Pixelation
General => General Discussion => Topic started by: Helm on November 10, 2009, 03:36:10 pm
-
I've been thinking about the critique process and how we could attempt new things, so here's one. Does anyone have specific questions about how a specific active artist in the forums does things? If so, why not ask directly and perhaps they'll answer, perhaps someone else will answer, perhaps we'll have a bit of a dialogue. The difference I'm aiming at is between being told something and having asked to be told something. Critique as we do it in the pixel art forum isn't much of a dialogue (though sometimes it is) and I've always thought the learning process works better when it's directed by the one asking to learn. So why not try this? I will - of course - answer any questions directed at me (with examples if I get the time) and if you ask any active mods I'll try to get them to answer (with black sorcery) but other users it's up to their time and courtesy. Let's try to break away from the 'hi here's my art, fix it plz' mode of critique we've been doing for a while, not that it doesn't work, but other approaches also might work.
The more specific the question, the better. 'How do you make your pixel art' begs for a tutorial, and I've attempted something like this before (http://www.wayofthepixel.net/pixelation/index.php?topic=8110.0). But here let's be more specific about techniques, styles and processes, right?
-
(with black sorcery)
Aha! so that's your secret. ;D :lol:
Well, I like the sound of this idea, cause I really have a lot questions all the time... :lol:
...questions to the many veterans which can help a lot to the new / intermediate pixel artists (like me who only has 9 months doing it). As far as help being offered, I personally have learned a lot in overall by just reading other threads, which is something I been wanting to point out in terms of feedback... helping someone with edits and or feedback not only helps the person who's asking, but helps others as well who might have been wanting to know the same thing.
I think that this can help a lot, also what would help a lot would be to make a thread where people vote or request tutorials to be made... for example, there is little to none tutorials about pixel art animation, like fighting characters and other stuff out there... all you find is the same thing over and over, while there are soo many topics and things to know out there that no one really bothers to share. In order to maintain and keep the pixel art community growing I think new pixel artists need to know more about pixel art, and for the most part learning from people who has been doing this for a while, is the way to go.
Lol I hope this doesn't sound too confusing man... as far as me asking directly, I prefer to just make a thread and ask for the specific feedback, since I as pointed out above, I like it to be public so others who might have the same question, be able to benefit from it. :y:
-
How often do you moderators use references when pixelling, in comparison to when you first started? Do you find you have 'internalised' a lot of how for example a tree, or grass, or a face would look?
-
I think that this can help a lot, also what would help a lot would be to make a thread where people vote or request tutorials to be made... for example, there is little to none tutorials about pixel art animation, like fighting characters and other stuff out there... all you find is the same thing over and over, while there are soo many topics and things to know out there that no one really bothers to share. In order to maintain and keep the pixel art community growing I think new pixel artists need to know more about pixel art, and for the most part learning from people who has been doing this for a while, is the way to go.
I find that the best way to learn lots of things is to discover them for yourself, only by trial and error do you come to understand why techniques are important/better. Don't get me wrong, I think that C+C and advice/help along the way is great and important but an explicit tutorial is only going to teach most people to copy.
-
happymonster: That totally depends. I think looking at references even (or especially) if you think you know how something looks can help. It helps to prevent you from falling into formulas. Formulas work if you have to get something done or need to work in a specific style. But you should figure those formulas out yourself. And I think it often helps to simplify from nature. So yes, looking at reference when needed is good. I do it.
-
I'll post a little one-liner thing since I wanna reply in here without having to write an essay just cause I'm a mod : D
I use references all the time, I've got such a copycat brain that if I have no input before attempting to do a scene or something I often have to fiddle around a great deal to come up with small details and get proportions at least half right - but if I have some kind of reference to follow I can at least get some general guidelines in how to construct something or make it more realistic and less of a mess made out of the misguided symbols lodged inside of my brain. Heh.
On the other hand things such as 'a tree' have clear clusters of ideas in my brain which enables me to draw it without reference.
Actually I think one of the reasons to why I can do trees is because I made a one page tree-study one day. I just looked at trees around me intensely and drew them, I even drew from reference outside the the window of a speeding train - it was the ultimate lazyass studies, but man, it expanded my brain.
Thus I'd rather go to England and draw green lush fields for a week than google 'fields' and the like in google image, but hey, can't have everything. But if you're ambitious about your art then by all means do that or something like it - real life studies give you lots of internal reference.
But anyways, I've internalized most of the things that are within my comfort zone, I think.
On the other hand my comfort zone is small and sucky.
Soo I have to use lots of references when not drawing things that are either girls, robots, jellies or both. I mean all of them.
Haha, photocopier, great to hear from a guy with your name that you should copy less and explore more yourself, haha!
Anyways I dunno, it depends on the person I guess? I think that other people's edits of your own images can take you much further than just reading a tutorial - the feedback is the thing that separates them.
You can copy from a tutorial all you want, and the only feedback you get is the feedback you give yourself and that can be anything from awesome to supersucky depending on your self-confidence and experience.
