Posts for ShinyDoofy


Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
LagDotCom wrote:
I've worked out another pattern! 3DO: 1993
Wow! Somebody give this guy a medal for this astonishing research work! *applaud*
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Get a preliminary encode from here. Any publisher is welcome to use parts of it as they feel. I screwed up the very first frame (it's empty), but sound seems good at ~68kbps ABR.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Finally figured out what was going wrong, at least for SNES encoding (haven't tried NES again): in unix/nesvideos-piece.cc the intro length and its overlapping part is determined. Of course I didn't have the files under the path hardcoded there, but after either deactivating the logo or providing the logo files, the video looks just great. In case I oversaw it somewhere on the page, I'm sorry for the spam. If it's not the case, a little comment in the encoding HOWTOs might be appropriate, I guess.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Oh, the boredom... I'll try and encode this one until it's published.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
I just read about the SNES mouse protocol. Seems pretty screwed up to me, too Bisqwit, I found some code that might help you. Here's something general about the protocol and this file includes a snesmouse.c and .h for querying the mouse and reading its input. Might also be used to emulate one for an emulator?
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
alden wrote:
with the same input.
Blindfolded!
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
CrazyTerabyte: all I can do is read the code and try to understand what is does. Unfortunately writing is a whole different thing for me. Applying patches might at least partially work, but as upthorn pointed out, it's even harder to rewrite code that relies on something that's not supported on your OS and that you also haven't worked with at all. Isn't there one developer on the team that could be able to create such patches?
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
I'll try and take a look at which version I get to compile and which one I don't. /edit1 Ok, I got all four versions I tried to successfully compile and run! To start off, Gens "Grrl" v9f: there already was a precompiled binary in the tarball, no big deal. make clean && make and I got a gens that already supported playing .gmv files. Only thing to critique: no GTK, it's purely console based. /edit2: Getting a GTK interface was a pure pain. The GTK build can't play movies anymore, but hey, there's something to click around in! Get the patch here. Secondly and thirdly, Gens 2.12a with and without replay / AVI patch: without the patch, it struggles on several occasions. After stripping "-DGTK_DISABLE_DEPRECATED" from GTKCFLAGS in the Makefile, asm errors came up. This can be fixed by stripping "-O3" from NASMFLAGS, also to be found in the Makefile. Afterwards, it compiles great! But it couldn't play movies yet. After applying Bisqwit's patch and again removing the DEPRECATED parameter, it compiled and ran great. Even avi creating works great. Only thing: You'd have to recompile it everytime you want to change mencoder's options. Could be solved by also writing the video stream to a fifo file, I guess. And finally, there's Gens 2.15.2, also available from Sourceforge. This one provides a ./configure script, running it without any parameters doesn't harm anybody. It compiled fine, but doesn't support watching runs or creating AVI files. Maybe I can get Bisqwit's 2.12a patch to run on 2.15.2. We'll see. Oh, my setup: I'm running Gentoo Linux on x86, Kernel 2.6.26 gcc 4.1.2 glibc 2.6.1 nasm 0.98.39 GTK 2.12.10 If case you'd like me to check any other version, drop me a note. I hope this helps a bit.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
I guess he's referring to dehacked's microstorage.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
It's been almost a month since Gens 10 came out. Is there some kind of progress for the Linux build to receive an update, too? Is it even planned to also port the changes to a Linux version? From what I understand, the patch to create avis under Linux should be fairly easy to adapt (dunno anything about the ASM parts). Once again, I'm not a coder, but I'd be glad to help out in terms of testing the Linux build.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Thanks for the fix! :)
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
It is the updated one. The only thing missing of course are the logo and the little text boxes at the bottom in the beginning. The mmv's crc32 is "fb5f14c4".
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
I had nothing to do and felt like encoding it until it's completely published. Note though that there's no logo or anything as I don't have one. /It's so big because the video stream is lossless. Might be used by a publisher if he/she wants to.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
punkrockguy318 wrote:
I got sound once with "--sound 1 --soundq 1 --soundrate 22000 --soundbufsize 96" but it crackles (like a really old record that has not been treated well over the years).
Try a soundrate of 11000, that's the new default in SDL.
That sounds even worse for me.
punkrockguy318 wrote:
It shows *.nes files. It no longer has the *.zip filter because zipped roms are broken in sdl right now
Ironically, it does now. Strangely though, the key binding for my NES pad didn't work anymore and even configuring them again didn't do a thing.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
punkrockguy318 wrote:
Hello all, I'm the main SDL dev on the fceux crew, so I'm going to try to respond to some of your issues.
