But how do you put an photographic image of the tomato's skin onto the globular/spherical shape of an un-cut tomato model - come to that, what's the best way - angle etc, to take an photo of a tomato so as to have an image to put on a spherical (just about) model of a tomato???

Okay, best quess here...others with more experience may correct me at will.

First I would create the perfect 'red' skin coloration/material, and apply it to the entire object. 
Then create a 'Decal" from a photo of just the stem area, and apply it as a decal with a top-down orientation.

Any other input, people?  Does this solution sound reasonable?

UV-mapping would be my guess.......?
or is it......create a sphere, apply the pic and stretch it using the Texture Edit controls?

mmmm....just tried that.....all I got was the pic on the sphere but it had a white part on the back of the sphere, which you could hide.....maybe....

(It's the actual procedure of making it that I don't get.)

Like: "First I would create the perfect 'red' skin coloration/material, and apply it to the entire object." 
how do you create such a thing?  Do you mean use a procedural?

Cos that's why I wanted to use a photo - it looks more like the real thing than a procedural does.

Not sure how to use (or make) a 'decal', but I've used the "object top" orientation to put the image of the tomato slice on to the cut surface.

It really depends on what you are doing. You could create a projection from a photo of a round tomato and put it on a sphere. You need a photo with even light so there are no big specular white spots on the tomato. I'd take about three photos of the tomato to fill the whole screen top to bottom then crop them square so the tomato touches on each edge. This should be a large image 2000 pixels or greater so you can doctor and reduce the final texture later.

There are several paint programs with distortion filters. I remember the flaming pear plugins called it christmas ball. Paintshop has a menu in effects called distortion effects. In this is a filter called Polar Coordinates. This will let you  covert from polar to rectangular. Convert all three photos. Create a blank image with an aspect of 1 high by 2 wide. Crop the center of the converted photos to eliminate the reminants of the non tomato edges that still may be hanging around in the image. Piece these three sections together on the blank image to create a spherical map. Save often so if you get completely messed up you can go back to an earlier version. Clone or blend the edges so the transitions don't stick out. Make seamless tiling can help but only on the outer edges. You can also use aa thing called the highpass filter to take bright regions out of the texture while still preserving fine details.

Help?

Aspect? 

1 what?

2 what?

Oh, and you can see most of the polar converted image there - does that look right to you?

I see some funny little pale bits at the bottom of the image - probably the remnants of the edges of the round tomatoe - that were showing background, rather than tomato - should I just crop that edge off?  Is that what you mean?

Or do I need to crop higher than that to eliminate the shadows too - the ones that were at the bottom of the tomato in the original?

(I've taken shots of the stalk terminal, so I have that part covered... literally.) Lol!

But the skin is just a procedural, not a photo.

Fran you might try kicking up the specularity a little to make a shiney area like on the photo of the tomato skin

2 width to 1 height somewhere above 600 pixels width to 300 pixels height . There's a place to adjust these. anything that even sounds like maintain aspect ratio should be switched off. and 2:1 is what I ended up using all those yearons ago when I was experimenting with mapping to a sphere. The thing you'll realize and rather quickly id imagine is that as you approach the poles you have to exagurate the horizontal pixels more and more as you're putting the same amount of data into an ever smaller space. What do they call it again? mercanter projection? something like that.

BecSchm,
Specularity is already at 100% - and turning it down doesn't seem to have any effect either...

Um... played with it a bit and it seems taking the bead out of the procedural on Specular, helps.

I still want to try the photo though.

**bikermouse,
**
Ah, er, um... so the actual size of this new image would be whatever the combined width of the 3 original images makes up to?  yes? so long as the aspect ratio remains 2:1?

Gonna make it darned difficult to squeeze 3 images into the space for just 2... (shakes head) I'm having a hard time understanding this bit.

I mean, why do you need to use 3 images?  Why not use 2?  then it'd be a lot easier to get 2 images wide and 1 image high, in a new 2:1 image.

So how do you squeeze 3 into 2?  just change the aspect ratio of the 3 images until they all fit into the new one?

