AuthorTopic: fruity shading experiments  (Read 17053 times)

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #20 on: July 07, 2012, 02:47:15 pm
Maybe use the sawtooth / square wave style dither to break up the banding at the top?  Definitely looking better though.

Tourist
Thanks:)
I've tried it, it shouts HELLO I'M DITHERING!!!!! WHAT IS YOUR NAME?? ... I LIKE PIE! (seriously, the result is like I actually pasted a checkerboard pattern over it, even though there is no literal checkerboard pattern in the image. actually looks much worse than the banding does. ) When I get access again on Monday, I'll post an example.
I think this is because the direction of the dithering is at odds with the actual angle of the plane.

That said, after adding the next subdivision level, the banding will reduce drastically.
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #21 on: July 27, 2013, 06:16:50 am
RISE FROM THE DEAD!

I just picked this up again in the past few days, thinking I can also include vegetables/ 'vegetables' (eg capsicum). Haha, that plum is looking pretty blurry to my now more-developed artistic sense, oh well.

Old:


New. with Tasmanian purple garlic added (a mere 18 colors inc bg. Here's a ref photo; I used a physical ref myself.):


Trying to improve my ability to use clusters (and clusters of clusters) to emphasize particular parts of shapes. A partial success on the garlic, I feel. I would have liked to convey the 'succulent bulginess' of the individual cloves better, but struggled with the overall low-contrast nature of garlic skin.

Also used my new 'adv_spread' script for GrafX2 to generate intermediate colors when needed. It's like a vastly improved version of the builtin Spread function. When you select adjacent colors as the endpoints, it automatically inserts a color between them and adjusts the picture accordingly, so it's very easy to use with the 'ramp subdivision' rendering method I'm using for this series. It interpolates in Linear RGB, LAB, or LCH colorspaces, but isn't yet published on the web.

I only just looked here and saw I said that I'd subdivide the lemon another step; I've just begun on that.

EDIT: Wow, this is a real test of my cluster-stacking abilities -- I must be out of practice. Happily, even though it's a greater effort than I recall, I'm getting a much more satisfying result, quicker. It seems to take quite some skill to stack up clusters in a regularized/conforming way without introducing inadvertent dents or bulges, but I'm finally getting there.

EDIT2:
Here's the finished subdivided lemon. I'm pretty happy since I banished most of the banding without damaging the sharpness that much. 23 colors in final.



Bonus: Silly impromptu lemon typography:


EDIT3,4:
And here's the next addition: a pineapple. Still WIP:



EDIT5:
Nearly done with the pineapple's basic rendering:


Currently 19 colors, I'm struggling to imagine how I can usefully add more than about 3 colors, looking at the current state of it. Pointiness needs few colors, apparently.

I also have a capsicum worked out and just waiting for rendering.
« Last Edit: July 31, 2013, 08:52:08 am by Ai »
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #22 on: August 02, 2013, 12:55:37 pm
Capsicum:


Early WIP, only 15 colors.. I seem to have hit a real 'Amiga feel' with this one -- I guess capsicums naturally have the right aesthetic qualities to fit that.

EDIT,EDIT2:


Yeah, this is going well. (23 colors. The 'recursive shading' thing gets a bit more complicated when you're reusing an existing ramp.)

It looks like the GrafX2 quantization code did something unexpected to my palette at one point, as I noticed I now have a few consecutive duplicates in older ramps (plum+cherry). I can fix this, I guess I'll just have to stay away from quantization for this piece.

EDIT3:
NEW:


Yeah, I'm getting quite happy with the planes in this now. Reached 26 colors. 3 or so more red shades seem like they'd be useful. This one might manage to get up to 32 shades.

I started automatically versioning this using 'watch' and a little shell script I put together 'gitwatcher'. Hopefully the playback/slideshow will be enlightening.

EDIT4:
NEWEST:


Nearly done with this one, I reckon.
33 colors at the moment, no obvious additions to make.





Most recently I've been learning to sculpt in Blender, so I made a little materials/shape mockup to help rendering the carrot;
from that reference I produced this:



Currently I feel it's too dark (partly caused by my patchy 'reuse existing ramp' skills and a semi-failed attempt to stack the full set of shades one by one instead of recursively refining.). Maybe a little too knobbly, too, but I can live with that as a mere stylistic choice.