Welcome to TASvideos :)
punkrockguy318 wrote:
What videolog are you talking about?
--videolog is a parameter that Bisqwit (this site's owner) added to .98.12 in order to be able to play a .fcm file and encode the video at the same time. The game's speed was determined by the speed your computer could encode the raw video and audio streams with mencoder so that no frame was lost or anything. This way avi files were made to publish tool-assisted speedruns on this site.
punkrockguy318 wrote:
Lua is particular about its version numbering; it breaks compatibility from minor release numbers. Lua is officially distributed as lua5.1 and even debian distributes it as lua5.1.
I didn't know about the version breakage, but lua is definately not officially distributed as "lua5.1". Check the official Makefile for 5.1.3. However the case may be, let's not fight over it. I'll just keep on patching SConstruct and the problem's solved for me.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
First things first, I'm on Gentoo Linux. I'm running kernel 2.6.26 on an 3800+ AMD dualcore (x86, not amd64). gcc 4.12, glibc 2.6.1. Thus, I'm using the SDL version.
qFox wrote:
I've been able to reproduce the crash on loading a zipped rom once, but no more after that. Kind of odd. Can you try to reproduce the crash and tell me exactly what steps you've taken?
Sure. I just checked out revision 809 and compiled it using "$ scons". I take any random ROM and zip it with "$ zip -9 romname.zip romname.nes" to create a file called romname.zip, In this case, I created fullmoon.nes with this IPS file.
$ crc32 fullmoon.nes 
63a94c78
$ zip -9 fullmoon.zip fullmoon.nes 
  adding: fullmoon.nes (deflated 29%)
$ ./fceux fullmoon.zip 

Starting FCEUX 2.0.2-interim...
Loading fullmoon.zip...

An error occurred while loading the file.
qFox wrote:
The .html doc now contains proper switches. It also contains a nice red warning that it's an ancient doc. Please keep this in mind. Feel free to update it to the current FCEUX if you want to, for Windows users there is the fceu.chm. The other documentation comments fall under the same category; that doc is ancient.
Great, thanks! :)
qFox wrote:
Config file: There's no backwards support for older versions simply because there are so many versions (not just in the FCEUltra branch).
Ok, makes sense.
qFox wrote:
Doesn't google give you enough info on keycodes?
The only thing I could find so far is this listing. It only features the names for the specific keys, not the numbers you guys use.
qFox wrote:
I'm adding the movie support as a feature request but I'm making no promises here :) [...] --special: I've reported it
That's great to hear. Looking forward to it!
qFox wrote:
Sound: I think this is still an issue with the SDL build and I'll wait with reporting it untill we're satisfied with the SDL build. When we are and if the problem persists, please report it again :)
I got sound once with "--sound 1 --soundq 1 --soundrate 22000 --soundbufsize 96" but it crackles (like a really old record that has not been treated well over the years).
qFox wrote:
gfceux: although I'm not sure what state gfceux is in, I've added your comment to the bugtracker. I think we are still looking for somebody to help us get that project straight. If you'd like to help please let us know :)
I'd be happy to test. Just noticed something: the browser for ROMs neither shows .nes nor .zip files which I think it did some time ago.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Raiscan wrote:
And without access to exact timing of when the NES turns on, your whole idea becomes pretty impossible to.
The thing is that a fcm file lacks the information on what frame the console actually asks the controller for its buttons the first time. The protocol for querying an NES controller is quite simple and thus it's easy to emulate one on a microchip. But there's the problem that the console might not always ask the controller about its state and thus you can't count the frames the game has actually run. So you can't know which input you're supposed to give to the console. Yet if the console really did ask the controller on every single frame, no matter if doing so would change something or not, then it's just a matter of how much memory the chip has. This might be a thing NES emulators "feature", that the controller input is given to the console although it didn't ask for it.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
X2poet wrote:
Uh... yeah. What?!
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Just 'cause it shows a Canadian? Or maybe because fart jokes are out nowadays?
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
According to SF bug 2037733, zipped .fm2 files may be used to watch a run. Trying to use such a zipped run results in a segfault for me. I tried submission #2042 for this and built the .fm2 using the Windows build of FCEUX 2.0.0. /Any progress for the things I mentioned on page 3 so far? :) //Also, why are zipped runs supported, but not zipped ROMs? ///I know I'm getting annoying, just wanted to add something about the Lua thing: when installing it by hand the filenames do not contain "5.1", they're called "lua.h luaconf.h lualib.h lauxlib.h" and "liblua.a".
