sorry, my sister downloaded those songs, so I have nothing else to convert. To be honest I prefer the theme to zelda. :)

Hey, could you please upload it to CalcGames? I'll post an update on it about it being a complete release.

Very Good Job! Woot!

Wow, the sound quality on that is great! However, the higher quality of the sounds people start using in their programs, the more space it will consume. I wouldn't recommend using your calculator as a (rather bulky) iPod, but the notion of playing actual sound on your calculator is very cool nonetheless.

If the drivers for USB drives come out, this could go even farther...gigabytes of space and higher quality...the next generation of mp3 players(sp?) could be coming!

TI could also start including card drives such as a SD card reader like the HP-49. 2 GB of music on my calculator. Too bad there are not enthusiast for HP calculators as there are for TI, my dad has the HP-49. Oh, and yes, I know that there is something after the 49, I just can't recall it.

The 49G+...that's what I use in junction with my 89T...

i have a 83pse and had to clear all mem to install the app...the quality was worst....although the size was over 500 000 bytes...the song was , yes , really screaming! worst...worst...worst....but its one of the best ideas i´va ever seen...respect!

Excuse me?

this is for the 84+, not 83+!!!

Uh... The Ti-83PSE is compatible with the 84's(except 4 USB) and it runs this program JUST FINE though it can't hold much music. Wow, NECRO!

Absolutely amazing!!!

Yeah, congratulations Jim e!!!

I don't suppose a 68k port will be happening, will it?

I don't see why not. btw how does this program get so much better quality sound then prevous programs?

See his first post about it on RevSoft (link to revsoft above)

that is quite vauge. all you really can send to the link is on and off. it would be nice if he realesed his scorce and/or a better description of the algorithm

All he does is turn it on and off really fast.

But that's how ALL sound programs work...

How much CPU time (duty cycle) is being used to reproduce the sounds?

Is there any remaining CPU time to build this functionality into a game?

Bout 84%~68% cpu time is consumed by modulating the link port. And 16%-32% is used to actually read from the memmory. These figures are affected by the sample rate. However all I really need to do is make sure that I have a suficient number of clocks cycles per sample to flip the link port on then off with 256 degrees of accuracy. So at 15mhz playing 32khz audio I have 468 clocks available. 200 of those are used to read from memmory, adjust the pointer, handle keypress, and execute proper modulation code. So this leaves 268 clocks to modulate the link port, This is the standard turn link on, wait, turn link off, wait, play next byte. This is called Pulse-width modulation(see link), its the same method used in pretty much all gray scale attempts.

This would be impossible for in game use. But there is technically time for compression. Especially if its something as quick as RLE, in which I end up having time to spare. With my test on RLE I was a able to compress the audio up 75% by applying thereshold to compressing code. The sound was fine, but I killed sychning so that caused static. I left it off because I did not want this project to end up on the shelf of no return.

Yes, I'd love to see it come out on 68k!

i doubt it will be released on a 68k. 68k asm is completely different than z80 asm. It would have to be completely rewritten. That being said, if it is possible on a z80, its certainly possible on a 68k, you might even be able to add some compression support on a 68k.

This is the greatest program...I am changing all my songs for the calc.

So far I can get a 70 second song with singing, at 22.05KHz 8bit mono. This is about cd quality... no static or nothing.(Note this app takes 94 app pages and you are left with 0 arch memory.)so if you want to make a song 70 sec., 22.05KHz,8 bit, and mono, you need and 84+silver and memory cleared.

Then if I do the same song at 16KHz I can get 90 seconds but there is a little static

This is a great program!
now if there is a way to make the apps smaller then that would be great so that we can fit more than one song. But this is a start.(or Ti needs to put a few gig on their calcs.

If the USB driver comes out, you could store the music on USB disks and have the next generation of mp3 playback devices!

All the code for reading files from flash drives already is completed and has been released. The problem is there is nearly zero interest because in order to use a flash drive, you have to buy a ~$12 adapter off the internet.

I never thought that would be such an impetus to using flash drives, but it seems that no one wants to spend any additional money.

I immediately purchased the converter when I saw the demo for the flash drive. I just wish someone would write more software for it.

I tried to use it with my lexar jumpdrive (PD128-00-522 if you care), couldn't get it to work. But from the looks of things I don't think I'd have time to read from the usb. Perhaps if the user would be willing to accept a skip noise every 2 seconds or so I could buffer a good deal of the wave. Theoretically though there is time to do video. Of course that means bigger file sizes.

Oh by the way, love the commenting in your source. :D

Maybe if you spread the buffering out a little more?

Can you cut the play quality by maybe 10-20% and make the file read faster? What we need is a WAV/MP3/... reader app that can access and read from a flash disk/hdd disk and play back the songs like a mp3 player (sp?) and has a nice format. Probably a lot of work though.

Um, do you really think that a calc could decode an MP3 in realtime?

who knows? maybe someone will come out with it. anything is possible. calcs are equal to the early computers, some even faster.

Yes, but I really doubt the PC or even XT could decode MP3s in realtime, especially when most of the processing power needs to be used for cycling the outputs.

A 486DX2-66 is the general minimum for playing mp3s, and that's with a regular soundcard (not link port toggling).

Exactly. So unless you REALLY overclock your processor by a factor of ~10 (and I doubt it would work at that speed) and install a 486 instead of a Z80 or 68k (good luck trying to fit it in the case!) and TI adds a pretty advanced ADC with buffering and DMA and stuff to the hardware, MP3 decoding isn't in the future of graphing calcs.

I HAVE IT!!!
Use linear predictive coding!
Ever hear of a lossless codec called FLAC? Basically, it predicts the next audio value based on the previous samples, and stores the difference. Uses a lot less bits. MUCH easier to implement than MP3 anyway. I actually think this might be do-able.

I meant impediment, not impetus.

I was gonna say...

Yes but, if you're going so spend $300+ on an ipod, what's $12?????????

If it would work, it would make sense. Also, there are cables that can go from any plug to any plug -- ref: RadioShack "JDIŽ GoldX QuickConnect 5-in-1 USB Cable Kit6 ft." GXQU-06 -- $29.99 and it could be used for other things...more worthwhile...

Any chance for an app that could print to USB printers and read NoteFolio vars? (I mean on the calc, not the computer)