|
Mata posted:I would suggest spawning the snowflakes with some horizontal movement to the left as well.. When the snow moves straight down the bird feels pretty stationary, with the text being in motion.. I noticed this as well, but I think it may be useful for artistic reasons in this case for the bird to 'feel stationary' despite navigating obstacles. Real nice, clockwork.
|
#
?
Jan 18, 2015 04:39
|
|
|
|
| # ? Jun 10, 2024 12:36 |
|
|
Around lunch time: At quitting time:  ... though now I'm back in the office for a third shift, since everyone else went out for the night. That's my plot/clue flow, which I'm now throwing back into the doc, and then I'll write it and WOO DONE FINALLY. (it's... kind?... of a screenshot - designer work screenshot)
|
|
#
?
Jan 18, 2015 05:20
|
|
|
Cryolite posted:I think he had a kid. Oh... drat. Well, I hope it's not another 15 years or so until he gets back into that. He should just sell the engine as it is, because god drat. Shalinor posted:Around lunch time: I really hope that we can get a white board towards the end of the project, because all I have is a thick-rear end notebook of grid paper. It would really help! I actually do have a couple pages in there where I rearranged stickies that are rooms in the game. LP0 ON FIRE fucked around with this message at 05:34 on Jan 18, 2015 |
|
#
?
Jan 18, 2015 05:30
|
|
|
 Move generation in what will become a Scrabble-type game cheat/AI. The GIF is kinda slow, it works quicker than shown here. Currently the scoring is just a random number, so that's up next. Update: Now it's completely working, here's 4 AI playing against each other:  Next up is designing/creating the web interface so a human can actually interact with it. ArcticZombie fucked around with this message at 23:42 on Jan 29, 2015 |
|
#
?
Jan 19, 2015 00:56
|
|
|
TZer0 posted:
That's super cool! If I ever get up the ambition to make a game (won't) this is the kind of game I'd make.
|
|
#
?
Jan 19, 2015 20:18
|
|
|
Here is a cross-post from my project.log thread. I examined my BeagleSNES to-do list this past weekend and address several of my outstanding development items. I ended up accomplishing far more than I had hoped to in only a few hours. Five minutes of awesome: https://www.youtube.com/watch?v=Gg8kcdPKgSc  Among other things: 1. I broke BeagleSNES out from the main GUI process. Now, BeagleSNES launches from the main GUI as a separate process. Virtual terminal support is incompatible with the framebuffer driver (i.e. I can't chvt to a different terminal and open /dev/fb0 again), so I had to completely shut down the video subsystem and EGL context in the main GUI, launch the SNES emulator, and then reinit the video and EGL on return. The SNES emulator has a non-GPL license, so I can't build it directly into the main binary with the rest of the GPL code.  Still, you need to respect the software licenses of the various emulators, and I bend over backwards to try and do so.2. All four platform emulators (GBA, GBC, NES, SNES) all pull the same keymap config for the gamepads from the same place. The short version is that I intercept joystick events in each emulator's event loop, look them up in a map, and then generate keyboard events that the emulators actually act on in their event loops. One XML config file for all of the emulators! 3. I fully integrated the EGL support into the SNES emulator. It actually performs much better because I am bypassing the software pixel doubler blitter and just dumping pixel data raw into the texture. Doing the same thing also sped up the GBA emulator a lot. 4. I fully integrated the pause/save state dialog into all of the emulators. If you knew how differently each of the emulators implements save state support, well... you'd have an idea of how messy a job this ended up being. 5. There is now a major reduction in save state and SRAM writing time for all of the emulators. Before, I was naively calling sync() to make sure the files were written to disk. This is super slow. Now, I fsync() the file and the directory that the file is written into. This is much faster, and I owe this tip to the Linux System Programming book I mentioned before. Even better, I think I can set a flag when the directory has been fsync()'d to then use only fdatasync() on the saved file for even more speed. My 0.7 release was pie-in-the-sky for a while, but I now realize that I am far closer to it than I thought.
|
|
#
?
Jan 19, 2015 22:32
|
|
|
 Made a dumb HTML5-based music player. The original goal was to have it on Windows Phone as a replacement for my dead iPod Classic, but the Windows Phone SDK is one of the worst things I've ever seen. I reverse engineered the Foobar2000 file format so it could read from my standard music database directly. That's why the track numbers aren't in order.
|
|
#
?