However, the material and sculpting mock did really help a lot to shake out how I wanted this to look, and I'm getting faster all the time. Digital sculpting is  so amazingly fun it's almost beyond belief :D

I'm also pretty struck by the strong difference in definition between the old (items on the left) and the new (items on the right)

EDIT:

Fixed the darkness problem and the texturelessness.
« Last Edit: August 11, 2013, 11:38:07 am by Ai »
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Crow

  • 0011
  • **
  • Posts: 647
  • Karma: +0/-0
  • Technicanimal
    • View Profile

Re: fruity shading experiments

Reply #23 on: August 11, 2013, 10:58:58 am
Old account restored, posts merged; have my post as a bump~
Discord: Ennea#9999

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #24 on: August 14, 2013, 07:34:53 am
Thanks :)

Here's my latest progress:



I've got a few different versions of the tomato there -- I was trying to avoid being bland and mocked up a striped tomato ala Russian Black, but when I went to render it I couldn't get it to work, so I ended up doing this sort of swirly sparkles thing, end result being a mix of 'fleshy' and 'glassy' feel.

I'd like to re-emphasize though, that Blender's material system and sculpting is really great for getting a strong idea of how I want a particular item to feel. 'Subsurface scattering' option is especially useful for anything fleshy (ie. most fruits :)

I've done a lot of drawings for this by now, when I look at my list, I see I've sketched:

* Apple
* Apricot
* Blackberry
* Cherimoya
* Grapefruit
* Cashews (yes, nuts are classed as fruit)
* Jackfruit
* Mango
* Peach
* Pear
* Persimmon
* Plum (yeah, I already rendered this but I can do much better than that!)
* Cherry (ditto)
* Quince
* Raspberry
* Strawberry
* Broccoli (okay, not fruit, I ended up including some veges too)
* Sweet potato (ditto)

The original list are not prioritized because the old sketches are crappy esp WRT planes  :lol: But those fruits are included in plans -- I actually based the total list on wikipedia list of fruits, so it's quite inclusive.

I'll show or link some of those drawings here when I have a large block of time available for scanning.

Some interesting things to learn here -- I wanted to reuse the capsicum color ramp but mostly couldn't because it was too desaturated to give the right intense tomato red. That kind of issue doesn't come up with low-color icons, since all the colors are approximate and cover broad areas, there's more leeway for inaccuracy (as long as it's an -interesting- kind of inaccuracy, for example the wild exaggerations of color, pattern and value encouraged by EGA palette.)

Still got about 110 palette slots free -- a few colors were reusable for the tomato, and most colors from the kumquat were reusable for the carrot.
I guess I'll start a 'part 2' image when I run out of palette.

EDIT:
Fixed up some of the texture + excessive glassiness issue:



Reasonably happy with this now, it still visibly needs  about 10 more colors (currently clocks in at 21 colors FWIW)
« Last Edit: August 14, 2013, 02:32:33 pm by Ai »
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #25 on: August 18, 2013, 12:25:57 am
I thought if I changed my process a little I might actually get some crits.


Working in parallel on 3: Raspberry, Persimmon, Strawberry.
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #26 on: August 23, 2013, 12:43:42 am
I chucked it onto a mid-grey BG and cleaned up any selout:


Also refinement, and early apricot and mango. Tomato is done, I'm happy with how it's smooth while still having character.
« Last Edit: August 23, 2013, 12:46:06 am by Ai »
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #27 on: September 03, 2013, 11:44:54 am

Update, Head-exploding strawberries edition. Slowly getting there.
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.

Offline ptoing

  • 0101
  • ****
  • Posts: 3063
  • Karma: +0/-0
  • variegated quadrangle arranger
    • the_ptoing
    • http://pixeljoint.com/p/2191.htm
    • View Profile
    • Perpetually inactive website

Re: fruity shading experiments

Reply #28 on: September 03, 2013, 11:57:45 am
Some of them are quite nice, but I think you should try to get solid volumes without as much colours as you are using right now.

I think the bell pepper is best atm, just as far as overall volume and lighting goes. The tomato has nice lighting but could probably need a bit more definition, plus the green stem bits at the top. The raspberry and strawberry look largely invented and do not really have good volumes/look pillowshaded as far as the overall shape goes.

I might attempt some edits at some point, but atm I am busy as well as having troubles with my wrist :/
Keep at it :)
There are no ugly colours, only ugly combinations of colours.

Offline Ai

  • 0100
  • ***
  • Posts: 1057
  • Karma: +2/-0
  • finti
    • http://pixeljoint.com/pixels/profile.asp?id=1996
    • finticemo
    • View Profile

Re: fruity shading experiments

Reply #29 on: September 03, 2013, 02:26:37 pm
Thanks for the critique :) I agree with all of your crits completely, actually -- I'm probably better off restarting the raspberry and strawberry from reference (they were originally done from relatively-recent memory, as I was trying to improve that skill. Volume was never really right at all and I just shoved them around frustratedly). In general I think I start with 4-5 shades (after doing flats) and yeah, 3 or even 2 would be better. Good to be reminded of what is really important.

It might also be good for me to use contour-fill a little less, as it gives an addictive feeling of effectiveness but often rather approximate results.
If you insist on being pessimistic about your own abilities, consider also being pessimistic about the accuracy of that pessimistic judgement.