I'm using fmod and there is a function FSOUND_Stream_GetTime(stream) which returns the number of milliseconds since the beginning of the track. The only problem is that it's not 60milliseconds to a second and so forth. Instead it's just a single number from 0 to infinity. How do I break this single number into minutes, seconds, and milliseconds in the 60:60:60 format. Sorry if this is a bit offtopic.

yeah the problem I'm getting it's coming out like 00:00:78 instead of 00:01:18, etc.

That's because there aren't 60 milliseconds to a second. Nor should there be. The prefix "milli" is used to denote 1/1000ths, not 1/60ths. To get seconds, divide the milliseconds by 1000.

What Matthew wrote should get you what you need. If, for some odd reason, you really want 1/60ths of a second, you should be able to convert the code that computes the variable "ms" into something that gives you 1/60ths of a second.

alrighty that worked.. what's wrong with the 00:00:00 format instead of 00:00:0000, personally I like it all 1/60.

The ms display won't be 0-99. If it's 0-99, they're centiseconds.

Personally I've never seen seconds subdivided into 60; it's usually 100 or 1000. Sometimes it's 24.

I'd recommend minutes:seconds.centiseconds or minutes:seconds.milliseconds (note the use of . not : ).

int ms = (LONG_LONG)al_duh_get_position(dp)*1000>>16; /* :P */
int secs, mins/*, hours*/;

secs = ms / 1000; ms %= 1000;
mins = secs / 60; secs %= 60;
/* hours = mins / 60; mins %= 60; */

/* By all means replace this with text_printf().
 * %02d will print a trailing zero for <10.
 * %2d will print a trailing space for <10.
 * %d would not print a trailing space.
 */
printf("%2d:%02d.%03d", mins, secs, ms);
/* printf("%2d:%02d:%02d.%03d", hours, mins, secs, ms); */

If you'd like to use centiseconds, replace 1000 with 100 and ms with cs everywhere, and change %03d to %02d. (Using FMOD's function, divide by 10, obviously.)

hey enthen does DUMB support FFT output, I heard you a long time ago talking about finding the algorithm and whatnot, but have you actually implimented it yet? Oh and I think I will take the mm:ss.ms where ms = 1/100s method.. it looks nicer.

hmm interesting replies.. I never thought that displaying a timer would be so complex.. maybe I should specify a config file for my media player that allows the user to set the time format.

[edit]
Oh and I added mp3 id3 v1.0 reading for the media player as well.... only I use fopen and fread which is really slow... my 300 or so files take about 12 seconds to load which is way to much... I need to get this down to about 2 seconds or people will think the program hung up, only I don't know if allegros packfile reading is any faster.
[/edit]

ahhh brilliant.. that should work.. though I can't really impliment that at the moment because of HW.. but brilliant non-theless I would of never thought of that.. it would update just like winamp does as well... possibly I could do it for only those in the view of the playlist. It's very doable as well, not much to change...

Ack! PLEASE stop referring to 100ths of seconds as milliseconds! They're not milliseconds, they're centiseconds!

No, DUMB does not do any FFT or similar algorithms yet. I implemented it in ML, which has a tendency to be one of the slowest languages around (although apparently it "can now run as fast as C"... ) Anyway I didn't really understand what I wrote.

DUMB's .mod support is on its way though. DUMB CVS also does length calculation. It won't take much more work to allow you to stop a module from looping (it'll use a system that's already in place for the length calculation). So progress is being made

I'm dumb!, I really didn't mean to turn this thread into a DUMB plug thread as well

hehe... pushing other 'intelectual' languages are you paul? I'm dumb!.

Yep, it was Poly/ML. Cool, I'll try O'Caml then (it should be installed on here, though it might be an out-of-date version). However, unless the compiler can really work miracles, I doubt it'll do better than gcc on an algorithm that can work in place in a procedural implementation

Why do you mention g++ in particular? It's my understanding that compiled C++ code used to be slower than compiled C code (assuming the source file is understandable as both), but not any more...

For reference, here's the thread in which I implemented FFT. (I wonder if I should clone Steve Terry now )

By the way, XMMS is much faster than Winamp at reading the ID3 tags. Hooray for Linux's file system!

Thomas: If I was an intellectual, I'd reply "yes" to that.

Ben: I mentioned g++ because of this: http://www.bagley.org/~doug/shootout/ ... O'Caml did very well.

Anyway, definitely look at O'Caml. It's doubly cool because O'Caml is also a scripting language as well as a compiler, as well as giving you the productivity increases that you already know about from using ML.

Steve: Cool. I just thought that would be the best way, short of rewriting everything to use threads

Paul: It just seems you like those languages that are mostly used by scientists (who don't code much) and not by 'pro' coders.