Jan 20, 2015 04:09
|
|
|
I just submitted BeagleSNES over at the HackADay tip line. Again. I just hope it won't be passed by. Again. Third time is the charm, right? I tried to give them all of the bits that they need, so let's just hope they check it out. Anyway, the beagleboard.org folks were kind enough to offer to send a GamingCape my way to get it working with BeagleSNES:  The problem is that the cape is short a few buttons. Near as I can tell, it is only the d-pad contacts for the joystick and the two thumb buttons. Even for the Gameboy and NES, you need the start and select buttons in addition to the B and A buttons. Anyone have any ideas on how I could handle mapping the controls properly? hendersa fucked around with this message at 22:50 on Jan 20, 2015 |
|
#
?
Jan 20, 2015 21:50
|
|
|
hendersa posted:Anyone have any ideas on how I could handle mapping the controls properly? It has an accelerometer so you could maybe use shake or something like that for one command?
|
|
#
?
Jan 20, 2015 21:55
|
|
|
hendersa posted:The problem is that the cape is short a few buttons. Near as I can tell, it is only the d-pad contacts for the joystick and the two thumb buttons. Even for the Gameboy and SNES, you need the start and select buttons in addition to the B and A buttons. Anyone have any ideas on how I could handle mapping the controls properly? Probably a terrible idea but could you somehow overload the analog joystick? The dpad is only four inputs, could you assign those plus the two+ missing buttons to various ranges of the joystick?
|
|
#
?
Jan 20, 2015 22:49
|
|
|
Trabisnikof posted:It has an accelerometer so you could maybe use shake or something like that for one command? pokeyman posted:Probably a terrible idea but could you somehow overload the analog joystick? The dpad is only four inputs, could you assign those plus the two+ missing buttons to various ranges of the joystick?
|
|
#
?
Jan 20, 2015 23:10
|
|
|
hendersa posted:Actually, that might work. Roll the thing one way for select, the other way for start, and give it a good shake for the pause menu. I bet that'd make it really fun to do the Link's Awakening A+B+Start+Select combo that brings up the world map.  Is the thumbstick clickable? If so, you could map stick+A to Start and stick+B to Select and have most games covered. Seems like it would be reasonably natural to control.
|
|
#
?
Jan 21, 2015 01:39
|
|
|
Includes 3D Gyro, accelerometer, and magnetometer. Only two buttons. 
|
|
#
?
Jan 21, 2015 02:37
|
|
|
Onion Knight posted:Includes 3D Gyro, accelerometer, and magnetometer.
|
|
#
?
Jan 21, 2015 05:05
|
|
|
Shalinor posted:I don't suppose they're planning another edition or anything? It seems like a ridiculous oversight. Comments from the creator: Max Thrun posted:The start and select buttons are handled with Up+A+B=Start and Down+A+B=Select which works okay. The other buttons for other systems are not being handled. The lack of start and select are both a result of oversight and lack of room to fit them. I originally only planned on playing gameboy + nes on this so the other systems are just extra. This project was totally designed and built in 6 weeks so there wasn't a whole lot of design time up front. If I were to ever try to actually make a product out of this I would definitely add a dedicated start and select and maybe even shoulder buttons up top. 6 weeks isn't enough time to figure out that two buttons isn't really enough 
|
|
#
?
Jan 21, 2015 08:02
|
|
|
Scaevolus posted:6 weeks isn't enough time to figure out that two buttons isn't really enough 6 weeks of design and build time for an electronics project like that is not a very long time. Honestly, looking at that site. If it truly was designed and built in 6 weeks, it's insanely impressive. Considering they created the device for a two button use case, it's hard to fault them for it 
Buffis fucked around with this message at 09:26 on Jan 21, 2015 |
|
#
?
Jan 21, 2015 09:23
|
|
|
Buffis posted:6 weeks of design and build time for an electronics project like that is not a very long time. But, you know... that would take time and  My goal is to work around the shortcomings in software and get more people using Max's hardware.
|
|
#
?
