I recall seeing a website somewhere in cyberspace that described a sort
of digital multiple exposure technique that involved taking 3 shots of
the same scene (IIRC the example involved some moving water). One then
took the three images into Photoshop or whatever and made one of the
images red, one blue and one green, then overlayed the 3 into one image
that created a "false color" effect on any parts of the 3 images that
differed from the others.
I may have the details wrong, but that was the general idea. I lost
the link in a computer hiccup a while back. Anyone familiar with what I
am talking about? There was a technical name for the technique.
>I recall seeing a website somewhere in cyberspace that described a sort
>of digital multiple exposure technique that involved taking 3 shots of
>the same scene (IIRC the example involved some moving water). One then
>took the three images into Photoshop or whatever and made one of the
>images red, one blue and one green, then overlayed the 3 into one image
>that created a "false color" effect on any parts of the 3 images that
>differed from the others.
>
> I may have the details wrong, but that was the general idea. I lost
>the link in a computer hiccup a while back. Anyone familiar with what I
>am talking about? There was a technical name for the technique.
>
> Thanks for any info.
>
>-Al A.
This method was originally invented in the film era. You would take 3 exposures
with b/w film, using a red, green, and blue filter over the lens for each
exposure. When combined on color print paper only the objects that moved would
have color hues on them. Anything that was motionless would be a full-color
image. Later when color film became more popular, the technique was revised to
use with any color negative or slide film. A rectangular card was made with the
3 filters placed sequentially on your filter-slide. The card was physically
dropped inside of a long filter-holder in front of the lens during a 1/2 second
or so exposure. The card itself becoming a 3-color camera shutter.
Start out by using a burst mode with your camera on a steady tripod to make 3
exposures of the same scene just a few fractions of a second apart from one
another. Or be fast with the shutter button and take 3 photos in rapid
succession. Make sure you don't move the camera during the exposures. The images
must be perfectly aligned. Just things that move in the image should change
position.
Use any editor that allows you to split the RBG (or CYMK) layers of your 3
photos. When done you should have 3 (or 4 with CYMK) B&W images for each one of
your 3 photos.
Select the blue-channel layer of your first image, copy it to a new image.
Select the red-channel layer of your second image, copy it to your new image.
Select the green-channel layer of your third image, copy it to your new image.
(or if using CYMK layers, do the same with each color channel, don't forget to
add back in your K layer from one of the photos)
Then use the editor's merge (or combine) RGB layers or merge (or combine) CYMK
layers on your new image. Be sure to select the right layers for the right
colors. It might help to label them (in your layer palette) as you copy them
since they will just appear as b&w images until recombined.
Or if you can just copy each one of the color channels from each of the 3 photos
(without doing a separation) to a clipboard, and paste them into a new image,
that's even easier.
Most any decent editor has these features.
It's a fun technique to learn what all the color channels and layers are all
about, but the novelty of images produced this way wears off after you do and
see just one. At one time I thought it might add an extra touch to some
mountain-lake photos, but it only made them look tacky and destroyed how nice
they were in the first place. It reminded me of an elvis on velvet painting that
you'd buy at a garage sale for 50 cents. A fun exercise but worthless for any
useful visual merit.
BaumBadier wrote:
> On Mon, 09 Jul 2007 22:03:35 -0400, "Al A." <alanganes@comcast.net> wrote:
>
>> I recall seeing a website somewhere in cyberspace that described a sort
>> of digital multiple exposure technique that involved taking 3 shots of
>> the same scene (IIRC the example involved some moving water). One then
>> took the three images into Photoshop or whatever and made one of the
>> images red, one blue and one green, then overlayed the 3 into one image
>> that created a "false color" effect on any parts of the 3 images that
>> differed from the others.
>>
>> I may have the details wrong, but that was the general idea. I lost
>> the link in a computer hiccup a while back. Anyone familiar with what I
>> am talking about? There was a technical name for the technique.
