Black and White filters on Digital Cameras

I tried using my Black and White contrast filters on my Canon G2 for the purposes of

demonstrating the effect. I set the camera on a tripod and took a meter reading. I

switched to manual and set a shutter speed to include the required filter factor. The

resulting images were taken as colour JPEGs and then converted to B&W in Photoshop.

 

The result was that the stronger the filter the worse the contrast. The red produced a

very flat image. Examining the original image I found the red channel was mostly

blown out. The blue channel was mostly suppressed. I presume the flat image is then

the result of the Bayer interpolation - the dark blue and the light red compressing the

tonal range across the whole image.

 

The B&W mode (which I used to visualize the final image) produced the effects I

expected to see but the image captured in colour and then converted bore no

resemblance.

 

I certainly don't want to reopen the can or worms about whether you can reproduce

all the effects of B&W filters using the channel mixer.

 

Does anyone use B&W filters on digital cameras ? Can someone give me a solid

explanation of this effect ? What does the B&W mode on the camera do - just use the

luminosity from each pixel ?

 

As a side question I know the G2 uses the CCD to meter, does it use only the Green

pixels ? Does the histogram display only the green channel ?

The best way to shoot B&W with digital is to shoot color.

 

Shoot color images all the time, don't waste your time with the B&W modes on any camera. Take the color image in to photoshop and you can select channels, blend channels and do so many things that you have more capability than if you were to put color filters on the camera. It's a very versatile method.

 

If you shoot in color and want to see what it would look like with a red filter, just remove the blue and green channels in photoshop. There's also a channel mixer where you can choose how much of each channel you want to include to produce a B&W image.

Please do me the favour of reading my post before responding.

I've often wondered the same thing. For instance, it's easy to get blown highlights in a blue sky in digital. You expose for those highlights, and you've got a lot of photoshop work to do to get decent shadow detail. And it's still all pretty compressed.

 

Seems to make sense that a filter reducing the blue light would compress the full RGB histogram to more acceptable levels, allowing more subtle shadow detail to be captured. Given the limited dynamic range of digital sensors (and slide film for that matter) it seems like it might be a valid approach. I've always questioned the theory that colored filters are no longer needed because of this dynamic range issue. Unless you are shooting multiple frames and merging them of course.

 

I'll try it out too. Probably depends on the strenght of the filter. But maybe there's some other stuff going on we haven't considered.

Merging exposures is really the way to go if you can but for a filter solution I would

go with an ND grad. I have 1, 2, and 3 stop neutral density filters in hard and soft

transition precisely for holding back the sky in my landscapes.

Don't shoot me I'm naiive but is this an effect of auto white balance?

You were correct the first time.

 

Using filters basically screws up the Bayer demosaicing process. Since you are shooting raw, you are sending bad data into the JPEG compressor, where it gets baked in permanently.

 

Now, your Canon G2 is 2mp. 0.5mp red, 0.5mp blue, 1mp green. When you shoot with a red filter, all you've got is 0.5mp of data, which then gets mashed up by the camera's Bayer demosaicing process. If you shoot in color mode, all the channels get to contribute, and you get at least twice the resolution, and less artifacts.

 

So at this point, using the channel mixer gives you a much better image.

 

Follow the other suggestions, like use a ND grad to hold the sky back a bit and keep it from blowing out, and you're there.

Hi Alistair, I think Joseph is right. Think about it this way: Since almost all digital cameras are essentially color devices, the question about using the filter is a bit like asking if there are B&W filters for color film - something you obviously wouldn't want to do. I'm afraid I don't know what's happening in the cameras that are equipped with a monochrome mode. My 10D won't shoot monochrome, but I've made a few nice B&W prints with it from time to time using Photoshop. I think monochrome gets neglected all too often in our digital cameras, scanners and printers - all seem very color-only oriented. Best wishes . . .

I need to sleep more. Sorry.

 

Auto color balancing will screw it up, so take a manual reading (prior) or use a preset.

 

And as stated the camera is going to apply its standard equations to convert the image data to a color image and if everything is in a single color channel the cameras go haywire (some try to compensate, etc).

 

B&W mode in a camera just reads every pixel and saves it as a grayscale file, really similar to doing a basic grayscale conversion in photoshop.

 

In theory the camera manufacturers could build in the option like you have in photoshop, only read the red pixels, or the green or blue, or combinations. But in my first ignorant post, that's easily done in photoshop and probably not a feature many camera manufacturers want to implement.

 

I'm going to bed now. Been out shooting too much this weekend.

I am still not sure that I know exactly what the B&W mode does. I will try a couple of

test shots and see. The image on the screen didn't seem to match what I got by

simply desaturating. Under strong filtration I would expect a speckled appearance if it

just records the luminosity of each pixel.

 

The Canon G2 is a 4MP camera.

 

Thanks to all that responded. This was an academic enquiry since I use a film camera

and film for "serious" work. I didn't want to burn film to produce some pictures to

illustrate the use of B&W filters for our photographic society's web page.

I don`t think You can overcome the Bayer pattern. The BW mode on digis probably records the image in color and then throws away the color data - kinda funny since CCDs can only "see" luminosity and all the color is made up by the Bayer pattern and some software - actually to me is almost a miracle that the colors come out so similar to reality.

 

So probably using B&W filters in front of the lens just screws up the Bayer pattern and the software that computes color.

 

However, post the test in B&W mode i use the g3 and am curious.

<i>I don't think You can overcome the Bayer pattern. The BW mode on digis probably records the image in color and then throws away the color data</i>

<p>

There is one way that gives you a bit more control: save the image in raw format, then decode it using Dave Coffin's dcraw with option -d to turn off color interpolation. The resulting file reveals the individual elements of the Bayer pattern in grayscale. This gives you a bit more control over grayscale interpolation, e.g., you can deinterlace this image in four different ways (horizontal/vertical odd/even), keep the four deinterlaced images on different layers, and play with different ways of merging them.