74 gun ship by Gaetan Bordeleau - 1:24

[Landlocked123]

Gaeten,

 

Your attention to the highly technical details of digital photography are quite evident in your results.

 

Best,

John

[Gaetan Bordeleau]

Than you,

 

Mathematical properties of light have always interested me.

Here is a little test I did this morning, just for fun:

I measure the color accuracy and the intensity of neons and lamps in the workshop.

 

My conclusion would be the following line:

In home, if I really want to capture the true colors of a model ship boat, a flash reaching at least 5000 ⁰K must be use.

That was the hard way, now the easy way, just do it outside.

 

-2 kinds of light source neons and lamps

-Lux values closely related to the distance : the closer to the source, the stronger the value

-Readings are stronger with lamps, distance is about the half of neon readings

-Tests to be exact would need readings at the same distance, this is not the case here.

- For a daylight  of about 5300 ⁰K, the closer I have is a Dazor lamp with neons

-To reach 5000 ⁰K, a standard daylight value,  a flash must be use

 

Possibly new LED lights have closer color to what we see during a normal sunny day.

 

 

[Gaetan Bordeleau]

Not really working on the mode ship actually, just driving  on the photography learning curve.

Just doing some experiment in tracing.

Simplified version  to report a height on  a stick.

Marking with a black rope, it does easily follows the curves.

I was too lazy to install a bigger vice for 1 hole only, I built a ‘’jaw extender ‘’.

 

Actually doing some experiments  with global markings using a projector.

If I could use the image and do all  the tracing, it surely would  be faster.

 

[SJSoane]

Hi Gaetan,

So are you projecting the sheer drawing onto the side of the hull, for marking off things like the wales?

Mark

[Gaetan Bordeleau]

Hi Mark,

 

Actually, I am trying to see if it is possible!

If possible using this method would means less errors in lines transfer like wales, deck lines, gun port openings.

 

[SJSoane]

Hi Gaetan,

 

I look forward to seeing the results of your experiments. This would indeed be a very useful tool.

 

Mark

[mtaylor]

Interesting method, Gaetan. I too am curious how this works out.  Fascinating idea... 

[cog]

14 hours ago, Gaetan Bordeleau said:

Hi Mark,

 

Actually, I am trying to see if it is possible!

If possible using this method would means less errors in lines transfer like wales, deck lines, gun port openings.

 

As Mark wrote: Interesting method.

 

If you project an image through a beamer on a straight surface, you will be able to trace the lines without a problem. However, if you project on a slanted or curved object, you wll get distortion either in the vertical, horizontal or both. When I look at your gun ports, those may not be subject to the distortion, or minimalistic considering their size, but the lines from stem to stern are definitely, so the place ment of the ports may/will probably be off, as will the run of the deck, etc where there is a curvature/slant of the hull

Edited February 22, 2018 by cog

[Gaetan Bordeleau]

Hi Carl,

 

Projection on a curved object : you will get distortion either in the vertical, OR, horizontal or both.  

 

Here is how I see it :

The main problem is at the bow.

If I look the curve of the bow from a top view, we see it in true lenght.

If I look  the curve at the bow from side, we see it shorter than what it is really.

 

For this exercise, what is important  is the view side : if the projection exactly fit on the model; lenghts are good and what is important here : the height of the curve is true and the goal here is to set points at the good height.

 

[Gaetan Bordeleau]

I traced fewer lines; outside shape and decks on a plastic sheet.

Tried a fast transfer.

 

Next step is to compare with transfer method of 1 coordinate at the time.

[cog]

Thanks for  the language lesson, the advantage of not being a native English speaker is that I can make mistakes, although, I try not to 

 

I agree with the bow  part being the hard one to solve. Is it possible to transfer a 2D rendered image to a 3D projection? 

Edited February 22, 2018 by cog

[Gaetan Bordeleau]

Carl, I surely will not give you an english lesson because it is not my primary language.

You ask a good question and I do not really know the answer.

 

At first sight, I would say that it is not possible.

In fact, I do not think that it is what I do.

Here is how I see it: 

It is like to transfer a 2D drawing on a 2D paper, the work is only in 2D. No measures are taken for the third dimension; depth.

 

 

I remade the lines with more precision. Still have to check the lines.

 

Height between decks is not the same for each one. In a house, usually all the height of ceilings are at the same height, but apparently not on the french 74.

 

I traced in red, on the drawing, parallel lines to the keel at the lowest point of decks around at  the middle. Overall, I would separate decks in 2 : fore and aft.

 

The fore part are parallel to the keel.

The aft part is much higher and in angle.

I think that the engineers at that time were bright. I am imagining the next sentences, so I do not know if it make sense or not :

Aft volume is smaller than the front half.

When in water, aft part is deeper than the front one. I always thaught that it was because it was heavier but I do not really know why. May be because of the lest, may be because a bigger volume will float more! I have never tried it but I guess that if I put 1 ton of lest fore and 1 ton of lest aft, aft part would be deeper because of the volume effect.

 

But anyway, aft part being shifted deeper and fore half being shifted higher, the height difference on decks become  more equal fore and aft. Where there was  almost no height difference at the fore parts we now have more.

Water from both ends now flows better  towards each end, and this is the camber. We can see it on the last drawing; before and after.