Jan 21, 2015 17:50
|
|
|
2D water for an iOS game I've been working on. Really psyched about this. It's one of those things I didn't feel like ever doing, because it felt impossible since I didn't even begin to know how to approach it, but I feel that I've had a lot of success with it. I ended up using a series of horizontal points, and each of them have spring physics applied to them. In the gif above, I have an invisible object hitting the water at a certain speed, and the physics of the spring propagate to the left and right. I noticed it was very processor intensive to have the code loop through on propagation that has very low values, so I cut them off at some point and set them to 0. This stops all the water movement entirely though, so to keep the water "moving" I applied two sine waves moving in different directions and rates before it draws. This was very inexpensive. I overcame other challenges though, like tessellation, since I couldn't draw any concave shapes using the GPU, so I brake the water down into the least amount of convex shapes every draw. Another challenge was optimizing what polygons to apply the math to and update the least amount of array indices. It only draws on the current screen size, and does the math a little bit outside the bounds. The start and end point indices in for loops are found based on the screen's center at that time before they run, so it doesn't have to go through all the indices of the arrays. This lets me make the length of the water virtually endless without any performance hits. The frame rate is really lovely in that gif because I was running that in the simulator and capturing with Quicktime on not such a fast Mac Mini. Running on the iPad though it's always 60fps.
|
|
#
?
Jan 25, 2015 21:24
|
|
|
I was screwing around with C# console apps and Mono earlier today, and created a simple Awful Bookmark Updater.  It parses your SA bookmarks. After you select which ones you want update notifications on, it will start spitting out, well, notification updates. Right now it's set to every five minutes. Now, that's not all that interesting (my WinRT programs both do this. The only difference here is that I'm using the core library those programs use outside of it) But this little app also uses text to speech. Drastic Actions fucked around with this message at 01:43 on Jan 26, 2015 |
|
#
?
Jan 26, 2015 01:39
|
|
|
hendersa posted:Five minutes of awesome Man, I'm really impressed how snappy this thing is, especially recalling/saving save states.
|
|
#
?
Jan 26, 2015 01:54
|
|
|
LP0 ON FIRE posted:Man, I'm really impressed how snappy this thing is, especially recalling/saving save states. Sadly, I will forever be compared to stuff like RetroPie and immediately dismissed because I don't support as many emulators. I see hundreds of message board postings of people using RetroPie and comparing emulator config settings with each other to try and get various emulators running well. 
|
|
#
?
Jan 26, 2015 15:38
|
|
|
hendersa posted:Thanks! I am trying hard to make it polished and a good user experience. I think that a lot of other emulator projects are trying to enable as many emulators as possible and have lots of start-up scripts and the like to get everything going. My philosophy is to polish everything up and get it tuned and configured for the end-user to make the system more like an appliance or console. While installing 20 different emulators with different config files and directory structures has its place, I am going for a much cleaner experience. Why not pre-tune the settings for the hardware and centralize the configuration files? I kind of wish you had a windows version of your work, because to be honest I agree with everything wrong with trying to get emulators to behave/work well. Even Retroarch has huge issues with saving configs/paths/things acting half-baked. Still loving your work and following your thread. 
|
|
#
?
Jan 26, 2015 15:48
|
|
|
LP0 ON FIRE posted:2D water for an iOS game I've been working on. This is really great. LP0 ON FIRE posted:The frame rate is really lovely in that gif because I was running that in the simulator and capturing with Quicktime on not such a fast Mac Mini. Running on the iPad though it's always 60fps. Not sure how well it would work on the mini, but in Yosemite you can capture directly from the device in Quicktime >> New Movie Recording, and I've had really good results that way.
|
|
#
?
Jan 26, 2015 16:12
|
|
|
hendersa posted:Thanks! I am trying hard to make it polished and a good user experience. I think that a lot of other emulator projects are trying to enable as many emulators as possible and have lots of start-up scripts and the like to get everything going. My philosophy is to polish everything up and get it tuned and configured for the end-user to make the system more like an appliance or console. While installing 20 different emulators with different config files and directory structures has its place, I am going for a much cleaner experience. Why not pre-tune the settings for the hardware and centralize the configuration files? I really don't feel like you have to make a device that supports every emulator in the world. I don't know all the details of your hardware, but IIRC it supports a pretty wide array of old school NES consoles and handhelds. Care to bring me up to speed on this? I quickly gazed over your site, but couldn't find it. lord funk posted:This is really great. Thank you and thanks for the tip! I'll have to try that to see if I can capture the full width of the iPad with a good frame rate.
|
|
#
?