>>
>> Thanks for any info.
>>
>> -Al A.
>
> This method was originally invented in the film era. You would take 3 exposures
> with b/w film, using a red, green, and blue filter over the lens for each
> exposure. When combined on color print paper only the objects that moved would
> have color hues on them. Anything that was motionless would be a full-color
> image. Later when color film became more popular, the technique was revised to
> use with any color negative or slide film. A rectangular card was made with the
> 3 filters placed sequentially on your filter-slide. The card was physically
> dropped inside of a long filter-holder in front of the lens during a 1/2 second
> or so exposure. The card itself becoming a 3-color camera shutter.
>
> Start out by using a burst mode with your camera on a steady tripod to make 3
> exposures of the same scene just a few fractions of a second apart from one
> another. Or be fast with the shutter button and take 3 photos in rapid
> succession. Make sure you don't move the camera during the exposures. The images
> must be perfectly aligned. Just things that move in the image should change
> position.
>
> Use any editor that allows you to split the RBG (or CYMK) layers of your 3
> photos. When done you should have 3 (or 4 with CYMK) B&W images for each one of
> your 3 photos.
>
> Select the blue-channel layer of your first image, copy it to a new image.
> Select the red-channel layer of your second image, copy it to your new image.
> Select the green-channel layer of your third image, copy it to your new image.
> (or if using CYMK layers, do the same with each color channel, don't forget to
> add back in your K layer from one of the photos)
>
> Then use the editor's merge (or combine) RGB layers or merge (or combine) CYMK
> layers on your new image. Be sure to select the right layers for the right
> colors. It might help to label them (in your layer palette) as you copy them
> since they will just appear as b&w images until recombined.
>
> Or if you can just copy each one of the color channels from each of the 3 photos
> (without doing a separation) to a clipboard, and paste them into a new image,
> that's even easier.
>
> Most any decent editor has these features.
>
> It's a fun technique to learn what all the color channels and layers are all
> about, but the novelty of images produced this way wears off after you do and
> see just one. At one time I thought it might add an extra touch to some
> mountain-lake photos, but it only made them look tacky and destroyed how nice
> they were in the first place. It reminded me of an elvis on velvet painting that
> you'd buy at a garage sale for 50 cents. A fun exercise but worthless for any
> useful visual merit.
BaumBadier wrote:
SNIP
>
> This method was originally invented in the film era. You would take 3 exposures
> with b/w film, using a red, green, and blue filter over the lens for each
> exposure. When combined on color print paper only the objects that moved would
> have color hues on them. Anything that was motionless would be a full-color
>
<SNIP a bunch of great info>
Thanks very much for the detailed reply. It is exactly what I was
looking for.
I actually am thinking that it may actually be useful to me in a work
project that I am on. So while the general interest and fun of it
appeals to me, I may end up not using it for its aesthetic value. I have
an application where we need to determine if something is moving (very
slightly) and what is not. Hard for me to give particulars, without
saying exactly what I'm working on, which I'm not supposed to do. There
may be better ways to accomplish it in this instance, I don't know, but
this is available, and we have the gear necessary (camera and photo
editor) to try it, for only the time invested. A few hours of
experimenting should give some indication if we have any chance of using
this effectively.
Thanks again!!
Al A.
PS- And thanks for the links, Corry. That first one was one of the
ones I was looking for. Now why didn't I remember the term
"Multitemporal Technique"? Silly me...
> I actually am thinking that it may actually be useful to me in a work
>project that I am on. So while the general interest and fun of it
>appeals to me, I may end up not using it for its aesthetic value. I have
>an application where we need to determine if something is moving (very
>slightly) and what is not. Hard for me to give particulars, without
>saying exactly what I'm working on, which I'm not supposed to do. There
>may be better ways to accomplish it in this instance, I don't know, but
>this is available, and we have the gear necessary (camera and photo
>editor) to try it, for only the time invested. A few hours of
>experimenting should give some indication if we have any chance of using
>this effectively.