 

I also tried to see the optical effect of a curve line and a kinked one at the bow, but did not get good results.

That is enough dreaming for today.

[druxey]

Like Carl, I doubt if projection will give accurate results. Too bad; it would save a lot of time! Even projecting on to a flat surface will result is a degree of distortion due to lens. This will only get worse when projecting onto to a curved surface.

 

I've used threads for some years to line off planking runs. They are easy to view and to adjust, unlike wood battens.

[SJSoane]

Hi Gaetan,

 

I have been thinking about your projector experiment. I have attached a diagram showing what the issues might be.

 

In the first diagram, showing the project and ship hull from the side, if the light source is a single point, then the image lines will radiate away from the source. An image at X is projected to a greater Y on the projected surface, but the image will be an even greater Z at the hull.

 

The second diagram shows the setup in plan. If the light source creates parallel rays, the image on the projected surface will be accurate, even around the curve at the bow. But if the light source is a single point, as in the third diagram, the distance on the hull at Z will be greater than the distance at Y on the projected surface. and the curve at the bow will not be accurate.

 

You could avoid this problem if you could take each point on the projected surface, and then project it onto the hull exactly at right angles to the projected surface. Maybe this could be done with your laser beam.

 

Just some thoughts in your very interesting experiment!

 

Mark

 

 

 

Edited February 24, 2018 by SJSoane

[Gaetan Bordeleau]

Hi Marc and Druxey and Carl

 

I think you are all in the good direction!

[Gaetan Bordeleau]

Marc clearly explained why using a projector is useless.

So the conclusion of this project is very easy now:

[cog]

Gaetan,

 

I think you will be better at it than I (English), else you may polish my French.

 

I think it is feasible, but not with a projector. I have something brewing but I will probably need to make a drawing to get it explained.

 

The quick version: You would need a spot light to set the points on the hull. If you have a drawing of the scale model in 2D in front of the model, and you have pierced the line of the deck, and e.g. the corners of the gun ports, you could set the exact corresponding points on the hull when you place a spot light at 90º angle to the keel at the level center of the pierced spot. {edit] so you get a narrow beam from the pierced hole to the corresponding point on the hull.[end edit] Does that make any sense to you?

Edited February 24, 2018 by cog

[Sailor1234567890]

My understanding is that the distortion will be dependant on the scale and size. Small rowboats have been projection lofted but when you start dealing with multiple decks, complex shapes etc. it becomes too much and you'll never get it to properly transfer the correct shapes. Nobody has been able to successfully projection loft a large complex boat for the reasons you've discovered. I do recall an article in Woodenboat magazine a number of years ago about projection lofting though. 

[Gaetan Bordeleau]

Carl you understand many things.

 

Let"s see if I understand what you mean:

 

The problem with the projector is that there will always be a projection angle in the ray of light.

To correct this each point is set  by first setting the light in front of the point to set  and at 90 degrees of the keel and at the same height level.

The goal is to project a point without distorsion. The light must be set for each point...

 

drawing: light 2d drawing  beam light through the drawing to the 3d model.

 

[druxey]

That's about it, Gaetan. The ray from any point must be normal to its point of origin. And the projected target at right angles to the ray point.

[mtaylor]

For the point light, I'm thinking the laser would be perfect.

[cog]

Il ne faut pas exagérer, chèr Gaetan

 

That's exactly what I meant. So the projection will work, but only when you comply with certain requirements. 

 

Mark, that would be a nice precision tool for it

[druxey]

However, the point of origin of the laser will have to be precisely aligned with the point on the draught, as well as be horizontal and at right angles to the plane of the draught, as well as the draught precisely in line with the model, which will need to be precisely parallel to the draught.... Aaaagh!

[cog]

Exactly ... Druxey ... piece of cake for Gaetan

[Gaetan Bordeleau]

Almost Carl,  this method being the long version,  I will only use the short version:

 

Height transfer on a stick

The stick is set at 90 degrees in both axes

a horizontal laser line transfer the line on the model

 

But it would have been nice to do it a bit faster

[mtaylor]

Sounds like a good plan, Gaetan.  

[cog]

20 hours ago, Gaetan Bordeleau said:

But it would have been nice to do it a bit faster

 

hear, hear

 

fortunately, this work of art is not a mere model, but amongst others lessons in patience, and creativity ...

Edited February 26, 2018 by cog

[Gaetan Bordeleau]

Marking of decks and gun ports opening is on the way.

 

To do the marking of the top rails,  the stick which keeps all the frames close to each other, had to be lowered. The lower  the stick is, the greater the distance between frames has the tendancy to want to expand. This is why tie wrap were added temporarily and this is why we always set these sticks at the top.

[Gaetan Bordeleau]

For the Gun port openings, I would say, there are 2 ways to align each opening part. An upper part and a lower part are inserted horizontally. I did trace the parts inside and outside, as it can be seen on the last picture of the previous post. The alignment of the gunport can be done on one or the other part but I think greater precision can be achieved with the inside line because this is the part which will be seen, If we check the alignment , we will check it with the inner part not the outside one, planking being installed or not.

[BANYAN]

Very nice and accurate work (as usual) Gaetan; that hull is looking very good.

 

cheers

 

Pat