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Again, you're welcome :)
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Ok, so summing up my thoughts/findings for revision 761: When running scons, it looks for "liblua5.1" in line 55 of SConstruct. That doesn't work for Gentoo Linux since only files named /usr/include/liblua.{a,la,so} exist. Thus, Gentoo people would have to remove "5.1" and possibly even fix the include and link parameters in line 65 (-> "-I/usr/include/ -llua"). I like the idea of having such a thing around, yet I think it should be optional because a lot of people will probably only use the emulator for regular playing and/or watching a run. I know this point really is whiny, but: "src/cheat.cpp:928:2: warning: no newline at end of file" fceultra/fceu/documentation/fceultra.html supposes only a single dash for command line parameters such as "-pal x" although calling fceux itself clearly shows two dashes (although both versions seem to work). The input configuration documentations differ: ./fceux shows "--inputcfg, -i", but doesn't list gamepad{1,2} as need parameters. The html document instead lists "-input1 x and -input2 x". So if you start fceux with one of these, it asks for "GamePad #4", no matter whether you chose input1 or input2. Although I haven't tested it yet (what with?) there's emulation for the Family Keyboard. I'm almost certain the standard binding of the `/~ key for the @ sign and other non alphanumerical keys such as the +, [ or \ key won't work on non-US keyboard layouts. This would make a SDL keycodes files even more worthy of having to rebind these. There's a new config file syntax, which is nice. But is there a way of importing the old config file instead of having to setup it all over again? A little help file listing the keycodes would be nice so user's can edit their key bindings individually. Regularly zipped ROMs are not accepted ("An error occurred while loading the file.") which imho is a huge regression from .98. In case you're trying to implement this again, bzip2 and 7z support would be highly anticipated, I guess, because they can decrease the ROM' size drastically. .fcm files are not supported anymore (it exits with a segfault), which I can live with. I'm just afraid that people who are new to this site might have a hard time finding the option to convert these fcms of older runs on the site. So far I haven't found a way of converting "old" .fcm files to "new" fm2 natively under Linux. Until now, I'd have to start the Windows build with wine and have it converted. Movie creation. So far there's the "--soundrecord" parameter to write the sound of a game to a file, but apparently there's no way of also grabbing the video. This is no regression (as I've only found Bisqwit's .98.12 version actually supporting this), but I'm sure there will be people creating runs with v2 and this leaves the avi creation up to Windows people. If I activate OpenGL it runs on something like 10 to 15 fps or even lower, which is really a pain and I have no idea why. nVidia GeForce 7800GT, 173.14.12. I'm not willing to use a beta driver that is not officially supported (which would be 177.13 iirc). The former "--special(fs)" parameter for setting a video filter is missing. When providing it nonetheless, fceux seems to forget about its former sound configuration and still wants to play sound. When trying to get sound, it hangs itself with the following message: "Audio write: Input/output error". When this happens, the video windows pops up just like before, but stays gray. It seems that fceux can then only be shutdown by SIGKILL. I haven't been able to get any sound at all thus far. I'm using OSS 4.0.1016. gfceux uses find_binary, but without modifications "only" looks in PATH (an environment variable). Although I also like the feature of having it automatically installed via "python setup.py", this in my opinion is bad for unexperienced users who then can't easily remove it again. In such a case they would have to delete the files by hand via a terminal. Try explaining such a thing to a purely desktop fixated and typing-unknown-commands-into-something-I-don't-like-because-it-ignores-my-mouse-scares-me user coming from Windows. Clearly a .deb/rpm/whatnot package would be the overkill; still one simple command to remove it again is easier for those people. Back to my initial point: personally I don't like scripts spreading files all over my filesystem (yeah, I'm exaggerating, I know it's just /usr/local) without being able to get them removed again in a matter of seconds. Although it wouldn't be a problem for me, it still would be some kind of annoying imho. I hope my newly made comments sound at least a little better than my previous ranting.
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
Experienced Forum User, Published Author, Former player
Joined: 12/5/2007
Posts: 716
That's really odd. They both work flawlessly for me with mplayer. What I noticed though was that with vlc the sound stutters like hell. So I just redid the avi with ffmpeg instead of avidemux, this one works great on mplayer and vlc: http://files.filefront.com/meganewavi/;11399096;/fileinfo.html Sorry for wasting your traffic