:scared:

The polar to rectangular transformation warps the round photo into a square map (mostly). Tomatos aren't perfectly round. The bumps are what create those white spots on the bottom of the image. You will have to crop those off. You make the image big because the pixels at the edges get stretched more than the center. You want enough pixels so you don't wind up with huge squares on the edges.

A rectangular map is just like a world map. It takes a certain distance to go from pole to pole. It takes an equal distance to go from the edge you can see on the left to the edge you can see on the right. It takes that distance again to go around the back side of the sphere from the right edge back over to the left edge you can see again. The map image needed to wrap around a sphere is twice as wide east west as it is tall north south.

Say your first tomato images are 2000 pixels square. Go into the File menu and create a new image 3600 pixels wide by 1800 pixels tall. Go to your tomato image 1 , copy and paste as a new selection into the left side of the created image. because tomato 1 is bigger than the created image you can shift the bottom image around to remove the white blobs on the bottom of the tomato texture. You can paste image 2 next to this or on the right side of the new image. you can also select the 3rd image using the box select tool with feather set at 20 pixels. This will make the transparency of the edges blend from 100 to 0 in 20 pixels. When you paste this over the other image it will produce a blend instead of a sharp line between images. The idea is to shift things around until you get a continuous rectangle of skin. I say three images because you want to use the best parts and two might not be enough. It's okay if bad parts or white blobs on the edges go to waste.

Your original photo is too hot. All you are after is the colors for diffuse and ambient channels. If you include the specular white spots you will have problems lining up the texture with the lights you create in a scene. You want to use the specular setting in bryce to create your hotspots. 

You need a tripod or just put both on a table so they don't move. Don't use a flash or have any direct lights or lighbulbs shining on the surface. You can get good diffuse light by hanging up a white sheet around the object.

Fran,
If you manage to get what Electroglyph is saying done for a texture; I made a sphere .obj a while back that is UV mapped to lat./long. coordinates. A 2:1 lat/long panorama will fit to it seamlessly.
You could just squish the sphere and jiggle :) some verts around to make your tomato shape. I will mail you the .obj if you want it. It is hi-res, 10,000 polys (100 segments, 100 slices). OK, it's not really a sphere; but it looks just like a sphere, and behaves mostly like a sphere (the latitudinal edge loops are broken, i.e. there is a non dimensional hole, resulting in 100 identical verts, at each pole)...
Some of the polar polys could be deleted (and put hole material, in Wings) and then you could model some leaves and a stem in their place, which could use any kind of texture independent of the skin. Or you could just put leaf/stem model on top of it. Just don't tamper with the sphere's UV Mapping, or extrude any faces from it directly.

Not sure I understand:

1. if you're going** take three pictures ** the way BecSchm's picture suggests, you will really have to watch your lighting** **so that they are equally well lit **(i.e. no shadows no bright spots.)
   **

2. as has been suggested if you do it that way you will have strech the image to eliminate the 'white spots'.

3. you should end up with only one image. 

NOTE *****************************************************************************************
the 2 to 1 deal isn't exact for anything but a sphere but it should be ok. 

knowing that 2 PI radians make a circle you can figure that what you see at any time would be PI radians (or a little less) you can figure the height and width to be equal at this point in a sphere so the vertical would also be pi radians; therefore the ratio is exactly 2:1. 

in an oblate sphereoid the height is less than pi radians  so the ratio would be slightly larger. I don't think you need to worry about whatever little differences are there for your current purposes. 

also if i were to use a texture based on a photo I wouldn't use 3 photos just one and crop it to the point where what you were going to use was all tomato(no 'white areas') you will notice a line along where the horizontal edges are projected to your sphere - you'll need to blend these in photoshop.    

Bikermouse  sorry if I am causing any confusion.  What I meant to suggest was that once Fran gets the photo image applied as electroglyph describes, she can then put the shiney spot back on to the tomato by using the specular setting as on the second tomato in my image.  

Ah!

I was just wondering something, since I'm not sure how to do this following (or not following would be more accurate) the instructions...