I remember posting a pixel piece here ages ago, being very proud of what I was able to achieve with my total lack of experience, feeling I was totally satisfied with it.
And then someone said 'very low contrast on that one' and just upped it a great deal and posted an edit of it, and it was like having been struck by the Baseball Bat of Lucidity, because suddenly I SAW how contrast works in pixel art and how colour transitions that look harsh on 5x zoom blended into harmony at 1 and 2x.
It was an insight that helped me with ALL of my art, and not exclusively pixel art.
So yeah I dunno what I'm saying, I guess I'm saying that tutorials are fine but feedback is better? And that you don't need to learn EVERYTHING by trial and error, because sometimes peer feedback will give you the same insights in the tenth of the time it takes for you to discover it yourself.
After that, it's all about using the knowledge for yourself though. But actually I've noticed my pixel art skill growing just from LURKING pixelation for months and not even pixelling myself. But then of course I have to doodle and scribble all the time - gotta keep your general art sense alive.
Anyways tutorials are great for inspiration, I think! To just see how st0ven could make awesome super detailed dragons by just drawing a silhouette and then filling it in without almost even changing the silhouette - it baffled and inspired me and while I eventually learned I couldn't do exactly that on his level of refinement at least trying that process would give me some new insights. Also tsugumo's original tutorial totally made me feel like 'I CAN DO THIS THING' whereas if I hadn't seen it ... I wouldn't be here. : D
Ramblings end, this was such an essay but without a point. I hope someone gets something out of this. Also photocopier, I realize I'm really saying the same thing that you are. Heh. So don't take this as some kind of arguing against you!
Cheersbyeorsomething
-
Thanks guys! That's useful..
-
I have my own question if it's not too much trouble - how do you flesh out the stick figure once you have a nice motion that your happy with?, If you've got any tips at all it would be very helpfull and much appreciated.
-
Antifarea:
As far as help being offered, I personally have learned a lot in overall by just reading other threads, which is something I been wanting to point out in terms of feedback... helping someone with edits and or feedback not only helps the person who's asking, but helps others as well who might have been wanting to know the same thing.
I'm glad this is so, sometimes I do extensive edits for newcomers with 2 posts that never get back to me and/or update their work (I don't mind if they do and don't integrate my points, I mind when they never show up again) and it's kinda discouraging, it helps to think that other people might be benefiting from it. And of course, on the personal level, *I* am benefitting from it. I've grown tremendously, not as an artist but as a craftsman in Pixelation just by practicing the end level of pulling together a piece. The first few levels of how it's done could use a lot of work but boy, have I learned to do the last 5% heh
I think that this can help a lot, also what would help a lot would be to make a thread where people vote or request tutorials to be made...
Personally I think this won't work, because a tutorial is something that requires extensive work and often I've found that one has to start from the bottom up and explain everything (like I tried in my attempt in Ramblethread) so it's more like... life undertaking than 'oh someone asked how to do a fighter animation, let me whip up some examples right quick".
happymonster:
If it's in my comfort zone, I don't use any reference. If it's outside, I tend to go to google image and just look at a hell of a lot of stuff and try to absorb as much as I can, but unless explicitly stated, I don't trace or 1:1 copy what I'm seeing. I come from a comic art background where what is prized - in my opinion - more than fidelity, realism and plausibility in rendering, is that the artist conveys his uniqueness in every part of what he draws. So whereas I might look for example, at pictures of cars when I want to draw a car, I will not draw a specific model or brand car, I will draw a Helm car, and I think aesthetically that's more of a priority than just doing the job of the illustrator with realism and plausibility, if you get what I mean.
What I found I gain from perusing reference before I start drawing is that it just sparks up the imagination even further because it reminds you of all the little things that a real thing has about it that you'd forget if you started with 100% imagination. If I say 'draw a kettle' to my brain, it might forget the bottom rim, or to add a curvature to the nose, or to add a split tip to it or whatever thousands of years of kettle engineering have found to be useful to have on there. Once I remember that stuff, I can then reinterpret them in a Helmy way.
Vercingetorix the Galois:
Your question is an animation question, right? I'm not an expert, which can be illustrated perfectly by how I do not use stick figures at all, I work with vague silhouettes when I start to make an animation and once I'm happy with the key frames, I will pixel them fully and then use parts of them rotated and then cleaned up and redrawn to make in betweens. I am a low level animator, I don't know many tricks besides easing a frame in and out, basic anticipation and if there's anything I'm trying to learn is that sometimes less frames/fluidity is better. Animation gurus will answer your question more fully, I hope.
