29 minutes ago, Chris Ross said:

You could take your frosted glass and place it in a cardboard mask and have a light source behind it then take the fresnel and move it back and forth to see what size spot it forms. to get the light into the tube you want a spot the same diameter as the tube. This you c coul use to determine how far to place the fresnel from the tube.

That is more or less what I made in the past. with 3d Printed piece to try different position
IN AIR....that is the problem in water the light/lenses work not the same.

31 minutes ago, Chris Ross said:

If the light goes into the tube the cone is still going to try to expand so maybe paintin the inside white so it bounces around in there might help?

32 minutes ago, Chris Ross said:

One example of what you are trying to do is the "better beamer" strobe attachment for bird photos it narrows the strobe beam from something covering the field of a 100-200mm lens to something covering a 300-400mm lens it quoted as increasing strobe light by 2 stops or so. See this link:

I bought some mirror paint bomb the idea is to put

1) before the frostet glas (or a very lite frostet glas) a beam concentrator like that

Painting the inner part of the snoot before the frostest glass like a "mirror" will help.
The inner form need to be calculate to concentrate the light on the frosted glas.

The first part will help to minimise the lost of light on the source.

Part 2 after In the "tube" It can be painted like mirror... but also there it will help to calculate the form of the tube/cone what ever you want to name this to reach a maximum of the ray to be // at the end of the "tube".

If it is not good enough used some "lenses" as collimator to reach a good directed flow of light in front of the snoot.

1 hour ago, Chris Ross said:

You could look into raytracing software but I think it's a fairly steep learning curve setting up software to model what you are trying to correctly.

I need to make test with the 3D modélisation software I use to look if the raytracing possibility are good enough to achieve the target of a good calculation of the result.

With to software I used before (at work)I believe it was possible.

I'll make test with this

find on Amazon... the bomb are wiating to be used since 4 months :(

Edited 8 hours ago8 hr by CaolIla