Jan 26, 2015 16:45
|
|
|
LP0 ON FIRE posted:Thank you and thanks for the tip! I'll have to try that to see if I can capture the full width of the iPad with a good frame rate. Keep in mind that it won't capture sound by default. But you can use Soundflower to reroute the audio to an input in Quicktime. edit: Actually never mind. You can select iPhone/iPad as audio input. I was thinking of capturing footage on the Mac itself. Superschaf fucked around with this message at 17:28 on Jan 26, 2015 |
|
#
?
Jan 26, 2015 17:26
|
|
|
They had to add the iPad capture feature once they created video previews for the App store. Making videos before that was such a pain, and required external video capture hardware.
|
|
#
?
Jan 26, 2015 18:05
|
|
|
LP0 ON FIRE posted:I really don't feel like you have to make a device that supports every emulator in the world. I don't know all the details of your hardware, but IIRC it supports a pretty wide array of old school NES consoles and handhelds. Care to bring me up to speed on this? I quickly gazed over your site, but couldn't find it. From the project.log thread a while back: hendersa posted:Right now, I have plans for supporting four platforms: NES, GBC, GBA, and SNES. I use the VBA-M, FCEUX, and SNES9X emulators as a base. In terms of emulation processing requirements, it is: In general, I should be able to support platforms from the 16-bit era without too much difficulty. Unless the emulator is super cycle-accurate at the expense of performance, I can tune the emu and profile it to get it running at a reasonable rate without enough dropped frames for it to be a concern. This is one of the big advantages that the BBB has over the RPi, since the RPi runs at 700MHz. People overclock the crap out of their RPis, though, but it voids the warranty and can cause flaky performance. Since overclocking the Pi is so popular, there are actually kernel command line options to do it on that platform. You can even get the Pi up to 1GHz, but most people clock it to 800Mhz or 900Mhz. The GBA is an interesting case because it starts to hit the point where dropped frames become noticable on the BBB. I've tuned it as best as I can, and there isn't any scaling in software, so I think it is in a pretty good spot right now. The GBA is a 32-bit platform, but the clock speed of the emulated processor is a much more important factor than the width of the bus. I would expect consoles like the SEGA Genesis and Master System to run comparable to the NES and SNES performance, though I have not tried them. Other handhelds, like GameGear, Neo Geo Pocket, N-Gage, etc. will probably emulate just fine, though I have not profiled any emulators for them to get performance benchmarks. The N64 is a bit of a question mark at the moment, since there are different approaches to emulating it. High-level emulation (HLE) uses knowledge of the graphics software APIs of the N64 to hook standard Nintendo dev library function calls and then route them to a native equivalent. So, if a game wanted to call a function that pushes three vertices to the GPU to render a triangle, the emulator would look for a program counter jump to that function's address and replace it with a call to a native OpenGL function that does the same thing. This way, you aren't emulating the GPU cycle-by-cycle, but rather replacing its functionality. CPU-bound operations are still emulated cycle-by-cycle, though the GPU is far more of a bottleneck than the CPU is. Back in the day, the defunct UltraHLE emulator could run N64 games on a 200MHz Pentium with a 3dfx card. Modern N64 emulators are more of a balance between cycle-by-cycle emulation and HLE. Dynamic recompiling and caching of instructions and registers are used, which cuts down on a lot of the cycle-by-cycle emulation overhead. Folks have gotten Mupen64 running on the Pi, which means I should be able to get it running on the BBB. Maybe I could leverage work already done on the Pi's OpenGL ES support for the BBB? Anyway, that is far down the road, as I want to get everything running well before I start adding more things in. Beyond the N64, PSX, PS2, Dreamcast, Gamecube, Wii, XBox, etc. are pipe dreams, I think. People demand these from me all of the time, and I just sigh and try to explain that it is highly unlikely these consoles will be emulated on a board like the BBB. Aside from the processing power that is needed, you also need to store the game data somewhere. Yes, you could have a USB CD-ROM drive or something but really... what is the point? I will generally be very nice and try to explain these things, but I get the impression that most people demanding these things are 12 year-old kids. Once I say that I will not even look at these platforms, I often get told that I'm just lazy or stupid or something. Go figure.
|
|
#
?
Jan 26, 2015 18:53
|
|
|
I just found out that BeagleSNES got a mention in the "Ten Amazing Projects for the BeagleBone" chapter of the new book BeagleBone for Dummies: I'll need to send them a thank you message for including my stuff and their kind words.  Edit: I also just found out that the book I wrote showed up in the Amazon listings today with a placeholder page. Looks like it will be available in February after all!
hendersa fucked around with this message at 21:42 on Jan 27, 2015 |
|
#
?
