HABBO ENLARGEMENT - Marlec / GER:Donay

[GER:Donay]

Hey! :)

Haven't been here for a long time and also I haven't done much new things for you to show...
But yesterday I finished my newest enlargement. I'm not so proud of all parts of it, the head is from an older enlargement and the feets are looking a little bit strange :/

All in all I'm quite happy with that enlargement, hope you like it too.

[Samantha]

Looks ok, although I find the shading a little weird/lack of in some cases. The hands are the downfall for me the longer finger looks out of place.

[A]

Looks really good, well done and a nice attempt!

[oftenMiceElf]

Looks nice Donay. I love seeing these kinds of projects...


Indulge me for a second. The following isn't really directed solely at you and is more stuff I've thought about over the years regarding this sort of work in general. I do, however, hope that you (or anyone who's interested in this sort of art) can glean something useful from it though- especially since you seem to enjoy this style.

I've always found "enlargements" to be a fascinating concept. Maybe that sounds like a silly statement but when you think about what each pixel represents in terms of resolution it is pretty intriguing. If pixel art is basically a low-definition (resolution) representation of something at a higher definition it means that each pixel is approximating a whole lot of information to make it look like the thing it's representing. If you take an image and reduce enough you'll get a single pixel that's a median of all the visual info from the larger.



Ok, that's pretty obvious and maybe not all that interesting (though, to practice it is a nice challenge) but it setups up what's going on in these enlargements; which is to add back lost information from some prior (possibly never seen) reduction.

A great example of this process is what Capcom did with the HD remix of StreetFighter 2

Show Hide

You can see the higher def version isn't just a cleaned up 1:1 scaled graphic. It's a re-imagination of the original and pretty much every bit of detail that can be added back, keeping within their style, is. It's a bit easier for Capcom because they have a long history of what their characters look like outside of their low definition pixel interpretations. Habbo on the other hand doesn't have that. True, we do have some officially released "larger" images of Habbos (like this for example,

Show Hide

and they're a little creepy) but there isn't a clear line delineating "Habbos actually look like this". And for that reason I find user created re-imaginations of the avatar particularly intriguing because there's not really anything to fall back on.

If someone on my team were to tackle something like this here's a small list of things I'd point out before they started or what I'd be looking at:

1) If you're looking for a straight 1:1 larger graphic that's not really adding detail it's probably easier to use one of the many pixel scaling algorithm out there. Lots are displayed here: https://unindented.org/articles/pixe...arison-part-1/

You can also get a rough approximation in photoshop by scaling using Nearest Neighbor and then applying a Median Noise Filter. You'll have some clean up to do though.


2) Our avatar can be pretty odd. I won't lie. Use the larger pixel dimensions to fix the odd stuff that's present in the lower definition asset. Especially the hands and the odd overlapping that happens sometimes. If it helps at all, when I draw an avatar (pixel or otherwise), I'm thinking of a combo between Calvin+Hobbes and Speed Racer so there's a nice balance between soft/round and sharp. It also helps pull the face away from the creepy side of things (seriously, if you watch that youtube link you'll know what I mean).
random examples:
(first habbo is about 300%)

(middle habbos are about 150%


3) Draw/render hair in clumps or ribbons vs strings or lines. It's easier to visualize that way and closer to how hair is.



4) Don't be afraid to completely drop the original avatar shading. Especially on clothing. Habbo uses all greyscale assets that get colored via code. It's often lacking a real sense of light information for that reason.

Putting that all together (more or less), a small edit:

Show Hide

[Kyle]

NMe's edit above shows that though your art is really something great in its own right, the heavy dithering on every aspect of the work has forced it to lean far more towards pixel-by-pixel and resulting slightly flat enlargement than what it has the potential to be - a scaled up but partly reimagined habbo. It's so pristine, you truly have been assisted by the best!


Sent from my iPhone using Tapatalk

[myles]

just use the zoom in function on habbo and ur done?

[Drew]

@oftenMiceElf; Brilliant! I absolutely enjoyed the Summer School Q&A so this is just wonderful. I hope you'll post more in the future!

