

by adjusting the raw white-balance point (which is the most important point)

Darn man, can you talk "rookie colorist language" please? ;-)
I have no idea what you're talking about! (where's that highly expected manual-to-be of yours ) )
Cheers,

Michael

 

[paulherrin]

haha, those numbers don't surprise me, but there seems to be a balance between subject/object and background in trying to get the color temperature set in rcx and the simple fact is that underwater may be the most difficult position for color-accuracy.

so, rookie language (if it's even possible)... basically, white balance is the most important setting in processing an r3d because that is the reference point for interpreting the raw data. it's like when a child sees a color, but doesn't know what to call it... you have to tell the child what color they see. with underwater footage, there is no one wb to rule them all. i've used numbers all over the place for shots that were in the exact same conditions. just the reality of trying to get a good interpolation, you've got to understand that graeme is studying light and color scientifically (hence, color-science) and as y'all know better than me, to study something with the scientific method it must be repeatable, and conditions are always changing and all over the place underwater due to an endless number of factors that are hard to control for. so i assume he's trying to build a database of real world examples (which is good) to do a best guess on that end. i'm just guessing myself on most of this stuff, i could be way off... and there's a whole nother possible 'issue' of olpf's and sensor engineering that factors in, which is where my knowledge goes dark.

 

[Michael Maes]

haha, those numbers don't surprise me, but there seems to be a balance between subject/object and background in trying to get the color temperature set in rcx and the simple fact is that underwater may be the most difficult position for color-accuracy.

so, rookie language (if it's even possible)... basically, white balance is the most important setting in processing an r3d because that is the reference point for interpreting the raw data. it's like when a child sees a color, but doesn't know what to call it... you have to tell the child what color they see. with underwater footage, there is no one wb to rule them all. i've used numbers all over the place for shots that were in the exact same conditions. just the reality of trying to get a good interpolation, you've got to understand that graeme is studying light and color scientifically (hence, color-science) and as y'all know better than me, to study something with the scientific method it must be repeatable, and conditions are always changing and all over the place underwater due to an endless number of factors that are hard to control for. so i assume he's trying to build a database of real world examples (which is good) to do a best guess on that end. i'm just guessing myself on most of this stuff, i could be way off... and there's a whole nother possible 'issue' of olpf's and sensor engineering that factors in, which is where my knowledge goes dark.

OK
Yes, "of course" I know what White Balancing is and means. ;-)
I only thought you were referring to another step in post to do WB, another software-function.
Now I use the droplet in RCX as a starting point.
(Well first I correct the exposure if needed)
I try to set my WB more and more when I'm shooting.
This particular shot I did the in-camera WB on that spot, pointing the camera down to the sand from, say, 3-5 feet distance.
I had my lights on, else it's useless to set the WB... haha...

Then again, that probably produced way too much signal in the red channel?

 

[paulherrin]

haha, sry...

well, i think it's good to set it if you can while you're shooting to match as best you can what you're seeing, but it's going to be a tradeoff probably between water color and your subject most of the time.

there are a lot of different methods that work once you get debayered. you could do multiple color passes and combine them, but what i do is get the foreground looking good in rcx/raw and then adjust the water color to where i think it should be, or to where someone tells me it should be... otherwise you can go somewhere in the middle and adjust everything, which in reality is usually how it kinda ends up. the point is, you get a good image. whether it is accurate or not usually has to do with how closely the cinematographer is calling shots tot he colorist, and how well developed and nuanced his or her color memory. in absence of that is only image research. so since i wasn't sitting with you, i should have looked at a bunch of pictures of grouper from that area and done the best i could on both fish and water-color, but that wasn't my point in the exercise. that would only make sense with a sequence, really, where you can also cross-compare different shots/subjects on the same dive/day.

 

[Tom Hamilton]

Introducing color into a world that, with increasing depth becomes increasingly monochromatic, is bound to be an unusually complicated task. That red fish at 50 feet is only red because you shine a light on it that makes it so. Shine a different light, it becomes a different color. It is, naturally, what it is, but we insist on changing it, make it somehow more pleasing. How is that in any way connected to the real world, the one we see in air, the one our eyes were designed for? It is all interpretive in my view, the observer changes that which is observed, through the tools he uses to make those observations.

Very much looking forward to your tutorials Paul, and will be in contact some months from now with a request for assistance. Cheers.

 

[Michael Maes]

Grading is such a difficult matter, that's why there are specialists like you!
Colors u/w are what they are, true, but.... they don't include magenta.... haha... or should I say :-( ?
Happy shooting/grading you all! )

 

[paulherrin]

absolutely good philosophic points, tom... i embrace activism, reconstruction, transformation - that the observer does in fact change the observed and hopefully for the better, or maybe they change the conditions for the observed to succeed (like, say... promoting underwater biodiversity, appreciation, cleanliness... even through subtle methods as simple[ish] as capturing the grace of underwater creatures. or making a tutorial for you guys that makes even better underwater images appear in the world . the camera lens is a powerful observer for change. in my opinion, electric underwater lights (not to mention the scuba tanks, housed cameras, scooters) are very 'natural' given the innate propensity and curiosity of mankind to explore, create, develop, build, change their environment... to me it's not so much about making something look more beautiful than it was, it's seeing something inside that is beautiful and bringing that out from beyond our perception of light and color so that what is seen relates to the subtleties of our subconscious, symbolic memory colors. and underwater is technically more complicated because the eye adjusts when you're down there, so the camera is over-objective. what you think you saw is more important than what you 'actually' see through the camera lens. copies of copies and turtles all the way down. now, that's not realism... everyone and every project has a different perspective, and that's awesome

 

[paulherrin]

Colors u/w are what they are, true, but.... they don't include magenta.... haha... or should I say :-( ?

haha, now there's a healthy dose of realism