Mech

[odedrt9]

Hey 

I tried to do some kind of mech thingy:



I'm yet again not really confident about the shading as I suck at shading things
I think the overall shape is fine.

Hmm that's it for now, I know there isn't much to say because it's a small sprite but I'd like to hear your opinions and critics.

Edit:
So I searched for a cylinder like metal plate shadings and I came up with these:
And

I used Starbound's space train for reference as you can see the top of the train is cylinder like.

[0xDB]

I would suggest to use primary (as in real world photos or life drawing) references instead of referencing from another artists rendition of anything because if you do the latter, you end up just copying their style and their mistakes (if they made any).

[odedrt9]

Hmm I understand what you mean although I did look for real references, I couldn't find any good ones...

But what do you think?
I'm still not really satisfied with the result. I don't know why exactly but I feel like something missing..

[0xDB]

What do you think is missing? If you're not satisfied the result, keep experimenting and changing it until you're satisfied.

If you don't know what's missing, you could start by writing down what this Mech's purpose is going to be in the world in which it operates. Maybe then, the things it is missing will come to you.

[astraldata]

 

In art -- as far as dealing with the visual construction of things -- there's a rule that's very important to remember -- "form follows function."

If you've no idea the function of this machine, there will always be something 'off' in your rendering of it -- that lack of purpose (or 'function') will always be visible to both you and your viewers subconsciously. That alone could explain what is 'off' sometimes when you feel like you've rendered something correctly.

In the case of your drawings here, what is 'off' is indeed your rendering -- but your rendering of the 'form' doesn't follow the functions of shape and light, and therefore it doesn't indicate the target form very well.

Nothing wrong with Dennis's suggestion if you need more direction in the design, but I do want to add that 'just experimenting with it till it looks right' just doesn't *ever* help you learn **why** something's 'wrong' in most cases. Rather than prevent something similar from looking wrong in the future, it only helps you learn how to make something like it not look *as* wrong. You could do that for years and never grow once in your understanding of the basics of its construction or even art in general if you follow this path (I know this because I went down that path for many years!). That 'experimental' line of thinking is only completely applicable in the concept or prototyping phase of a design -- but if you can't render things, you need to come to terms with *that* issue first and then practice practice practice (hopefully that's what you're doing here!)

Regarding pixel art, references will help you with construction and *maybe* general color schemes -- but rarely with the underlying rules of rendering (this is why just drawing lines of varying shades in an attempt to mimic your reference isn't helping you here).

It takes already having a grasp of representing the logical rules of shape and light in relation to colors to understand why another artist chose to use the colors and placement he did to render a form. Of course, the closer to abstraction that artist's rendering is, the less this matters -- but it *does* always matter to some extent -- no matter the style of rendering chosen -- even in straight black and white you need to understand light and shape and how this affects the color of materials in a scene in order to understand what values (or even just the type of line) to use to represent these things effectively.

------------------------
 C+C  +  My Process
------------------------

Now, I want to go over the differences between both our images to highlight the reason for my changes:

1) Your image's form wasn't very clear -- I gave a more clear light source (top-left) and spread the metallic sheen across the cylindrical forms more in the center, which made the light fall off more to the right (which you can tell by my increased usage of the darker value as it graduates to the right).

2) Your image seemed too 'flat' and 2 dimensional -- I added additional contrast to your colors to make the closest and more lit portions of the shape 'pop' forward more with the increase in brightness values. I also made the graduations in shadow values more distant from one another (increasing the contrast in those as well).

3) The 'thin' cords of the 'helmet' (I assume is what it is) weren't rendered effectively -- 1px lines shouldn't ever be used to define anything on their own, so I added shadows in order to allow them to both have more form *and* be visible on both light and dark-colored backgrounds (thus their high contrast). I also separated them value-wise from the rest of the helmet so that they could stand on their own and add interest to the design as a whole.

4) Colors were very dull and boring (straight grays as your other colors didn't help this fact either) and you used 100% saturation in your reds to try to offset that dullness which (unintentionally) burned the eyes in the process 
-- First, I ensured proper value contrast **--> independent of color or hue <--** in your bright-to-dark values (you can grayscale the image to check your color contrasts until you develop a sense for this on your own), thus making your forms more shapely and ensuring the separation of the portions of your image that are meant to be separate, and only then did I tweak the hues.
-- However, although I knew that I wanted a red hue for the base color, I needed to decide whether I wanted it the scene lighting to be warm or cool and limit myself to only those hues on the spectrum when rendering the colors.
-- At that point, I selected the highlight hue and shifted it toward the yellow color since I decided I wanted warm, and the rest of the colors in-between, I shifted toward the red I wanted, and then toward blue for the shadow colors (while not getting *too* blue because my colors are meant to remain *warm* and therefore can't stray too far toward actual blue). As you can see, the image looks a lot more vibrant due to this hueshifting, but that wasn't the end of it
-- I increased the saturation of the gray-highlight and put it as a blue-green and just let the purest grays of the deeper shadows pull and absorb their warm 'color' from the red pixels around them. Since gray draws in color around it like a sponge, it also dulls those colors around it if it's used too much.