I found that creating polar thingie using the photo of the tomato - a new one taken today in daylight - after cropping the image so there are no edges to the tomato and using as square a cropping border as I could get - I then got a polar image with no funny whitish bumps along one edge.

My question is, do I need the edges of the tomato?  Or will just part of the tomato skin image do?
I'm in the meantime now taking more pics:

1. hanging a sheet over the window
   (rather dark cos now it's raining)
2. With room light out
   (what tomato?)
   Having immense trouble focusing, so I'm also trying manual focus, which is hard cos my eyes aren't too good, especially when craning my neck over and down cos the height of the desk & chair ratio is naff - (even with using books to raise the tomato on a pedestal)
3. With room light on
   (oh that tomato)
4. etc etc...

What do you think?

......S'okay. I've tried again and as the rain stopped even though it's getting late - I think I've managed to get a good variety of better ones.  Now to go back to the instructions.

Well I (now) understand why the 2:1 ratio.

Actual size of 1st image: , the rest of the photos range from:

1354 to 1522 wide. And from

1156  to 1304 high

So...  with 1465 x 1140 and 1513 x 1149 and 1417 x 1228  as the sizes of my 3 photos, cropped, turned polar and cropped again – what size do I make the 2:1 ratio big new image?

Even if you get a good texture and apply it to a smooth sphere it's going to look wrong. Part of what makes us see a tomato as a tomato is that it's not perfect. there's always little feet, or bumps that stick out or flat sections where it was laying.

I made a couple of different whole tomatos and one cut into seven sections. The sliced one is spread out because the skin and each face are seperate objects. There's no good way to make this a bryce file without making it too big for me to upload so I left a way for you to grab the pieces in Bryce. The zip contains obj files with materials and names applied. By materials I mean that the skin,  faces, stem, leaves all have names and you can slap your images on them. I used bryce procedurals for the actual image.

http://geocities.com/electroglyph2002/freestuff/FransTomatos.zip

And yes I did put some small imperfections into one of the whole tomato models, and it did look a lot better that way.

I was thinking of giving mine away as a freebie when I'd finished playing with them. If I ever get a skin done for them.

Any chance you can explain what size image I need - re my message above, with 1465 x 1140 and 1513 x 1149 and 1417 x 1228 as the sizes of my 3 photos?

I'm afraid I just have no idea how to figure out what size new image I need.

Above is the image of the slightly dented tomato in Bryce - and I know the bump is too high on the terminal, and it's too shiny. But you can see the effect of the denting, so I'm off to make more dents.

Ah i see you noticed a tomato is reflective :-)
Now to get the right type of specular effect, you will need to blur the reflections, and make them a little less strong.
I would guess to set the specular halo to (off the top of my head) the following HSL settings:
H 5
S 80
L 220
If this messes up the color of your specularity, you could probably lower the S value.
After that its a matter of reducing reflectivity untill the effect is right.
If the highlights are too sharp, increase the L value of the specular halo, if they're too blurry, lower them. I dont expect the correct value to be much lower then 210 though.
All this is ofcourse assuming you're using premium render mode.

All your photos should be the same distance from the tomato. All your photos should have an aspect of 4 wide by 3 tall. 

When you take a photo you will wind up with something like the picture above. Crop the tomato out of the center before doing the polar coordinates transform (dashed white lines). The polar transform stretches the edges of the circle (yellow lines) into a square (blue dashed lines). 

In the photo the portion of the tomato directly in line with the camera at the center is straight. The edges (yellow lines) are curved by the shape of the tomato. The polar transform unbends this image and flattens it out so you can put it on a piece of paper like a world map. The center is fairly undistorted but it gets worse the farther you go toward the edges. That's why I say use three images because you can loose some of the edges.

When you have three images cropped into rough squares and transformed make your final image about 10% shorter that the shortest picture by about twice this much in width. Your shortest image above was 1140 so make the image you paste into 0.9x1140=1026 tall by 2x1026=2052 wide.

Make it 1024 x 2048. There are computational advantages to using 2^n power.

