AuthorTopic: Second Cluster Study - Knights of the Round  (Read 36689 times)

Offline Reo

  • 0010
  • *
  • Posts: 200
  • Karma: +0/-0
  • aka MechaBarbarian
    • MechaBarbarian
    • http://pixeljoint.com/p/11318.htm
    • View Profile

Re: Second Cluster Study - Knights of the Round

Reply #40 on: January 05, 2013, 02:55:47 am
WOW! This has been incredibly insightful, and fun! ;D

Here are some general thoughts on some of the edits!
Mine first: I agree my sprite doesn't really fit with the game anymore. After having done about half the work on my edit I glanced at Helms edit ( for less than a second ). And thought it looked different enough from the original that I should just go all out on it, looking at it now I might've been misstaken. I'm generally pretty happy with how my edit performs. As you say there is a pretty cool effect going on. One place where the scanlines seem to have altered the sprite though is the ''folds'' on his right boot. They appear much more uneaven. Some of my AA seem to have rendered pretty useless. As a side note when searching for references for this I  found little look in pictures of people holding actual swords in this postition, and turned to baseball instead


@Helm I'm truly impressed with the face in yours. That is a true eyeopener. It's sort of a sub-pixel effect. There might be a few pixels that stand out other than those you mentioned. Like the pink pixel over his left arm, generally single pixels seem to work against the scanlines . But generally it is one of my favorites. You're always very good at giving your work certain character and atmosphere, 'tis no expection. It also has a fantastic clarity to it which makes it one of the best imo.

@Dennis edit is very interesting, especially the dithered mail. It looks almost like fishscales. It looks really cool when it's horizontal but the effect is lost on the diagonals. Perhaps the angular shape of it make it seem a bit lifeless. Like a statue.

@Cure's edit is also great. And works wonders. It fits into a game perfectly while at the same time looking like it comes straight out of some medievil tapestry.

@St0vens is really incredible. Fantastic clustering. Very elegant solution to the red flap. Looks almost painted rather than pixeled. Even at this resolution!

@Pistachio and @Dex also deliver strong and interesting edits!

It would be pretty sweet if  there was some program that would allow you the edit the sprites while seeing a scanlined preview of it.  Again great results from this!

Offline Conzeit

  • 0100
  • ***
  • Posts: 1448
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile
    • CONZEIT

Re: Second Cluster Study - Knights of the Round

Reply #41 on: January 05, 2013, 04:14:07 am
Dont you guys think you're kinda overdoing it with those scanlines?


from the times when people have experimented with photographing CRTS in pix, it seems like the scanline wouldnt be more than half as wide as the pixel itself, or atleast not as opaque...


I've experimented with this kinda thing a bit and I've tried to recreate the fringing and the rgb offsetting that happened in this

I can send you guys a PSD of that if you want, but I pretty much just copy the layer 3 times set each layer to a RGB channel and then motionblur it sideways to create the fringing. The motionblur was 3 pixels in this 400% resized image
« Last Edit: January 05, 2013, 06:43:53 am by Conceit »

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: Second Cluster Study - Knights of the Round

Reply #42 on: January 05, 2013, 04:45:44 am
That looks like from a really bad monitor, way fringe.

But yeah, the overall thickness of a scanline should be based on lightness level of the pixels on it. I am kinda trying to emulate that to a degree, but it is not super easy just in PS.

On the C64 for example single pixels or 50% hires dither in dark colours like blue and brown to black will be eaten into significantly by the surrounding black.

And RGB offsetting and all that is not something that is automatically inherent in CRTs, it just happens when the tubes get magnetised or are otherwise decalibrated (tho a bit of it adds to the charm I think)
There are no ugly colours, only ugly combinations of colours.

Offline Jeremy

  • 0010
  • *
  • Posts: 139
  • Karma: +0/-0
  • howdy
    • pixellerjeremy
    • http://pixeljoint.com/p/19218.htm
    • pixellerjeremy
    • View Profile

Re: Second Cluster Study - Knights of the Round

Reply #43 on: January 05, 2013, 04:47:54 am
Ended up with 16 colours+trans. Whoops :yell:



Legs are different sizes too. Bother.

E; sword hilt is wonky too I think i'm blind

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: Second Cluster Study - Knights of the Round

Reply #44 on: January 05, 2013, 05:10:04 am
Here we go. I came up with a better as well as cleaner scanline process for this. I will go to bed now and hope that I can remember the process of what I did when I wake up. It is a bit more complicated what I did before, but way more accurately depicts glow based on relative lightness.

There are no ugly colours, only ugly combinations of colours.

Offline Conzeit

  • 0100
  • ***
  • Posts: 1448
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile
    • CONZEIT

Re: Second Cluster Study - Knights of the Round

Reply #45 on: January 05, 2013, 06:34:47 am
O.O Give us teh sekretz once the activity is done =) awesome job ptoing!

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: Second Cluster Study - Knights of the Round

Reply #46 on: January 05, 2013, 12:58:28 pm
NEVAR!!! Actually, I am just starting of doing it, and I think I got an idea to make it even better!

But yeah, I will post a rundown of it in here.
There are no ugly colours, only ugly combinations of colours.