[0xDB]

No offense to Dennis, but 'just experimenting with it till it looks right' just doesn't *ever* help you learn *why* something's 'wrong' in many cases -- rather than prevent it from looking wrong, it only helps how to make it not look *as* wrong. You could do that for years and never grow once in your understanding of the basics of art (I know this because I went down that path for many years!). That way of thinking is only completely applicable in the concept or prototyping phase of a design -- but if you can't render things, you need to come to terms with *that* issue first and then practice practice practice (hopefully that's what you're doing here!)

None taken. What you write shows me, I should have been much more verbose, so you (or anyone) would not have interpreted what I wrote in the way you seem to have.

I was not suggesting experimenting and changing it "mindlessly" (as you seemed to read it) like randomly changing shades or individual pixels until it would "mysteriously end up look right".

I was suggesting to mutate the whole thing, changing the structure around without even thinking about the finalized rendering. That's why I suggested thinking about purpose and the world as that might reveal if it needs additional features or different features, like guns or thrusters or antennas or whatever it may need for the world it "lives" in and for doing whatever it does there.

That experimentation and changing process is what you refer to as prototyping in the design phase. In that phase, it is in my opinion unnecessary to already know how to render things accurately for the final version. You can just use basic and simple elements like blocks, spheres, cylinders, etc. (or even simpler 2D elements, sticks and boxes).

Once you have the design down you can proceed to rendering. If you already know how, that's a huge bonus but if you don't, you don't need to practice that before you can make a solid design. Also, before you know "what" you want to render, you can not know "how" to render it and would not know what references to look for. So there's only a small disagreement about sequence of work- or rather learn-flow here.

A design can be as simple as a 2D stickfigure scribble combined with a coherent description of what it's supposed to be/do. It does not require deep rendering skills to make a basic design.

But well... I think we have a similar point of view on these things, I just can't seem to find an end to writing and felt an urge to clarify but I'll cut myself off here as I don't have anything to add at the moment.

[Amorphous]

My two cents.

[...]'just experimenting with it till it looks right' just doesn't *ever* help you learn *why* something's 'wrong' in many cases -- rather than prevent it from looking wrong, it only helps how to make it not look *as* wrong. You could do that for years and never grow once in your understanding of the basics of art (I know this because I went down that path for many years!)[...]

Speaking as someone who hadn't received any type of formal art education until high school started a month ago, I strongly disagree.   Over the past two and a half years since I started drawing, "just experimenting [...] till it looks right" has gotten me from pixelling like this to as so. Obviously the latter is messy and unfinished (sorry about that ), but I think you get my point. I don't see why experimenting without a definite purpose is somehow a brick wall to artistic growth. For me, screwing around with a project until it improves has always been the precursor to figuring out why the previous version looked worse and the new better. Like Dennis said, "In that phase, it is in my opinion unnecessary to already know how to render things accurately for the final version."
Maybe our methods of learning are just drastically different, but tbh I'm kind of puzzled by your thought process, astraldata. Thoughts?

Anyway, as for the mech, visual clarity is a big problem. When I first clicked on this thread, I thought it was a helmet visor. I made two edits, one with a totally different style, and another that might actually be helpful for your purposes, odedrt9. Astraldata's edit has good points, but I think bringing out the highlights so much actually didn't add depth, because I can't tell which portions are higher than the others.

Hopefully something in there helps you get the mech up to your level of satisfaction!

[astraldata]

@Dennis
Ah, yeah, you're right -- it looks like I did misread you there. I had a feeling you might have been hinting at what you clarified, but I felt my point should be made anyway (just in case someone reading this was/is going down the path I had been before I started learning art seriously). I wasted a lot of time that could have been spent learning the construction of various things and building up a mental encyclopedia of things to reference that I could use in any design -- regardless of how that design is eventually rendered -- which would have made the 'designing' process you mentioned much easier for me in the long-run.

That free-form method of 'designing' is wonderful for coming up with new ideas in my experience, and it really isn't talked-about enough. But time spent learning how to construct and render many different things makes this process much easier (thus the reason why I focused so hard on emphasizing the importance rendering!)