-
I'm glad this is so, sometimes I do extensive edits for newcomers with 2 posts that never get back to me and/or update their work (I don't mind if they do and don't integrate my points, I mind when they never show up again) and it's kinda discouraging, it helps to think that other people might be benefiting from it. And of course, on the personal level, *I* am benefitting from it. I've grown tremendously, not as an artist but as a craftsman in Pixelation just by practicing the end level of pulling together a piece. The first few levels of how it's done could use a lot of work but boy, have I learned to do the last 5% heh
Yes, I know what you mean, and I have once been guilty of that, but in terms of not posting back, but I did updated the work though... I remember I think it was you or ptoing who helped me haha please bear with me my memory fails me as of lately, but I back then asked for help twice for an sprite, one for edge from star ocean 4 and nowe from drakengard 2... and I really started to grasp things easily after I got the feedback and understood what was really wrong with it, so yeah, it always helps, and I bet in most cases people just either forget to post back or just have something else to do at the moment, but I'm pretty sure most people who you edit stuff for or reply to, at the very least consider applying your tips to the piece. :y:
Personally I think this won't work, because a tutorial is something that requires extensive work and often I've found that one has to start from the bottom up and explain everything (like I tried in my attempt in Ramblethread) so it's more like... life undertaking than 'oh someone asked how to do a fighter animation, let me whip up some examples right quick".
Well either way, what really concerns me is to get feedback on animation... I know I seen it in other threads and like I said before I have gotten tips, but still, what worries me is the fact that there isn't really anything about animation out there other than that warner bros drawing thing from some guy that was an artist in the 60s haha... it's really hard to be honest, you don't even know where to start, so my question is simple... how do you guys animate frames? do you make one sprite and then edit it to change the expression? or do you just make each frame entirely from scratch? that's the thing though... most people don't know where to start, and mainly cause there isn't any info out there about this important thing, it's more like everyone goes the hard way, of learning on their own... and why is that if there are soo many people that already know? why not share it? it might take a long time, but think about it... it will help many many people who are in need at the moment haha... so yeah, do it. :)
-
what worries me is the fact that there isn't really anything about animation out there other than that warner bros drawing thing from some guy that was an artist in the 60s
Anything???? There are plenty of good books on the subject, and I've learnt about the basic elements from observation, experimentation, and reading on the 'net.
http://en.wikipedia.org/wiki/Animation#2D_animation
lists a few important topics (onionskinning, tweening, morphing) and provides some other stuff (eg a link to a video for animation techniques)
http://en.wikipedia.org/wiki/Keyframe
I plan to buy this DVD eventually; there are some clips excerpted from it available on the website:
http://www.theanimatorssurvivalkit.com/
I'd also advise you to get Deluxe Paint IIe (downloadable somewhere or other; DOS), ProMotion (trial/commercial ($25?); Windows) or Grafx2 (Free; Windows/Linux/DOS/Amiga), and play around with the function that lets you pick up parts of the sprite and use them as brushes... this makes moving objects / parts of objects dead easy so it helped me understand animation better.
Playing around with the animation features in Blender also helped me to understand keyframes.
So a breadth of experience is very important to acquire.
As to methods, I can describe how I animate:
(I use GIMP-GAP, which is a rather comprehensive animation/video editing add on for GIMP.)
I think out a few keyframes and draw them as stick figures.
I step through the animation and see if I'm missing keyframes, in which case I add the additional frames.
Using onionskinning is very important for this.
After keyframes are planned, I turn them into color-'coded' silhouettes using a 3x3 brush, anatomy knowledge and experience :) and clean up important edges with a 1pixel brush (important edges are those in the 'front edge' of the movement of a particular part, ie. the edge that is least far from the destination created by the animation).
Onionskinning is even more important here to be sure that the volume of the parts are consistent.
At this point I usually clean up one frame completely to act as a proportion/volume reference for the others.
I make sure that any secondary features (eg flowing scarf) are also cleanly and consistently animated here.
Now I finish cleanup so that every edge is clean.
I do the tweening; I look at the paths the different parts are taking in the keyframes and try to interpolate smoothly between them so that each part moves along a coherent curve or line. I avoid tweening very fast moving parts,
as they should typically snap between keyframes, or have just 1 or two tweened frames biased a lot towards either the end or start of the motion... so some duplication of a part from a previous frame happens here.
As to where to add tween frames.. it's just about getting an acceptably smooth representation of the motions so that relatively slow movements look smooth and fast ones look snappy.
Note that when you're tweening, it's even more important to consider each individual joint and its range of motion in the animation in order to get the most expressive result. So anatomy understanding is terrifically important here even if your character is super cartoony.
btw, have you read what Tsugumo has to say about animation? http://mirror.motherhamster.org/tsugumo/default.html
If not, why not?)
After tweening, the actual rendering starts. There's not much to say here except you need to develop solid lighting skills and the ability to concentrate on a style and apply it just as equally to every part of every frame.
I try to render the whole animation at once, rather than individual frames, in order to increase consistency and
spot problems sooner. (eg. adding each level of shading one by one to the animation, rather than one by one to the individual frame)
Copious amounts of copy+paste or brush picking tend to happen here :)
what else: I save frequently, and run the animation again every time I make a significant step forward in the creation
of the animation.
-
This is a cool idea for a thread. :)
Here's some questions. Sorry if I'm asking questions to somebody who isn't very active anymore. I've been kind of out of it myself lately, so I'm not real hip on who's still here or not.