Offline 0xDB

  • 0011
  • **
  • Posts: 873
  • Karma: +0/-0
  • Dennis inter-is.
    • dennisbusch_de
    • http://pixeljoint.com/p/1287.htm
    • 0xdb
    • View Profile
    • 0xDB

Re: Second Cluster Study - Knights of the Round

Reply #47 on: January 05, 2013, 01:43:25 pm
How do you apply the scanlines?
My process is probably similar to what everyone else is doing. I first blow up the image by 700x1200 percent (unfiltered pixel resize), then I make a 1x12 image where I fill the lower 5 pixels with black and keep the rest transparent. I then use the 1x12 image as a fill pattern for a floodfill on a second layer in the big image to add the scanlines. Then I add a barrel distortion effect and afterwards reduce image size again, this time with a filtered resize to get some additional blur.

Dennis' edit: The shoulder pad sadly looks like it's in front of the face, though I like the decorative idea. The chain-mail works remarkably well. The primary hand holding the sword is mushy as hell, though. Unfortunately, the groin edit, though good as a drape study, makes him look as if he has no pelvis and a tiny waist. More work could be done on the boots, more contrast.

Dennis' second edit: The yellower highlight was a double-edged sword. Contrast reads better, but the finish now is kind of golden, confuses the materials. Primary arm less mushy for sure.
@Dennis edit is very interesting, especially the dithered mail. It looks almost like fishscales. It looks really cool when it's horizontal but the effect is lost on the diagonals. Perhaps the angular shape of it make it seem a bit lifeless. Like a statue.
Thank you both for those hints. I'll try to make another edit taking those observations into consideration.

I'm afraid I do not currently have any feedback to give because pretty much everyone here is a way better artist than myself and my mind feels a bit blank atm, so my crits would likely end up being inane expressions of deep admiration for what I see without much food for thought or constructive pointers towards what could be improved (I might still try doing that later).

All I have to offer for now is a photo of a real CRT monitor set to 768x448 pixels at 4:3 screen ratio, displaying the edits which have been posted so far:

http://www.dennisbusch.de/shared/drawings/2ndClusterStudyScreenPhoto.png

That photo might be a useful reference while trying to perfect any digital imagery filtering methods to recreate the effect. I should add, that monitor is about 8 years old and has lost quite a bit of color brilliance already.

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: Second Cluster Study - Knights of the Round

Reply #48 on: January 05, 2013, 02:55:59 pm
Dennis: Why did you scale up by 700x1200 percent? That wont give you the correct ratio. You need to scale up 700x900% at some point to get the proper 4:3 ratio in the case of CPS games.

Anyway, what I did to get my scanlines.

  • Scale image up 200% with nearest neighbor (NN)
  • Make a new layer on top and put in scanlines, OFF ON OFF ON OFF ON...
  • Turn off scanline layer, apply gradient map from black to white with smoothness at 0% on the original layer.
  • Use this greyscale layer as a layermask on the greyscale, but inverted. so basically white pixels from the original will mask out the scanlines completely.
  • Scale up 700x900% NN
  • Duplicate a layer with everything you got now and make a greyscale gradient map at 0% smoothness and use it as a layermask on itself so that brighter stuff gets not masked out.
  • Merge that layer down into a new layer so that the layermask gets applied.
  • Now you have 2 layers, the bottom one with the scanlines and one which you got from a brightness mask basically
  • Now apply a motion blur 2 times to the top layer, 14 pixels vertical, 18 pixels horizontal.
  • Now scale down bicubic, so that the height of the image is exactly a third of what it was.
  • Scale down again 50% bicubic
  • Make a new merged layer from with everything and put it on top
  • Gaussian blur this layer, do the greyscale gm at 0% again and apply it to itself.
  • Set to soft light at like 50% or so.
  • Shift around the channels a bit. I shifted the R channel to the right 1 pixel.

Also, updated scanlines with all entries in original post.

« Last Edit: January 05, 2013, 03:00:06 pm by ptoing »
There are no ugly colours, only ugly combinations of colours.

Offline ErekT

  • 0010
  • *
  • Posts: 330
  • Karma: +0/-0
  • fistful of pixels
    • View Profile

Re: Second Cluster Study - Knights of the Round

Reply #49 on: January 05, 2013, 03:30:13 pm
Quote
Erekt's edit is top-darker, which is a bit strange. The tapering of the round ends of the gloves and the red flap does not help the sprite's flow. The looser tunic over the belt is a nice touch, with a pretty highlight after filtering. Angular arm holding the sword also seems to curve outwards as the red flap edge. Is this a pattern in your art, Erekt? Pants and legs could have used more love, to make the sprite cohere. Erekt could study Cure's edit.

Yeah, I tried to fix the difference in contrast and lack of attention to the legs in my second version. But it could use more work still. Not sure which part of the arm you're referring to exactly. The upper arm? The flap curve was a conscious choice but not the arm. The original arm felt too blubbery and too long so I tried to rectify that with an angular shape. But if I bend things in unnatural directions without knowing then that's definitely something I'd like to fix! I don't really see it for the arm tho, can you clarify? And thanks for the feedback, I appreciate it :)

@Dennis: I noticed something that looks like a structural problem with your man's left thigh. The er, bulge in the loincloth just below the belt, the straight angle, and the highlight on the leg leading all the way up to the cloth suggest to me a pole-ish shape that ends abruptly below the belt. It's more visible in the second edition. And maybe the leg could be brought back some more with darker shades? I like how you did the scales and the shell-like shoulder pad. Good stuff.