>
>Thanks again!!
For motion studies it could be fairly beneficial. I hope it works out for you.
You might like something a little easier to discern though, if detecting
movement is more important than "pretty colors". No need to tediously futz
around with all those color channels too. Just take 2 consecutive frames and
copy one as a layer on top of another (in any editor that supports layers of
course, most all of them do layers these days). In that new top-layer's
properties choose to make it as a "Difference" layer. Anything that moves in the
image will show up in stark contrast against the black that is caused by
everything that is NOT different between the two. Black = zero difference. Not
as colorful, but much easier to see the smallest of changes from one frame to
next. In the old days I used to use this when making web animations from
hand-held sequences and trying to realign one frame to another manually. Before
there became easier automated methods. Shifting, resizing, and skewing the layer
on top to create as much black as possible and only leaving the part that I
wanted as animated showing. Then, after shifting that new layer as needed to
align it as perfectly as possible, reverting that layer back to "normal" from
"difference". Stacking as many frames as needed, repeating this process each
time, then cropping the whole lot of them as one, to the largest
common-denominator.
If you need to record only the difference between the two, you could always
merge the background and "difference" layer, which is now mostly black, and use
it as a mask layer. Then apply that to your original image. Only those parts
that are different will show through the mask.There's all kinds of ways to go
about this.
BaumBadier wrote:
> On Wed, 11 Jul 2007 19:13:06 -0400, "Al A." <alanganes@comcast.net> wrote:
>
>> I actually am thinking that it may actually be useful to me in a work
>> project that I am on. So while the general interest and fun of it
>> appeals to me, I may end up not using it for its aesthetic value. I have
>> an application where we need to determine if something is moving (very
>> slightly) and what is not. Hard for me to give particulars, without
>> saying exactly what I'm working on, which I'm not supposed to do. There
>> may be better ways to accomplish it in this instance, I don't know, but
>> this is available, and we have the gear necessary (camera and photo
>> editor) to try it, for only the time invested. A few hours of
>> experimenting should give some indication if we have any chance of using
>> this effectively.
>>
>> Thanks again!!
>
> For motion studies it could be fairly beneficial. I hope it works out for you.
>
> You might like something a little easier to discern though, if detecting
> movement is more important than "pretty colors". No need to tediously futz
> around with all those color channels too. Just take 2 consecutive frames and
> copy one as a layer on top of another (in any editor that supports layers of
> course, most all of them do layers these days). In that new top-layer's
> properties choose to make it as a "Difference" layer. Anything that moves in the
> image will show up in stark contrast against the black that is caused by
> everything that is NOT different between the two. Black = zero difference. Not
> as colorful, but much easier to see the smallest of changes from one frame to
> next. In the old days I used to use this when making web animations from
> hand-held sequences and trying to realign one frame to another manually. Before
> there became easier automated methods. Shifting, resizing, and skewing the layer
> on top to create as much black as possible and only leaving the part that I
> wanted as animated showing. Then, after shifting that new layer as needed to
> align it as perfectly as possible, reverting that layer back to "normal" from
> "difference". Stacking as many frames as needed, repeating this process each
> time, then cropping the whole lot of them as one, to the largest
> common-denominator.
>
> If you need to record only the difference between the two, you could always
> merge the background and "difference" layer, which is now mostly black, and use
> it as a mask layer. Then apply that to your original image. Only those parts
> that are different will show through the mask.There's all kinds of ways to go
> about this.
>
>
>
Again, Thank you very much for taking the time to pass along such
through info. It is mostly a matter of finding "What parts are moving,
and which are not?" that matters, at least for now. The colors are not
important, really. I expect that if we get some useful results, it will
lead to all sorts of new inquiries to try out. All your info is most
helpful. I will likely give some of this a try sometime next week.
false color technique? 
Linear Mode
