blendroid - Sketchbook

Current Thumbnail:


Day 1: Started out with the “cute elephant” tutorial on CG Cookie, but, never being satisfied with just doing as I’m told, I will be creating a stuffed Pikachu.

Got this sketched out in about 1 hour 30 minutes, including time spent looking at Google image refs of knitted animal dolls. Doing my best to keep the topology in line while emulating stitched cloth (will become more evident as I finish the details, hopefully tomorrow).


Created this by following the “Cute Elephant” tutorial on Blender Cookie. I created a simple scene based on my desktop (the actual top of my desk), just to try and learn cycles a bit more. I’m not sure if I know how to use lights correctly yet, and I regret having tried DOF, cause this took three-and-a-half hours to render and my texture on the curtains got blurred out. I’m sure there’s some sort of compositor magic I could have used to do all of the effects in post, but unless there’s a quick fix for improving this, I’m done messing around with it.

At 4K res and 1024 samples, this rendered in 3hr 30min:
https://d2ip58kv7n8yjd.cloudfront.net/p/assets/images/images/000/164/314/large/noah-summers-20130917-pokedoll.jpg?1407833967

https://d2ip58kv7n8yjd.cloudfront.net/p/assets/images/images/000/164/316/large/noah-summers-blendroid-pikachu-setup.jpg?1407833700

That’s amazing. Especially the fuzz.

This is a very nice render. Yes, there could be a lot of issues like DOF or pp, but anyway, it catches the eye which is still the most important. Congrats!

Thanks, guys! Doing my best to continue creating on impulse. I figured I’d try recreating one of the LED signboards from the subway:


By the way, is it considered poor etiquette to post hi-res png files? I don’t frequent the boards enough to know what the accepted image-posting practices are (and the forum rules don’t seem to address it).

Cute Pikachu. And the signboard is very nice! Congratulations! Next step would be animate the sign, haha.

I think it better to post low-res images, but you can provide a link to the hi-res. :slight_smile:

POKEMONS!
Pretty sweet!

Thanks guys!

ren811: I do look forward to learning how to animate an LED sign. I tried modeling it realistically, so all of the light bulbs are in there; I’d just need to figure out how to animate their material properties (I’ve never gotten very comfortable with IPO curves and whatnot).

Jumping on the bandwagon with a Slender Man model. Just created the base mesh tonight.


Nice work! Love the subway sign. :wink:

Finally sat down to do some sculpting. Really loving dynamic topology, but the workflow is still not quite natural to me, so I’m doing my best.

After I finish all of the symmetrical details, I’ll get to finishing the front of the coat.

Also, I don’t know how that dent happened in the coat tails. That’s not supposed to be there.


So, here’s what I’ve done lately. I scrapped the old Slender model and started anew. I realize the cloth folds are pretty unrealistic, and there’s so much room for improvement, but I’m pretty burnt out on modeling this (if anybody has any super suggestions, though, I might implement them before I sit down and retopo/texture him).


I pretty much have his textures thought out in my head, but I’ve been distracting myself with hard surface modeling for the last few weeks. Apparently, everything I ever thought I knew about texture baking was a lie (Thank you Polycount! And thank you xNormal!), so I’ve been relearning that. I set out to make a pretty physically accurate model of this common Korean payphone, as a game environment asset. I absolutely love the design; reminds me a little of Art Deco style.

I’m so done with looking at it, but I’ve learned so much by forcing myself to finish it. I think I’ll keep going till I’ve got it to the standard I originally wanted. My current stumbling block is quickly making convincing sticker designs, which apparently cover American payphones pretty densely. (The phone is Korean, but the asset is meant for a game set in the States.)

The model only has a normal, diffuse, and bump map, yet, which are all pretty haphazardly assembled. I have no workflow for this kind of thing yet, but I imagine that if I keep blindly working till completion, I’ll have a better idea of where to start next time.

The Pikachu model looks excellent, as well as your other posts. Nicely done.

Finished playing around with the payphone model. Made too many bad judgement calls when unwrapping it, and there’s no point helping it now. On to the next project.


I’m currently making the hi-res meshes to a few varieties of Korean fire hydrants. The first one (my favorite of the designs) is finished, and the others will be modeled mostly derivatively. I though the insets in the top would be super challenging to model, but they actually didn’t take too long, and the topology’s even very clean. It’s like after all these years, I’m actually improving or something.


I used the sIBL import script to light the scene (using the Heliport Golden Hour HDRI from the HDRlabs website). Not the best way to show off a model, without a proper studio lighting HDRI, I guess.

Brilliant texturing on the phone BlendRoid. Excellent stuff in this thread, keep it up :slight_smile:

That slender man looks awesome.

Thanks Sumo, Kalinaki. :slight_smile:

Here’s some progress on my fire hydrants. I think they’re completely modeled at this point, though I plan to redo the chains on all of them (make size more uniform, and apply rigid body deformations to all of them).

Like the payphone above, all four are real-world South Korean hydrants; I just took some photos and modeled them by eye, with some proportional adjustments here and there to keep their silhouettes interesting when viewed from a distance.



I originally planned to make low-poly versions and texture right away, but I’m on sort of a modeling stint for now. I’ll be coming back to these later.


Noticed that the afternoon lighting in my apartment building’s garage would make a pretty sweet HDRI, so I grabbed my Panosaurus (only ended up getting an LDR). The spherical panorama is stitched from 36 images at 18mm focal length (Sony Nex 5n default lens).

Being as indecisive as I am, the sun moved too quickly for me to get all three exposure sets. So I tried for an evenly-lit LDR, threw the images in to Hugin, clicked “align”, and it came out almost perfectly. There are stitching errors along the vertical extremities of the image, and Hugin tells me that the fit is about as bad as it gets, but I couldn’t have expected better for zero manual input.

I plan to venture out again tomorrow, better prepared this time. Tonight, I will experiment with the number of manual control points Hugin would like from me in order to get a better stitch.

Any advice from Hugin users would be appreciated.

Giving up on wrestling with Hugin for now. My current focus is this KIA Soul model. Never modeled a car before, but since I’m focusing on hard surface (not to mention how many times I see a car and think “I’d love to try making that in Blender!”), I figure it’s time to learn.

Time to go to bed, though. Have to go to work in the morning. Adulthood… “Bah! Humbug!”



Messing around with rendering style. This just uses a matcap material and ambient occlusion, and a plane set to “only shadows”.

Edit: And toying with image hosting on ArtStation: