Path: chuka.playstation.co.uk!news From: Dev Support Newsgroups: scee.yaroze.programming.sound Subject: Re: Help! Date: Mon, 09 Mar 1998 14:07:26 +0000 Organization: PlayStation Net Yaroze (SCEE) Lines: 101 Message-ID: <3503F79E.33B6@Playstation.sony.com> References: <3501FF35.B85A176F@codeworks.demon.co.uk> <35037E50.B938CA54@netmagic.net> <3503C715.28C1E454@codeworks.demon.co.uk> NNTP-Posting-Host: 194.203.13.10 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.01 (Win95; I) Our sound man says (1) Once the VB data has been transferred to SRAM, the vb can be happily overwritten, but the vh must remain intact in RAM. (2) the 28 byte thing should be no problem, since the aiff2vag tools convert a sample and make it the right size. How reproducible is your problem? Have you tried it with different vh/vb files? Does it happen just for certain keyons? Jim wrote: > > Elliott, > > I believe the code+data locations are fine. I've checked the size of > the binary and have around 100k left for the stack. > I also have a debug macro in my code which checks the stacks location at > specific points and it doesn't appear to be going anywhere near the last > bit of data. > > My games will only start crashing if I start playing sound effects... > I'm also not intentionally trashing the .vh .vb location while the > program is running so they should be ok. > > Its just driving me insane, especially since I wanted to release Z2 > tonight. > > Cheers > > Jim > http://www.netyaroze-europe.com/~Sig1LL > > Elliott Lee wrote: > > > > Jim wrote: > > > > > > Hi All, > > > > > > Ok I'm loosing hair rapidly. > > > > > > I have sound effects in my program. And they work (and sound) > > > fine. But now my program appears to be crashing, Its easily > > > reproducible, so testing to see if it's working is trivial. > > > > > > If I remove all the sound api function calls the code works fine. > > > And everything works fine up to this one location in the game where it > > > will crash without fail. > > > > > > I seem to remember hearing that samples should be evenly divisible by > > > 28. Could this cause a crash? > > > > I thought this was just for looping. I use some samples that aren't > > evenly divisible by 28 and they work just fine. > > > > > or was it just referring to loops? > > > Are there any other Gotchas with the sound api? > > > Do .vh and .vb have to remain in main memory after SsVabTransfer? > > > > Dunno. > > > > > I've also had problems with a .vb straddling address 0x80100000. It > > > worked fine though if I moved it to 0x80120000. > > > > Are you sure that your code isn't walking into that area? I changed > > my makefile to make the code compile the program starting address down > > at 0x80090000. I took that address and added the size of the code to > > it to get the next unavailable address (rounded up a bit). I read a > > post in one of these groups saying that some variables are not allocated > > space right in your compiled image (as opposed to, say, DOS programs). > > > > On a related note: when downloading the main program to > > the black box, I get 4 status messages which say something > > like loading rdata, sdata, etc. What are they? (I have been > > looking at the addresses it gives and if any conflict with > > data I have, I move the data around...) > > > > For instance, an ordering table with 2^14 orders of Z-resolution would > > blow up your executable to outstanding proportions, so it is allocated > > when you download your main code to the Yaroze. > > > > But, I could be completely wrong about this point. :P > > > > > The executable is located at 0x80170000 and from checking the mapfile, > > > does not appear to hit the stack. > > > > That's pretty high up there IMO, but you should still have breating > > room. > > > > > Any insights would be most welcome. > > > > > > Cheers > > > > > > /Jim > > > > - e! > > tenchi@netmagic.net > > http://www.netmagic.net/~tenchi/yaroze/