Which meant that I had a fair bit to crop out before starting on the polar bit.  Thanks for the calculations, I'll save this whole thread I think, lots in it.
You're very good at explaining this, and patient too.

Rayraz,

"Ah i see you noticed a tomato is reflective :-)"

Oh you wait man!  I'm on 37% antialiasing with the procedural.... Looks pretty good but I'm gonna try another with the settings you suggest before doing anything else.

DvLenk6,
Aaah, yeah, see what you mean.

I hope i've been accurate, i didnt have a chance to test it heheh was just off the top of my head :-P

Quote - Ah i see you noticed a tomato is reflective :-)
Now to get the right type of specular effect, you will need to blur the reflections, and make them a little less strong.
I would guess to set the specular halo to (off the top of my head) the following HSL settings:
H 5
S 80
L 220
If this messes up the color of your specularity, you could probably lower the S value.
After that its a matter of reducing reflectivity untill the effect is right.
If the highlights are too sharp, increase the L value of the specular halo, if they're too blurry, lower them. I dont expect the correct value to be much lower then 210 though.
All this is ofcourse assuming you're using premium render mode.

I just went to have a go with those settings... only to realise I haven't a clue where to find HSL in order to adjust it.
I went into the mat lab, but (shrug)

Help?

ah ofcourse, its an easter egg, forgot its not an obvous tool :-)
You get an alternative color picker if you ctrl+alt+shift+click on a color. The default windows one to be precise. It gives a pallete, a color chart, and hsl and rgb numeric input. I used the hsl numeric input.

Theres also a few other color pickers with diff ctrl/alt/shift/click combo's but i dont know them off the top of my head.

Quote - ah ofcourse, its an easter egg, forgot its not an obvous tool :-)
You get an alternative color picker if you ctrl+alt+shift+click on a color. ....

You know Rayraz, I have tried the Bryce Help from the 6.0 manual and it really stinks! There is no mention of even basic things like how to get to the alternate color picker, how to make images of the DTE defaults pop up, how to use the terrain editor, etc.

Maybe we could talk Renderosity into dedicating a part of the backroom to some tutorials that wouldn't get lost in time like regular posts do?

I tried "ctrl+alt+shift+click" and all that happened was another texture popped up in the B channel.

yea we could do that :-)
I think theres a site that listed all these easter eggs... dunno the link, but they also had an explaination of TIR and TA and how to put a face on the moon etc.

hmm while clickin on the color-dot?? weird... lemme try 1 sec

hm ok i tried, for me it works? it says at the top in the mat editor:
Specular Halo, then a little ellips, with a color, and then the A B C D channels.
its the little ellips u gotta ctrl+alt+shift-click at.

Rayraz,
Well weird or what, I tried it again and - this time it worked.  Maybe I was clicking on the specular instead of the specular halo? (and there's no colour in specular - that's all coming from the IBL.) I could have been - since my brain is rolluxed up from fighting to get Nero to save my work to a CD for a hand-in deadline.

**electroglyph,

**There's some stuff on easter eggs - and a whole lot more in the Bryce Bible - i.e. "Real World Bryce 4" by Susan Kitchens and Victor Gavenda - even though there's nothing in there on booleans, or IBL or collapsing and exporting meshes, or multi-replicate - as those only came along in B6, it's still an invaluable tool.

In fact Rayraz's tip for specular halo is probably there too.  I just didn't think to look.

Attached Link: Bryce Keyboard Shortcuts

Did you have a test render yet fran?

Oh no, turns out I did wait for it to finish first, and here's the result:
Top image, is before changing the HSL, bottom image is with H6  S80 L230 - turns out L won't go any higher than 240.

Still not right, is it?

Did you turn on the blurry reflections in premium render settings? (and possibly turn down the amount of reflection some) because it looks like the reflections didnt get blurred at all :-/

And indeed it doesnt go further then 240. Actually, 240 will give a few nasty sharp-edged artefacts somtimes. I prefer to not go higher then 239 instead.