I could never knock the prototyping/mutating phase as the first step in a polished design -- but that 'polish' generally needs to be there at some point if your design is meant to define an end product -- especially if you're not the one executing the 'product' of that design. That is really my only 'bias' in the issue of 'full-rendering to convey a design' versus 'using a 2D stickfigure with a coherent description of what it is/does to convey a design' that you mentioned. It always comes down to function -- is the design there to function as a general *reminder* for you, or is it there to function as a specific *definition* for someone else (i.e. in terms of style/colors/detail-level/etc)? The level of polish/rendering of the idea that is needed to call that idea a 'design' can vary greatly depending on the design's ultimate purpose in relation to the end product. Until that 'idea' is fleshed-out/rendered/etc. to an acceptable level for its use, it is still in the idea/prototyping phase and probably should be regarded as a set of 'ideas' rather than an overall 'design'. These terms really shouldn't be used interchangeably.

Really, my main point here is that I feel it's important to make the distinction between what part of the process is the 'idea-phase', what constitutes those collection of ideas as a 'design', and at what point does the 'design' become the end 'product'.

My reason to make this distinction is that you can work on something for 20 years and never accomplish a satisfying 'design' if you don't have the criteria to make this distinction. This applies to more than just art. If you have the slightest hint at perfectionist tendencies, making your criteria earlier as to what each phase of "done" means, gives you an edge over others who can't/don't make this distinction. This will not only make you a better artist, but a better businessman as well -- (whether or not any of that is what you're going for. ^_^;) -- because it will make you more efficient since it opens your eyes to what's truly important in your end product -- even when, at the time, the end 'product' is the 'design' itself!

Just food for thought to anyone who's interested...


-----------------------


@Amorphous
I just want to first address the point about my highlights not adding form to the image because of it being difficult for you 'to tell what portions are higher than others' -- I agree that it didn't add as much depth as it could have -- but I also wanted to avoid modifying the image as much as possible so that I could give the OP an idea of how to improve the cylindrical form he said he was going for. I didn't add any further colors or greatly modify the ones he had -- this was simply a quick edit for clarity, not for showing off my skill to make forms. Your shiny highlights were a good touch on your mech, but I took it as a head-on view of a helmet/visor (thinking along the lines of Super Metroid or oldschool Transformers/Robocop) and I spent very little time/effort on making it 3d and giving it that extra flair. For the straight-on viewpoint and time it took me however, I think I did pretty good in adding some preliminary depth without changing the image too much.

As for your other issue regarding 'experimenting till it looks right' as being a bad way of learning -- I learned exactly the way you did at first. I'm not condemning it -- I'm just saying there's a better way.

I apologize if the way I put it is offensive, or if I came across as condescending or something. I really just feel that learning how to construct and render things at first is a better use of time when you're just starting to grasp art in general, because it gives you a mental database of references you can use at any time, no matter what style of rendering you use (pixels or otherwise). The 'experimenting' I'm saying is 'bad' is just tossing around the same old symbolic representations of what you *think* something looks like (or should look like) and then looking at it to try to understand why that isn't 'right' -- without any outside reference or study (most times, this is out of laziness). If you're not doing that, and you've *got* references you're studying closely (and *not* just trying to mimic them), then you are definitely learning.

The OP's issue with the cylindrical shape and its lighting came from trying to mimic a reference he did not understand, that was too different from what he was going for (and of which I'm still unsure of what he was going for with the cylindrical thing -- your edit is probably closer than mine as to his intention), and this caused him issues because he didn't understand the underlying principles of light or materials enough to render them in the style he was trying to achieve, that he found vaguely in the reference he chose. Had he studied multiple references of this style of metallic rendering (and 'experimented' with pixels while doing so), he would have become a bit more keen about how to render metallic things in pixels. Unfortunately, with this method alone, the issue still remains that it might still be a bit difficult to translate this pixel skill, to HD for example, without first understanding why each pixel goes where it does, and just what that dot or cluster means if it were able to be represented at an infinite resolution. If he doesn't understand why I used the columns of light, for example, to render a metallic cylindrical form with an upper-left light-source, all the colors and resolution in the world wont help him to light a rounded metal form coming off that column with those colors if he has no reference -- making it a very difficult task to come up with stuff from his imagination (like his original image) and render it.

Throughout the course of writing this, I feel like I've been able to gather my thoughts in an effort to formulate my words a bit better.

So, in summary, this is how I see it:

As said before, I don't condemn the act of experimentation without a definite purpose, nor do I consider it a hindrance to growth -- I only feel that studying the underlying construction of things (while experimenting with them or their construction) in an effort to help yourself to develop a mental encyclopedia of rendering methods and construction references that work for you (again, at the same time as your experimentation) is a more efficient way to learn art (pixel or otherwise) while at the same time toying around with it without any definite purpose. Purposefully studying how to construct and render things will help you achieve this encyclopedia of construction and reference more quickly, and gives you some extra toys to play with and combine when you're experimenting -- helping you to learn underlying rules and masterful tricks exponentially faster.