Jan 27, 2015 21:25
|
|
|
drat, $25 for paperback? Any chance you'd plan on offering a kindle edition or something?
|
|
#
?
Jan 27, 2015 22:52
|
|
|
Biowarfare posted:drat, $25 for paperback? Any chance you'd plan on offering a kindle edition or something? Most of the people that will buy it (like people prototyping Android devices who just want a few quick answers) will justify the cost because the sample code and collected info will save them a few hours of engineering and prototyping effort. If you have an engineer at a $50-$100 hourly rate, that adds up.
|
|
#
?
Jan 27, 2015 23:47
|
|
|
Pre-ordered just cuz I love your posts and I love to keep a pile of technical books next to my toilet.
|
|
#
?
Jan 28, 2015 00:38
|
|
|
bobua posted:Pre-ordered just cuz I love your posts and I love to keep a pile of technical books next to my toilet. 
|
|
#
?
Jan 28, 2015 04:13
|
|
|
Made a blender python script to automatically line up and uv-map polygons along the edge of a polygon. It's conditional, so it won't add waves in the middle of the water just because two water polygons meet there, for example.
|
|
#
?
Jan 28, 2015 15:36
|
|
|
Programmer Humor posted:Made a blender python script to automatically line up and uv-map polygons along the edge of a polygon. It's conditional, so it won't add waves in the middle of the water just because two water polygons meet there, for example. That looks straight out of Wind Waker. That's really cool!
|
|
#
?
Jan 28, 2015 16:25
|
|
|
LeftistMuslimObama posted:That looks straight out of Wind Waker. That's really cool! Thank you, that's what I used for inspiration.
|
|
#
?
Jan 28, 2015 16:38
|
|
|
The semi-relaistic WW3 simulator I'm working on continues to move further away from reality. Trying to replicate the Enders Game beam weapon efffect:  Blasting chunks off the planet:  This is my first thing using 3D, so it took me ages to figure out how to do the texture changes on specific parts of the models. The trick is to change the vertex colours and use a shader to lerp between textures based on each vertex's rgb value. Probably no secret for anyone familiar with 3D stuff, but there you go.
|
|
#
?
Jan 31, 2015 00:56
|
|
|
I think that first one looks really good except the angle is "off" compared to how it's coming in from the moon not exactly perpendicular to the surface so if it has some more momentum to the side there it'd be perfect, if that makes sense.
|
|
#
?
Jan 31, 2015 01:07
|
|
|
Snapchat A Titty posted:I think that first one looks really good except the angle is "off" compared to how it's coming in from the moon not exactly perpendicular to the surface so if it has some more momentum to the side there it'd be perfect, if that makes sense. I think it looks right the way it is: a lot of energy is delivered at a point and it explodes outwards symmetrically. Think about how craters on the moon/whatever are usually circular regardless of the impact angle.
|
|
#
?
Jan 31, 2015 17:12
|
|
|
I woke up this morning with dozens of mails in my inbox about BeagleSNES. HackADay finally accepted a tip about BeagleSNES and posted a blurb about it: http://hackaday.com/2015/01/31/passion-project-turns-beaglebone-into-standalone-super-nes/ I think the main reason why they accepted my submission this time is because it was filled with swearing. I figured that no one was going to read it (like the last few times I submitted), so I decided to go for broke. I guess that gets their attention? The main site editor sent me a message and apologized for missing the submission the last few times because they get "a lot of noise" at the submission mail address. I'm just happy because it gets the project in front of a different crowd (hardware folks versus the software BeagleBoard folks who now of it), so I might get a few new users, feedback, and good tips from people with a different set of views on things. 
|
|
#
?
Jan 31, 2015 18:02
|
|
|
|
| # ? Jun 10, 2024 12:36 |
|
|
hendersa posted:I woke up this morning with dozens of mails in my inbox about BeagleSNES. HackADay finally accepted a tip about BeagleSNES and posted a blurb about it: Congrats, you really been doing so awesome on this project. I'm curious about your swear filled submission. Would be interested in seeing it if you feel comfortable posting it.
|
|
#
?
Jan 31, 2015 18:31
|
|