Yeah, that's what I thought, but - silly me - I thought that HSL WAS turning down blurry reflections.  As I've not used that before, I didn't realise it was in the render settings...

Now you mention it, of course I do remember seeing in there.

Sorry, I'll go do that now.

Oh, I already turned the reflection down some on the lower of the 2 images in the message above.

....oh dear... oh deary deary me.  I have clicked on blurry reflections and Bryce is obviously thinking very hard about whether it will do this or not, since it is on the first pass, and it wasn't too happy about rendering just the grey background, once it got to the tomato.... sob!

hmm.. strangeness.. maybe hdri and blurry reflections dont like each other too much :-/
Whats the rpp setting you have? you can do testrenders at 4rpp to see if its anywhere close to looking as it should :-)

HSL is an alternative method of describing a color.
Other examples of methods are RGB (red, green blue) or CMYK (cyan, magenta, yellow, black).
HSL works with Hue (colortone), Saturation (color intensity) and Luminance (brightness). It's a much more human-friendly way of describing color compared to RGB.
I make sense of Hue values with the following rules of thumb:
0     = red
20   = orange
40   = yellow
80   = green
120 = aqua
160 = blue
200 = purple
239 = red

Yeah I know about CYMK and RGB but what's rpp?

rpp = rays per pixel, its in the rendering options of premium render. lower rpp makes for more noise, but less rendertime.

Quote - rpp = rays per pixel, its in the rendering options of premium render. lower rpp makes for more noise, but less rendertime.

Got to remember that... Thanks for the tip.

TS

RPP increases render times a lot more when you have transparencies and reflections.
Each time a camera 'ray' strikes a trans. or ref. surface, the ray is split into 'rpp' number of rays (it is a square grid, n^2 size); and each child ray goes about it's merry old way.
The child rays report back to the parent ' ray node' and an average color value is calculated. If the parent is itself a child, it then reports back to it's parent, and so on. More sampling rays also equates to more accurate Anti-Aliasing.
Bryce, by default, uses a fixed determinacy algorithm to determine what direction the child rays diverge from the surface polygon normals when they split off. When you activate the premium options (except True Ambience. it uses rpp distribution too; but is another animal altogether), it introduces quasi-random functions into the determinacy algorithm...'Jitter' of different types.

"Maximum Ray Depth" setting can have a HUGE impact on render times, especially when ray distribution is being used (rpp > 1).
The render setting "Maximum Ray Depth" is the number of times that an initial ray will split. Each surface that the ray encounters with either trans. or ref. is a split. So, if you have a sphere w/ glass material, you need at least 2 ray depth (one ray split, or 'penetration' for each side of the surface) for light to pass through the sphere. Just one depth for reflections; but another depth is needed for the transparency.
For best effect, you actually need 3 depth (number of transparency layers +1). +2 depth is slightly better quality (a few rays will bounce back and not penetrate the surfaces). Anything beyond that is wasted; increased render time with extremely minute increase in quality.
Less depth than number of transparency layers will result in surfaces not being sampled at all (pure black).
Just count up how many layers you need to pass through, add 1 (or 2), and use that number for the Maximum Ray Depth setting.

TIR (Total Internal Reflection) should theoretically limit the number of times that a ray is allowed to bounce (and split) inside of an object; thereby decreasing (or increasing) render times; but I have never noticed any difference when using it or not. Not in render time or quality.

er..  thats actually not entirely correct on the rpp thing there...
The RPP relates to the amount of points sampled within a pixel. So 4rpp will trace 4 points for each pixel, 64 will trace 64 points, etc. These points are then interpolated to determine the value of the resulting pixel. 64rpp does not mean that a single ray splits into 64 rays once it hits a reflective or transparent surface. Each point originally starts out with one single ray, which is why the value is called "rays per pixel".

In general, a reflective or transparent surface is slower because it requires more calculations to be done per ray as it bounces off, or goes through things instead of calculations comming to a hold upon 'hitting' a surface.
In the case of transparent surfaces, often a ray is split into 2 rays, one that gets reflected, and one that passes through.
Traditionally, for blurry effects these rays are substituted with a scattering of  multiple rays, though i dont know bryce's exact implementation. However I do have a theory;
Considering the high amounts of noise at low rpp for these effects, i assume no multiple rays are scattered from a single reflection or transmission, instead, the reflecting rays trajectory is most probably altered in a quasi-random fasion (at this point i do agree).