Hopefully you'll feel we see eye-to-eye now. In my opinion, it doesn't seem our values are really that different, but I guess I just put things oddly sometimes. For that, I apologize. I sometimes don't see how other people will take what I'm saying, so I have an unusually difficult time formulating my words properly. Hopefully this clears it up a bit though.

So just for lolz --

[0xDB]

The level of polish/rendering of the idea that is needed to call that idea a 'design' can vary greatly depending on the design's ultimate purpose in relation to the end product. Until that 'idea' is fleshed-out/rendered/etc. to an acceptable level for its use, it is still in the idea/prototyping phase and probably should be regarded as a set of 'ideas' rather than an overall 'design'. These terms really shouldn't be used interchangeably.

Agreed. Maybe we could start some sort of Glossary in the the Resources section where we work out some definitions of terms to convey these and other concepts with less words in the future and to prevent misunderstanding.

I sometimes don't see how other people will take what I'm saying, so I have an unusually difficult time formulating my words properly.

I think many people share those difficulties, myself included (though I like to think I've become better at it recently).

One thing, and this is meant to be constructive, you could try is to use less words (you will save time for yourself and for your audience/students that way). Keep it short and to the point in the beginning.

People who are receptive for details and who are willing to study or who would like to read more of your thoughts on a topic will then proceed and ask you to clarify things which are unclear to them.

I personally think it's a waste of time to pre-emptively try to explain and answer to each and every potential reaction/reception and to lecture on any topic before it is clear there is interest for details. Also, people tend to feel overwhelmed if you fire all of your information guns at once.

The gist of what I'm saying(and that's general advice directed at myself as well) is: Try to not predict/control/suggest how people will receive what you write or say. Allow them to take that responsibility for themselves. You might feel that communication suddenly becomes a lot easier that way. The more you try to formulate something so that it could not possibly be misunderstood, the more complicated it will look to the reader and more information has a tendency to lead to more potential sources of misunderstanding(depending on both senders and receivers level of knowledge), not less.

Like Dennis said, "In that phase, it is in my opinion unnecessary to already know how to render things accurately for the final version."
Maybe our methods of learning are just drastically different, but tbh I'm kind of puzzled by your thought process, astraldata. Thoughts?

I actually do not think our ways of thinking differ that much or even at all. Could just be lingo collision: same terms but with different ideas in the back of our heads. Splitting hairs probably does not add anything of value to this exchange of thoughts. Both ways/sequences of doing stuff are valid. I think neither should be taken as gospel and followed rigidly. Be open and free about experimentation and jump back and forth between studies for building mental libraries of things and open experimentation and freestyle scribbling. Everything has its' use. As long as you keep practicing both, you can't really do it "wrong".

[astraldata]

Well put, Dennis.

You really hit the nail on the head about me trying to answer every possible question before I hand it over to the other side. I too saw this as a potential waste of time, but I always felt that it's best to give too much information than too little to someone you're trying to help to learn, although I'm really starting to find that spouting out a reference book at someone doesn't really work well in practice either. ;P I'm really poor at figuring out just how much is too little and how much is too much though, especially on the internet where you can't see the other person's expression reflecting boredom or interest.

Also, you have a valid point about saying too much tending to cause confusion and, potentially, conflict. Apparently I caused that in this thread. Gotta work on that asap. I like most everyone on this board and hate to accidentally stir up anything with anyone here. Crossed wires really seems to be all too common with too many words for some reason.

That said, I think the idea of a glossary with definitions of common terminology is a great one. We've got some pixel-art specific terms setup, but since many people on Pixelation do pixel art seriously rather than just as a hobby, and many of these serious projects are game projects, I really think a general creative set of terminology that includes game-developer lingo might be a great addition to the common-knowledge-base.

A final point of yours I want to second is that of not being able to do something the "wrong" way. Granted, we may feel passionate about our own 'way' of doing something that helped us get ahead in our art, and we might prefer to offer that to others, but, it's best for others to do what works for them in the end. All I can offer is advice on a way that I felt was more efficient for me, and, theoretically, might be more efficient for others too, but, in the end, it's just another way of doing something that may or may not work for you specifically. Some stuff may still apply to your specific goal(s), so take of it what you will. I simply want to help others with pixel art. The mistakes and the advice given by pros (and regular users alike) helped me to learn pixel-art techniques way back in the day, so I'm just giving back now what I was given in the level of skill I have now. To help others succeed in something I excel at is what makes me happy.