Kennethfejer: What's your process for picking out colors? Your art always strikes me as "default" pixel art--all pixel art "styles" kind of stem from your traditional looking style in my mind--and the main reason is your color choice. Do you have any special reasons for picking the colors you pick? Any specific games you draw color inspiration from?
Panda: How did you develop your antialiasing technique? It's hard for me to eyeball which things to aa and how much aa to apply. You've got this down to a science it seems.
Helm: How do you decide when to include a pixel for accurate representation and when to omit a pixel for space constraints? This is very difficult for me. I can see how an object looks in reality, but there's either not enough room to show enough information or too much pixels being wasted on open space. You always seem to have the perfect balance. What's your trick?
Monsoon2d: What did you study to learn your dithering theory? You use dithering to a very satisfying affect, and it would be interesting to get some reasoning and story behind why do you it how you do it.
I'm sure I'll think of more questions for others later, but these are ones that stick out in my mind right now. There are specific pieces I'd like to dig up and ask specific questions about eventually.
-
Panda: How did you develop your antialiasing technique? It's hard for me to eyeball which things to aa and how much aa to apply. You've got this down to a science it seems.
A pact with the devil >:D
In all seriousness, I haven't done anything worth mentioning.
I've been pixelling on and off for around 6 years now, and eventually it just clicked.
As for what to AA and how much to apply, that is up to you I think. Depends on what you want to achieve.
Personally I don't like seeing blocky jaggies, but I don't like blurry stuff either, so I always try to find a middle point that works for both.
It all comes down to a mild OCD behaviour, your point of view on pixel art and taste, I suppose.
One tip though, keep checking a preview window or zooming in and out a lot to check if the AA you applied works. If it looks blurry or non-AA'd you've either AA'd wrong (too much or too little), or have the wrong values in your palette.
Oh and be patient, usually stuff doesn't come down in a day, so don't get discouraged if it doesn't work straight away.
Keep working at it, just do your thing, be merry.
-
Helm: How do you decide when to include a pixel for accurate representation and when to omit a pixel for space constraints? This is very difficult for me. I can see how an object looks in reality, but there's either not enough room to show enough information or too much pixels being wasted on open space. You always seem to have the perfect balance. What's your trick?
Usually I add more than I need and then when it doesn't work I take it away, heh. :) What's interesting is how some details - when you learn to think that way - are to slight to be ONE pixel fat, but then become TWO pixels. This might sound strange, but let me explain.
Let's look at this 'average girl face' that the internet has generated for us:
(http://www.locustleaves.com/attractive_average_girl.jpg)
Let's say we wanted to make something like this in 64 pixels width.
(http://www.locustleaves.com/attractive_average_girl2jpg.jpg)
(zoom in)
This is how photoshop resizes at that size. Look where features have blended together. Do you see a visible plectrum in that picture? Also if you were to draw it, would you have used a single pixel for each nosetril? Well photoshop has chosen to use two pixels for them, but softer.
This is the exact theory of manual AA that Ptoing has been posting here and there, just used for internal stuff as well.
(http://www.locustleaves.com/attractive_average_girl3.png)
Basically wherever a detail is overlapping between two pixels, it means its cluster should be on BOTH of the pixels, but the value (the depth) of the color the detail would have needs to be softened, do you see what I mean?
Of course you don't do all this for everything you have to pixel, but with experience you learn to create effectively subpixels between two pixels for finer detail, by involving two-pixel wide pixel clusters. It sounds counter-intuitive, but it really is not once you get used to it.
A very important thing the pixel artist must learn is that single pixels aren't as useful as the immediately seem.
-
Of course the advantage (or the bane) of the pixel artist is that unlike photoshop's resizing filters we can prioritize what details must remain visible, resulting in an image that is "sharper". But if done incorrectly it can merely result in an image that is abstract.
-
Panda: A pact with the devil! I knew it! :hehe: Thank you for the response! The basics of most of what you said are sitting in my brain somewhere--it's just a matter of making them common knowledge and habit I suppose. The biggest thing you've pointed out for me, though, is patience. I notice you say that it can take you days. How long did it take you for your bird avatar? This sort of brings up another question, too: Do you AA as you go or devote a "step" in your process for doing it, or what? For me it usually sits as a middle to end process, being the last thing for me to satisfied (or dissatisfied) with.
Helm: Very comprehensive response. Thank you. I understand the basic concept behind this kind of subpixelling, but where I run into trouble is when there's too much information to generalize into a happy soft combination.
For a basic example:
(http://img121.imageshack.us/img121/2349/examplef.png)
These eyebrows are simple enough. Pixels can represent them no problem. Maybe some more dynamic colors and it's damn near perfect in my book.
But now we go down and try to do the eye. The basics around the eye are easy enough, but then we run into a square that has: the white of the eye, the green of the iris, the black of the pupil, and a specular reflection of light. I am dissatisfied by my reduction of all that information. Each element feels crucial to tell the story of an eye. Could you shed some light on how you deal with a situation like this?
Also, again, thanks both for sharing!