Maximum ray depth is of influence for scenes that use reflections or transparencies, but also for True Ambience. The higher the ray depth, the more bounces of indirect light are calculated. This is clearly visible in renders. Try rendering a TA scene with ray depth 1 and with ray depth 2, you'll already see ray depth 2 gives a brighter image. anything above ray depth 3 is probably overkill as far as TA quality goes, though higher values can be prefered for scenes with detailed reflections.

TIR does affect render time and quality! In certain cases when rendering glass objects, TIR gives you visual details that would otherwise not be rendered, usually it also seems to brighten up glassy materials because of the added internal reflections. These reflections do indeed increase rendertime, and in specific cases the difference can be very significant. Especially for additive materials.

Well, rays originate from the camera. A ray hits a pixel on a surface and splits into however many rpp you have set; and go off according to whatever direction, based on the normal of the surface and other formulas involved. It's reverse raytracing;  because it acts opposite of the way a real world camera does.
You're right about the secondary rays not splitting. I thought they did in Bryce, but they don't. My bad.

As I think about the quasi-random aspect...
The randomness can only be caused by grid distribution of rays. There is no other mechanism to account for it. It is, after all, using Cornell's Hybrid distributive raytracer...
I think that without activating premium features, that there is no randomness to the sampling. Completely 'hard' shadows, perfectly mirrorized reflections, etc. It is only when the premiums are activated that the distribution becomes randomized.
You'll notice that are is no difference in the sky @1 rpp vs. 256 rpp; because the sky is not a surface and there is no distribution of rays; unless you use DOF. which leads me to believe that DOF is a camera effect.

Bryce does not bounce light; even with true ambience. TA lets nearby surfaces 'pick up' diffuse, ambient, and specular color.
I did lots and lots and lots of test renders with TA; With standard lighting and with IBL lighting.

The amount of color that a surface picks up is controlled by it's own ambience value.
The amount of color a surface 'emits' is controlled by it's illuminance. Ambient value causes 'self illumination' that increases the objects illuminance, but there is no actual luminosity being emitted, as with a light source.

Set a plane to 0 ambience, 100 diffuse; and sit a 100 ambient, 100 diffuse object on it, render w/ TA. There will be no color bleed effect in the plane (but there will be color bled into the ambient object from the plane). Turn up the plane's ambient and you get color bleed effect on the plane. If there was really light bouncing; then the 0 ambient, 100 diffuse plane would be illuminated by the bouncing light.
A 0 ambient, 100 diffuse (or specular) object on a 100 ambient plane will cause color bleeds also. Not as much as if the object had ambience to boost it's output; but if there is no ambience anywhere, there is no color bleed.
There is more TA effect w/ higher RPP because there are more rays being used. There are no multiple light bounces, because there are no light bounces to begin with. TA is a valuable rendering tool; but it is not global illumination (the ability to bounce light).

Global illumination (photon simulations) is not the only way to calculate indirect light. Radiosity does a similar thing, and so does final gathering, and probably more techniques.

The fact that a higher ray depth increases the amount of indirect light suggests either bounced light or an approximation is being calculated. I'll agree with you that it probably is an approximation and that its not physically correct, but the visual effect is still one as if light was bounced. It's just that the controls of it are a bit awkward, having the ambience control the amounts of picked up and reflected light. As with almost any premium render effect, they tried putting control over too many things in one channel at once.

The amount of rpp only affects noise levels, not the amount of TA effect. The ray depth however does increase the TA effect. With Ray depth im aiming at the same value that allows a maximum number of reflections and transmissions.

If you create a white surface, and have an object 'cast' a color bleed on it. Does the bled color change when the casting object's diffuse and/or specular color change? i was under the distinct impression only the ambient channel bleeds color...