Great job though, Donay, always lovely to see your work

[dbgtz]

Looks nice Donay. I love seeing these kinds of projects...


Indulge me for a second. The following isn't really directed solely at you and is more stuff I've thought about over the years regarding this sort of work in general. I do, however, hope that you (or anyone who's interested in this sort of art) can glean something useful from it though- especially since you seem to enjoy this style.

I've always found "enlargements" to be a fascinating concept. Maybe that sounds like a silly statement but when you think about what each pixel represents in terms of resolution it is pretty intriguing. If pixel art is basically a low-definition (resolution) representation of something at a higher definition it means that each pixel is approximating a whole lot of information to make it look like the thing it's representing. If you take an image and reduce enough you'll get a single pixel that's a median of all the visual info from the larger.



Ok, that's pretty obvious and maybe not all that interesting (though, to practice it is a nice challenge) but it setups up what's going on in these enlargements; which is to add back lost information from some prior (possibly never seen) reduction.

A great example of this process is what Capcom did with the HD remix of StreetFighter 2

Show Hide

You can see the higher def version isn't just a cleaned up 1:1 scaled graphic. It's a re-imagination of the original and pretty much every bit of detail that can be added back, keeping within their style, is. It's a bit easier for Capcom because they have a long history of what their characters look like outside of their low definition pixel interpretations. Habbo on the other hand doesn't have that. True, we do have some officially released "larger" images of Habbos (like this for example,

Show Hide

and they're a little creepy) but there isn't a clear line delineating "Habbos actually look like this". And for that reason I find user created re-imaginations of the avatar particularly intriguing because there's not really anything to fall back on.

If someone on my team were to tackle something like this here's a small list of things I'd point out before they started or what I'd be looking at:

1) If you're looking for a straight 1:1 larger graphic that's not really adding detail it's probably easier to use one of the many pixel scaling algorithm out there. Lots are displayed here: https://unindented.org/articles/pixe...arison-part-1/

You can also get a rough approximation in photoshop by scaling using Nearest Neighbor and then applying a Median Noise Filter. You'll have some clean up to do though.


2) Our avatar can be pretty odd. I won't lie. Use the larger pixel dimensions to fix the odd stuff that's present in the lower definition asset. Especially the hands and the odd overlapping that happens sometimes. If it helps at all, when I draw an avatar (pixel or otherwise), I'm thinking of a combo between Calvin+Hobbes and Speed Racer so there's a nice balance between soft/round and sharp. It also helps pull the face away from the creepy side of things (seriously, if you watch that youtube link you'll know what I mean).
random examples:
(first habbo is about 300%)

(middle habbos are about 150%


3) Draw/render hair in clumps or ribbons vs strings or lines. It's easier to visualize that way and closer to how hair is.



4) Don't be afraid to completely drop the original avatar shading. Especially on clothing. Habbo uses all greyscale assets that get colored via code. It's often lacking a real sense of light information for that reason.

Putting that all together (more or less), a small edit:

Show Hide

This was an absolutely fantastic bit of CC, though I shouldn't be surprised. Out of curiosity, do you browse graphics sections on fansite forums (be it Habbox or any other fansite) often?

[oftenMiceElf]

Out of curiosity, do you browse graphics sections on fansite forums (be it Habbox or any other fansite) often?

Certainly do. I enjoy seeing learning and the processes creatives go through- be that wips or general skill improvement over time. Also makes me feel a bit more aligned with what the community is interested in.


@GERonay I want to make it clear that I do really like your work and I hope that you don't take what I posted as discouraging (or thread hijacking) at all. I'm just nudging because you seem to enjoy enlargments :p

[GER:Donay]

Certainly do. I enjoy seeing learning and the processes creatives go through- be that wips or general skill improvement over time. Also makes me feel a bit more aligned with what the community is interested in.


@GERonay I want to make it clear that I do really like your work and I hope that you don't take what I posted as discouraging (or thread hijacking) at all. I'm just nudging because you seem to enjoy enlargments :p

Never thought about it in a bad way! I really like criticism in all way to improve my art.
And thanks