-
Panda: A pact with the devil! I knew it! :hehe: Thank you for the response! The basics of most of what you said are sitting in my brain somewhere--it's just a matter of making them common knowledge and habit I suppose. The biggest thing you've pointed out for me, though, is patience. I notice you say that it can take you days. How long did it take you for your bird avatar? This sort of brings up another question, too: Do you AA as you go or devote a "step" in your process for doing it, or what? For me it usually sits as a middle to end process, being the last thing for me to satisfied (or dissatisfied) with.
Oh, I meant it more like a "keep going at it and eventually you'll be able to do it".
But sure, some stuff can take me days. It is all relative as it depends on its purpose.
For personal stuff, I take all the time in the world and slack every now and then. For work I fully focus on whatever I'm working on (unless I have a bad pixelling day, that's it :P).
The falcon took me 3 days to finish, I started the sketch one night, the next day I was almost done with it, and on the third day I just went over some stuff, fixed some things and added the last touches. Overall it took me around 7 hours-ish.
And yup, I AA as I go.
-
How does
AdariasND Christie figure out the corect proportions for his sprites? I can never figure out how to tell how much space i should give to any given characteristic when working that small. :-\
-
well, since i seem to have my share of problems (i swear i do not intend any sort of vengeful irony in this statement), what are planes and what steps should i take to figure out anatomy or structure or whatever im supposed to do :blind:
-
This is a great idea for a thread!
@Rydin: You're still thinking too much on a per-pixel basis. You need to think of how elements need to be spread across pixels to imply the information "between" the pixels.
(http://i188.photobucket.com/albums/z212/gastrop0d/edits/examplef_ed.png)
This was actually really challenging to do. Could be a great activity idea!
-
Dear Helm,
What really is your working method for infusing such interesting use of color into your pixels?
You seem to purposely use color in ways the average pragmatic observer and artistic interpreter wouldn't think to use color, and the end-result is pleasing. Like properly spiced food, aside from the natural flavor, texture, smell, etc, the spice adds another dimension of enjoyment for the eater. Sometimes you harness color in this way, causing the overall expression of the piece to be that much more sophisticated.
And yet, as I ask, I can't help but answer myself that it's just an effect of artistic intuition, something not easily attained.
P.S. many artists here do this well, and I welcome a response from anyone on the subject, but Helm seems to consistently do it and even urge others, in his crit/edits, to make a practice of approaching color uninhibited by preconceptions.
Sincerely,
Mathias
-
EyeCraft, that's nice!! I was wondering how to do that size reduction myself after seeing the original images. So thanks for the pointers.. :)
-
Mathias: some painter said once that 'as long as you get the values right, then almost any hue will work'. That's the basis of how I use colors. I try to maximize the efficiency of the spacing of my colors in the range. The ground rules I apply when I make art is saying 'will this piece go from full black to full white or not?' Once I establish the lowest value I'll use and the highest one, and how many colors will be in the inside range, I then start to think of what exposure the piece will have (think in terms of photography) if I generally want it to have rich middles and fewer darks (as usual/reality) or if I want to do something lower-key or higher. Then I think of how sharp I want the piece to look (less sharp - will need more colors, so on). Usually I work with 16 color palettes spanning the whole range from black to white.
I draw something by making a silluette in some middle shade in my range, then I cut out the volumes like as if I was a sculptor or 3d renderer and when I've got the important (usually 3-5 value) sculpture ready on a mostly grayscale palette in my range, I start to tint the greys with the hues I need. The middle of the range is where the color identity of the piece goes so if I'm drawing flesh, most of the 'pure' flesh tones will go in the middle of the range. Usually I alternate between a saturated color, and a less saturated higher color (and a lower) and then a more saturated color higher and lower, so from the middle the palette gets color towards the extremes of the range.
Usually my darkest color is black, and my second darkest is a very saturated primary color. The reason for this is emotional; I associate the basic selections of the color wheel with various emotions and I urge others to do the same. A deep blue for me has a resonant melancholy, a purple is about to spring into action, a green is sickened or diseased or just unhappy, a red is strong emotion, a yellow is distance, so on. The piece dictates what I'll use.
The c64 palette showed me that greys are my gateway from almost any color to any other color. Good to keep in mind.
In highlights I tend to move towards neon colors, there's usually some neon blue near my pure white, I don't know why I do this, just seems to happen a lot. Pure white is used for speculars more often than for full planes of the piece, I guess remnants of my 'everything must be oily' demoscene fascination.
Keep in mind that if a color is very saturated, the eye sees it as a different value (could be brighter or darker, depending on the hue) than it actually says on the HSL slider. This is where a trained eye is needed so that colors don't punch through the planes. This is an issue that ndchristie has had in his art a lot when he was starting out with pixels.
Look at this piece now and see how all these principles go into action:
(http://www.locustleaves.com/longshord17-final.png)
-
This is a great idea for a thread!
@Rydin: You're still thinking too much on a per-pixel basis. You need to think of how elements need to be spread across pixels to imply the information "between" the pixels.
