White Balance vs Filter

[JamesD]

It seems (from what I have read, and I'm new to this whole videography thing so I may have the wrong end of the stick) that you don't need a filter to shoot underwater Videos, but you do need to manually white balance prior to shooting (and presumably any time you change depth)

Is that right? Also, assuming that's correct, can you do both (apply a filter below X depth and use manual white balance) to get a better shot, or is that overkill?

 

[ronrosa]

A filter is standard for an underwater video housing and you can get good footage with just a filter. Having access to manual white balance in addition to the filter is better.

 

[ChrisA]

Most consummer level cameras do not have access to the white balalnce function when placed in a housing. So it's a moot point which is best. The filter works OK but of couse can't be "right" for every depth and water condidtion. But I find using the filter plus some post-production "tweeking" of the color on the computer does well enough.

Then there is the question of "What is correct?" Should the color of the fish match what I saw through my mask or should it match what it would look like if I caught it and looked at it on land? There is something to be said for making it look like what I saw when I was there but then I did'nt see any reds and the red was really there. I prefer a compromise and have actually added "blue" back into some shots as the filter was just "to effective" in the very shallow water.

Next, when you cut shots together that were taken at different times and places you want the color to match, So"realistic" won't work. Color is really an art in itself.

It seems (from what I have read, and I'm new to this whole videography thing so I may have the wrong end of the stick) that you don't need a filter to shoot underwater Videos, but you do need to manually white balance prior to shooting (and presumably any time you change depth)

Is that right? Also, assuming that's correct, can you do both (apply a filter below X depth and use manual white balance) to get a better shot, or is that overkill?

 

[shadowr69]

It seems (from what I have read, and I'm new to this whole videography thing so I may have the wrong end of the stick) that you don't need a filter to shoot underwater Videos, but you do need to manually white balance prior to shooting (and presumably any time you change depth)

Is that right? Also, assuming that's correct, can you do both (apply a filter below X depth and use manual white balance) to get a better shot, or is that overkill?

If you have wb, then you can use that. If you have a filter and don't feel like using wb, you can use that.

You can use both if you want, yes that means color correct with the filter on the surface(make it look like no filter at all), then bring the camrea below with the filter still on and correct then too(not sure why you'd want to do this though). You'll have to realize that you get slight, and I'm pretty sure it's miniscule, degradation from the filter itself(the less you put in front of your lens the better from what I've learned). You also drop down a stop or so due to the light being cut from the filter. I don't think you'd get a better shot by using both because of what I mentioned above....WB with the filter on is like canceling out the filter....and basically cutting the light.

I come from surface photography, so video is new to me in a way, but the concepts are the same. More light == better image overall and wider depth of field(more things in focus).

-Mark

 

[JamesD]

Okay that sounds good, I believe my new video camera will allow manual WB through a housing (we'll see when the housing arrives) in which case it sounds like I can ignore filters and go with a MWB instead, next dive I'll have to play

 

[ronrosa]

Okay that sounds good, I believe my new video camera will allow manual WB through a housing (we'll see when the housing arrives) in which case it sounds like I can ignore filters and go with a MWB instead, next dive I'll have to play

Definitely play around with it. The manual WB might not be strong enough to correct without the filter. What camcorder and housing do you have ?

 

[shadowr69]

Definitely play around with it. The manual WB might not be strong enough to correct without the filter. What camcorder and housing do you have ?

I'm curious why you say that the manual wb might not be strong enough to correct without a filter. How would that be the case? have you heard of this before???

Would it be possible to test this case by using the u/w filter and doing a wb on the surface?

I don't mean to sound rude in questioning you, I'm just very curious now.
I'm wondering if there is something that I missed or missunderstood about wb. :06:

Thanks, :10:
Mark

 

[ronrosa]

I'm curious why you say that the manual wb might not be strong enough to correct without a filter. How would that be the case? have you heard of this before???

Would it be possible to test this case by using the u/w filter and doing a wb on the surface?

I don't mean to sound rude in questioning you, I'm just very curious now.
I'm wondering if there is something that I missed or missunderstood about wb. :06:

Thanks, :10:
Mark

There are limitations to a camcorder's manual white balance. For example, it's not going to turn something black into white or make a white slate look white if the lighting is really dark.

My TRV900 is a pretty good camcorder. Using manual white balance with the filter it makes the colors at 90 feet look like it was shot at 20 feet. I've never tried manual wb without the filter, but I imagine a cheap camcorder's manual wb will not be as strong as my TRV900's and therefore might not work well enough to be used without a filter, especially at deep depths.

I'm just a hobbyist with an opinion. Best bet is to try it out for yourself with your own equipment.

 

[ChrisA]

I'm curious why you say that the manual wb might not be strong enough to correct without a filter. How would that be the case? have you heard of this before???

Mark

Think about what the camera does internally....

Look at just a 2 by 2 block of pixels on the sensor. each pixel has a color filter over it. In each block of four pixels there is one red, one blue and two green. This information is (simplified here) stored as four black and white pixels and one color pixel. The color pixels id four times larger than the black and white ones and overlays then and color is rcorded as the direction on the color whel nd an intensity value. All of this is kept in only 8 bits of resolution.. There is some calulation involved here going from the sensor's RGB to the miniDV tape's YUV, block compressed format. The camera does the math using low presiion intergers, in other words there is a lot of rounding going on where small values become zeros.

Botto line is that if the camera tried to do an extream color shift, like say boosting red by a factor of 100 you'd have just plain wierd looking (posterized) results in the chrominance channel.. So the engineers limit the range of white balance to keep round off and quanitazation errors from showing in the results.

You can fix this by throwing money at the problem. If you have three CCDs you can then independently adjust the gain of the ampliphiers in each chip. Of course then you introduce noise but you combat that by using larger CCDs that produce more signal. Follow this through and you have a profesional level camera with three large sized CCds that costs a few grand Cheaper camera do the work inside the digital signal proceesor of course more money will buy you a better digial signal processor too. ONe that uses more bits or some day maybe even floating point But consummers demand small, low priced amers.....

The camera are just not tunned for UW light. For eample there are three color filters for each group of four pixels so they use two greens because human eyes see well in green light
in an UW camera we would lprefer two reds. Also the camers are designed with only enough WB range to handle daylight and indoor incandesent lights

 

[JamesD]

Definitely play around with it. The manual WB might not be strong enough to correct without the filter. What camcorder and housing do you have ?

Indeed, well the Camcorder is a Sanyo VPC-C5, and the housing will be (haven't got it yet) a EHS-500X, I've not tried either, and it's not a particularly expensive camcorder, so I'll have to see what I can do with it. White balance seemed to be quite good when I tried a few tests at home (pretending yellow was white, I got some nice blue skin tones!)

But we'll have to see, I'll let you know the results.