[tech-3d] Re: Stereo Base Calculation With a $20.00 Handheld

["Michael K. Davis" <zilch0@xxxxxxxxxxxx>]

Hi Boris,

In scouring the net for 3D galleries I came across your work for the first
time some months ago.  There's no question you are producing some beautiful
stereography.  I especially like "The Grinder", zoomed in, as shown here:

http://www.starosta.com/3dshowcase/3dimages/grinderPcu.jpg

Nice web page design too!


>From:   <boris@xxxx>
>Date: Sat Nov 18, 2000 12:28pm
>Subject: Re: Re: Stereo Base Calculation With a $20.00 Handheld
>
>Mike:
>
>you wrote in response to Allan Griffin:
>
>>Your assumption suggests you are convinced that a fixed separation camera
>>satisfies your requirements for "mid-field" subjects.  I personally don't
>>find this approach satisfying, but let's agree to respect each other's
>
>I also have started using variable bases for the mid range, in particular
>to make my anaglyphs work better.  I used to shoot just close up ortho
>(using the 1/10 rule! not that it ever existed), but as I have broadened my
>subject matter, so have my methods and my stereobase.  So far I have used
>just my intuition, but I have longed for a math/calculator tool for some
>time.

The HP 30S is the least expensive calculator I could find that fits the
bill.  Within fifteen seconds of turning it on, it will have prompted you
for the variables and given you the base, in mm.  All for $20.00.   I sure
do like it.

I was working with a really large spreadsheet (printed and laminated), but
it only handled various Nears, with Infinity always assumed as the Far.  (I
shoot mostly landscapes.)  Now I can accurately calcuate base for any Near
or Far, instead of interpolating from my spreadsheet.

[snip]

>I am not going to try to critique it technically/mathematically (It's been
>too long since HS algebra/geometry), but I look forward to programming my
>HP-15c one of these days and putting the equation to the test.  Given its
>flexibility (i.e. % of MAOFD variable input) and assuming it works I look
>forward to its utility.  Can I put in MORE than 100% MAOFD?  

Absolutely.  Just supply the percentage as an integer, for example: 80
instead of 0.80, or 110 instead of 1.10, etc.  The formula will accept any
value you want to supply.

>I am an
>incorrigible depth demon.  Can Mike help me re-write the program to take
>on-film-deviation (in mm) directly as an input, with base as the output?

No problem - here's the formula again, as entered in the HP 30S:   

    .026C*((304.8AB/(B-A))*(1/65-1/(609.6AB/(A+B))))

Deviation is the first argument, so you would just have to replace this:

    .026C

With this:

    (Viewer FL/Camera FL) * C

Where C is your desired deviation in millimeters.  Rename the variable to
d, if you like.  (C was previously the percentage of MAOFD desired.)

The constant .026 was derrived from the following, as explained in the the
original article:

    (Camera FL/30)*(Viewer FL/Camera FL)*(1/100)

    i.e. (65/30) * (78/65) * (1/100) = 0.026

For your purposes, where you want to specify the deviation directly, you
can drop the (Camera FL/30), which is MAOFD, and the (1/100) multiplier,
but you really should retain the (Viwer FL/Camera FL) modifier to
compensate for the stretch or squash caused by a mismatch in focal lengths.

If your Viewer or Camera Focal Lengths are not constant, you'll want to
make variables out of these.  If they never change, then just code them as
a constant, for example:   78/65 * C = 1.2C, for a 78mm viewer with a 65mm
taking lens.

The new equation would look like this:

    1.2C*((304.8AB/(B-A))*(1/65-1/(609.6AB/(A+B))))

Again, using this equation, C is the actual desired on-film deviation and
the constant 1.2 compensates for a 78mm viewer focal length being longer
than the 65mm taking lenses.  If you make the camera FL a variable,
remember to substitute the other constant of value 65 (in the middle of the
formula) with that variable.  

Enjoy!

Mike



-------------------------- eGroups Sponsor -------------------------~-~>
eGroups eLerts
It's Easy. It's Fun. Best of All, it's Free!
http://click.egroups.com/1/9698/3/_/520353/_/974690665/
---------------------------------------------------------------------_->