(http://i188.photobucket.com/albums/z212/gastrop0d/edits/examplef_ed.png)
This was actually really challenging to do. Could be a great activity idea!
Also I looked at this reply again and it stunned me how well it illustrates this point: when working with AA, looking at what you're trying to convey, in single pixels, is mostly worthless. Great, great image, Eyecraft.
-
That's brilliant Helm! Although I think your emotional reactions to the very dark and saturated primary hues are not universal. I never see yellow as equalling action, although I agree with some of the others you mention. You did say "for me" so this isn't a criticism, just an observations for others who want to use the same approach.
-
An elegant response, I'll be reviewing it as time goes on and certainly as I start on new pixels. I intend to post up some work when I have it and check it against what you've written. Thanks so much for being willing to share your experienced point of view on things and not hoard it like many, saying 'well it's just a trick of the trade'.
-
I agree, EyeCraft, it would be an awesome challenge. ;D
Your interpretation of the eye makes sense. It's frustrating because I know there's more information there to represent, but there's just not enough room to squeeze it all in. How I've done it in the past is that the most important descriptive detail wins the square. It seems, however, that it's more accurate when the average of the sum of details wins the square. Which makes sense because when a computer compresses the information, it doesn't care if it's an iris or a eye lash or a beard hair, it compresses it all the same.
Am I starting to get a grasp of how and why this works how it works? If so, it's just a matter breaking my old bad habits and applying this concept.
-
Thanks Helm, another concise colour process description ;D
That's brilliant Helm! Although I think your emotional reactions to the very dark and saturated primary hues are not universal. I never see yellow as equalling action, although I agree with some of the others you mention. You did say "for me" so this isn't a criticism, just an observations for others who want to use the same approach.
Actually, yeah. I started wondering to myself what, if anything, colours meant to me. When I was younger I had very definite associations, but as I've grown and worked more in art colours have become a lot more flexible to me. For instance, I used to hate yellow. It was sickly, vile, like vomit or urine. Green was soothing and appealing. Blue was cool, clear, pure. Purple was fluffy, warm, bright. Red, very hot, hard, dangerous.
But now for me the kinds of associations I have are more contextualised (usually with mixes of colours). If there's green shadows I find it a kind of romanticised alienating feeling, perhaps introspective feeling. I guess if I can expand those kinds of associations back to being more broadly associated with hues themselves, I might be able to make more interesting colour choices. Okay, I'm really just rambling to myself now. :lol:
I agree, EyeCraft, it would be an awesome challenge. ;D
Your interpretation of the eye makes sense. It's frustrating because I know there's more information there to represent, but there's just not enough room to squeeze it all in. How I've done it in the past is that the most important descriptive detail wins the square. It seems, however, that it's more accurate when the average of the sum of details wins the square. Which makes sense because when a computer compresses the information, it doesn't care if it's an iris or a eye lash or a beard hair, it compresses it all the same.
Am I starting to get a grasp of how and why this works how it works? If so, it's just a matter breaking my old bad habits and applying this concept.
That's a really good question. The idea I've had is you basically approach it with the intent to average it (can be hard depending on palette... just average to value in that sense, as Helm remarked, as long as the value is right you should be right). Then you might accentuate particular lines, edges, or patterns of colour depending on what you think is important to the total structure of the piece.
(http://i188.photobucket.com/albums/z212/gastrop0d/edits/examplef_ed2.png)
I was curious, so I did a computerised average of the same parts and compared them to what I had done. The average on the eyebrow has a tonne of colours and is basically a huge AA fest, in terms of pixel art. I exaggerated the line a little more in mine, mainly because I thought it would work and it would keep the pixels cleaner.
The eye is a lot more interesting. I broke it up in to two halves and compared them. The averaged one has a much better sense of eyelashes and pupil/iris. I think I tried a little hard to hint at the specular in my one and it has weakened the sense of the eyelash and iris because of it. I think mine succeeded in suggesting the lines of the eye, however. The averaged one has bled even more of the darkness of the iris into the bottom row of pixels, and it's really given a sense of it. That's probably the biggest thing that I missed.
So yes, approach it like averaging, but pure averaging doesn't always give the absolute best results. Just think about the subject, and bleed things across pixels to either "grow" it in size, or bleed it out to "shrink" it. :)
-
These side by side comparisons are awesome, EyeCraft. This is an illuminating learning activity, doing your own pixel interpretation of a photo and comparing the result to how the computer interprets the same photo on the same scale.
Man versus machine :yell:
-
Indeed, there's something to this process we should make an activity out of.
-
Not to derail this really interesting discussion, I've got a question about something that's a major obstacle in almost every pixel I do: Large planes (especially flat ones). I almost invariably find myself adding meaningless texture or something to break it up, but that doesn't always work for what I'm trying to do.
How do the rest of you handle large surfaces? Just have the foresight to non include them? :P
(Fantastic thread btw)
-
Not to derail this really interesting discussion, I've got a question about something that's a major obstacle in almost every pixel I do: Large planes (especially flat ones). I almost invariably find myself adding meaningless texture or something to break it up, but that doesn't always work for what I'm trying to do.
(Fantastic thread btw)
That's what I thought I should ask as well. Take the Another World/Out of This World backgrounds. How to balance small areas of high detail against large empty planes?
-
I have a question, how do you, Helm go about creating your volumes for your pixel art? I find it especially hard to visualise at a pixel scale. Maybe my traditional skills need more work but I would still like to hear how you tackle the subject if you have some time. It is quite hard to create realistic volumes at this level and things seem to work better if you go for simplicity in my eyes.
Cheers for reading,
Jason.
-
Quicksilva: What I do is pretty simple. In whatever I draw which I will then fully shade or color (meaning I don't do this for comic art, which is very high contrast and has different inking and shading tropes most of the time) after I have my general shapes and lineart pretty much there I faux-3d render it by hand. Meaning I actually connect tris from point to point and make the geometry more complex step by step.
I mocked something up very fast because I should actually be doing comics instead of tutorials, but a few images are worth more than just dry text. Let's take the below doodle:
(http://www.locustleaves.com/vol1.png)
Then I start doing this, just more detailed. Once you're used to it a bit it's pretty mindless. I might not go to as much detail for simpler bits where I know how I'll behave. The key point in all of this isn't to just do this every time, it's to *think* it even when you're not doing it. Abstraction of something we're used to think as a bunch of symbols to pure geometric planes. Less polygons are better than more polygons.
(http://www.locustleaves.com/vol2.png)
Then select just a few shades and group polygons that, according to the lightsource would be closer to the same shade than the next jump towards darker or brighter. This is pretty basic, they teach this at any fine art course (of course not this computery and much more proficiently, but I learned by myself pretty much so)
(http://www.locustleaves.com/vol3.png)
I won't keep going for the whole thing, just some explanations of how I handle basic volumes. Once all value range has been applied then according to whatever style I'm going for I might blend between triangles or I might not. In pixel art that's usually where I start dithering between clusters.
(http://www.locustleaves.com/vol4.png)
This is the later without the lineart. Obviously there needs to be more of a bridge between the higher values and the middle values, like 2-3 colors but I don't have the time to do this, I hope the concept is clear enough as it is.
(http://www.locustleaves.com/vol5.png)
-
Thanks Helm for the explanation and for taking the time to show it so clearly (complete with nice drawings, much appreciated :) ) I`ve seen people doing a similar technique before but you`ve just made it so much clearer. I must give this method a try. Thanks again. I have a feeling that it is going to be trickier than it looks though... Incidentally, do you do any 3D modelling that you base this technique on?
Jason.
-
no, not at all. Which is why also my faux-wireframe looks so bad in realistic 3d modelling terms.
The key thing for this approach to work is the next step after you have the wireframe. It's when you choose 2-3 tones and have to 'group' polys together by approximation and say 'these'll all be darkest shade, these middle shade, these highlight'. That's where the most intuitive jumps your brain will do are, but ONLY if the piece is solid on that level, when you go in and add more intermediate shades, will the piece retain its solidness.
-
Got one, not pixel specific though.
@ Ben2theEdge: what is negative space?
(I'm asking you specifically because from the search I did, it looks you actually mention it the most, but I'll gladly read anybody else's views too!)
I've been reminded of it about a piece in my PJ gallery (sorry to link to my own work, just convenient here) (wink @ ska!).
http://www.pixeljoint.com/pixelart/48570.htm (http://www.pixeljoint.com/pixelart/48570.htm)
(http://www.pixeljoint.com/files/icons/full/arbre_manupix_2.png)
I feel I got the point about composition in that piece (have I?), but the general notion is eluding me.
Is there more to it than "a nice empty space for good composition"?
-
I'll have a go at an answer, as my Art teacher's tried to beat it into my skull :P
Negative space is the detail-less bit around the tree, often the "negative" of the colours next to it. Wikipedia (http://en.wikipedia.org/wiki/Negative_space)
In that piece that ska linked to (http://pixeljoint.com/pixelart/27294.htm) the negative space plays a really big part, if the house were to just have a same-width border around the idea of the piece would not be conveyed. You've almost got that going on in your piece but it could be more drastic, something like this
(http://i48.tinypic.com/1ruaup.png)
THe eye immediately sees black in the sea of pale pink, then zooms further to the red and then is like "Oh cool, it's a bird"
It's just about balance of emptiness and detail to convey your message :)
-
Hey helm. It's been killing me. How do we define something like a base being 3/4 top down perspective? I've been trying and trying but I can't get it right.
Here are some of my previous tries
(http://i272.photobucket.com/albums/jj196/bgill31/Walkv3-1.gif)
I think I'm starting to get it with my latest
(http://i272.photobucket.com/albums/jj196/bgill31/tempgdsp.png)
I thought the shading would be more rounded if viewed from the top.
-
Thanks Jeremy!
Shame on me for not having searched wikipedia. :(
-
Hey helm. It's been killing me. How do we define something like a base being 3/4 top down perspective? I've been trying and trying but I can't get it right.
