Lostcast 166: Sprite Artless

  • LDG

  • Patron

    Thanks for the MCiB shout-out, and thanks for coming on my show! It was a lot of fun.

  • Tiger Hat

    NaN is evil. Not only is Matt right, and an error should be thrown because it presents expected behaviour, but it’s also internally inconsistent;

    NaN supposedly “Not a Number” but javascript recognises it as a number - gar! This is the sort of nonsense that makes me dislike javascript and PHP.

    @richtaur - there is a huge improvement in art-style between AWL and AWL2 - most notably, there’s a lot more personality in your newer work - I’d love to see you work on your animation skills rather than another style, but that’s my gig I’m the sort of loser who has watched the matrix films frame by frame :)

    I feel like AWL2 is quite removed from the “binding of isaaclike” feel - there’s much more close combat, and shenanigans, and honestly, whilst the art is good, it’s all very ugly, I’d much rather a Raga toy than anything from it.

  • Jammer

    Great show! For the record I do feel your aesthetic is unique, but I hear what you’re saying. I eagerly await to hear whatever voice your art finds. Also, Geoff, I know I haven’t dont even one portrait on here like I said I would, but I look forward to you and more people posting in the art improvment thread!

    You guys do great work, and thanks for the great cast!

  • Hi guys great podcast as always. I just wanted to say that the way to get performance improvments(at least I think) with a texture atlas, is to reduce the draw calls themselves. (what I mean is you batch all draw calls for one rendering update together and then call draw once and specify where what is drawn all together at the same time), but I am wearing a tigerhat right now, especially about how chrome handles stuff.

  • Tiger Hat

    I agree with @Vox - I think your artstyle is pretty unique(-ly awesome!).

    Going further, it’s not the reason I’d compare AWL1 to BoI. Maybe an unpopular opinion here, but I think AWL1’s gameplay is superficially similar to how a subset of BoI runs go. Based on the Let’s Plays I’ve seen of AWL2Soul Thief, it looks like you guys are enacting a huge amount of gameplay changes that seem like they might further the gap.

    ANYWAYS, thanks for the podcast and I’m looking forward to getting in on some of that Soul Thief Early Access action!

  • LDG

    Thanks for the thoughts and encouragement you guys :) Great fuel for next week’s podcast.

  • Tiger Hat

    Great cast as usual! I feel like you are still missing something with atlases, though perhaps the refactor work needed wouldn’t outweigh the benefits. I think that because you kinda half-assed the implementation (at least thats what it sounded like :) ) that perhaps that’s where the benefits hide.

    I dunno if you’ve heard of TexturePacker but it’s a really awesome tool, it will reduce the redundant sprites automatically, trim images, etc.

  • Tiger Hat

    @PartTimeIndie said:

    I just wanted to say that the way to get performance improvments(at least I think) with a texture atlas, is to reduce the draw calls themselves.

    This makes sense - and has the same theory as the raindrop scaling thing; is the mob list just an array of objects or is it sorted by type?

    If you draw all the inkwells first, and then move on to the inkspots, you may actually see some benefit, especially if your sprite renderer tests to see if a particular sprite sheet is already in memory. I don’t know how much control you get in JS over this sort of thing though.

    It’s what Unity3d does internally.

  • Jammer

    Just listening now, I find it interesting what you said Geoff about not using scale, rotate, save & restore. Just using the transform function instead. I know save & restore can be heavy, and you should keep those minimal, but have you found using transform vs scale to be much more performant? I wonder if scale & rotate are still batched anyways. Kind of lazily set data, and then actually operate when you call drawImage or fillRect.

Log in to reply