I'm not sure what you mean?
-
I need help showing that a sprite is in top down perspective. Like in gba rpgs like zelda.
-
I don't have any experience with top-down games but here's a few thoughts.
First of all, the first sprite is top-down enough.
Second, it comes down to context. Make some tiles and mockups and put the character on them with his shadow and whatnot and you'll see if it works or not. For these sorts of games I feel the main character having personality and being distinctive is much more important than perfect technical axonometric projection or whatever. Look at Phantasy Star IV for the Genesis. Those sprites are basically side-view, on a top-down overview and it works fine. I don't think the visual disparity is anything anyone got held up on.
Other users (Gil, Adarias, others) have much more experience with actually drawing these types of sprites, I have never made a sinlge top-down overworld sprite if my memory serves.
-
If it helps, I've taken a little goblin I've got modeled here and spun him around with a camera looking down at him from 60 degrees up. It might help you visualize it at least, even if you don't follow anything exactly, which I would suggest you do not do.
(http://www.gerla.us/paul/edits/GoblinSpin.gif)
(BTW, his proportions are close to correct, his feet and hands and head are a bit bigger than a normal human, and the arms a tad long.)
-
I use something aking to side view, but with enough tilt to give it that top down look. Basically, for mid-size sprites, it tends to look better if they're not too much top down. Pawige's render is an excellent example though, if you want the big inclination.
Right now, I use this (30°?):
(http://art.game-designer.org/pixelart/rotating.gif)
-
Has anyone noticed that with the RPG views, it doesn't seem to be correct. As in, most objects retain their front view, while the top of the object gets skewed more towards the camera. The front doesn't get skewed away. That may be why I think most sprites in this perspective are more frontal than they are top-down. Is it supposed to be like this? I'm guessing not, but it's a choice that is made to easily convey more important information.
(http://i102.photobucket.com/albums/m93/Dustin99/RPGviews.png)
-
I think a lot of it has to do with the artist's ability to use perspective. That, or it's just so ingrained in the genre that artists use it just because most everything before it has used it. There is the fact that it's a game and it's easier to have the information of entrances, exits, and player locations easy to see... But in most RPGs that doesn't hold water because there's not too much action when you're viewing things from this perspective.
It is still art, though, so it's subject to a lot of stylization in the first place.
The last time I did topdown rpg stuff with buildings I think I did the "correct perspective" (http://www.lolipopsicle.com/linky/inntree.gif) on both structures and characters, so take what you will from that!
-
Honestly, the 'right' perspective on the first example in my opinion looks horrible. It just seems wrong, even if it's not, that the whole back of the roof is missing -- very empty.
-
There's two kinds of perspective used in RPGs, which is something you have to keep in mind. Both are constructed differently. Novice artists use both mixed, don't know what they're doing, though.
First off, you have the perspective you show as "Correct". Orthographic projection, camera pointing 45° down. (Your measurements seem off, but I see you get the point)
There IS another one that's very commonly used and just as correct though. It's a variation of the cavalier perspective. Cavalier is where you take a normal front view (x & y are normal front view) and draw the z axis at a 45° angle out to the left or right to make it visible. The RPG variation draws the Z axis out to the top. The result is the "Traditional/Common RPG view" you reference. You also have to pick a foreshortening ratio for the Z axis (all values are valid, you just pick one and stick with it). Traditionally in RPGs, this value tends to be 1:2, meaning the lines running parallel to the z axis are drawn at 1/2 lenghth. Some go up to 1:1 though (which looks strangely elongated though).This is an equally valid oblique projection.
If my text isn't clear, I'll create some examples. The advantage of the second is that you can show the fronts of buildings normally, sprites are easier to construct, etc.
-
Aaand then there's the 'zelda perspective' where you just fudge everything but make sure to keep perspectives that express what is important about the environment.
-
Aaand then there's the 'zelda perspective' where you just fudge everything but make sure to keep perspectives that express what is important about the environment.
Well, it's certainly interesting, but Zelda: LttP shows that consistency is more important than "correct" perspective. You have to be a pretty good artist to eyeball it and get it right though, so I don't advise trying out styles like that when you're still learning the basics.
-
Thanks for the help on the old question, much to learn from here.
I've got a new one for, whoever.
Do you guys tend to draw stuff horizontally, vertically, or perfectly diagonally aligned with the canvas because it's easier to draw? Am I right in the assumption that sometimes things are easier to draw this way?
I drew an object that had for the most part, a line of symmetry extending from corner to corner of the canvas. I wanted to draw the same object but tilted slightly, so it was about halfway between aligning horizontally and diagonally, but found I had a much harder time doing it, because it was harder aligning the pixels the way I wanted them to.
Actually, this probably only applies to symmetrical things.
-
If I understand you correctly - yes, pixelers commonly use angles that the medium of pixels facilitates, such as 0°, 45° and 90°. You know how that works, you do pixels, I see ya!
Not following you on the symmetry part, though.