Need someone who knows C++ (or some similar language) and how dwi OR sm files work

[Tasselfoot]

I need a converter made... and it needs to be made in 3 days. Preferably less. It's for an outside project that I've been contracted to help on. It will be PAID.

It needs to take either .dwi or .sm (either one, doesn't matter which) and convert it to an xml sheet. The output is in milisecond timing, not framer timing like FFR. I have an example of the output code... so, please let me know if you're qualified to do this and capable to do this and have the time to do it.

If you want to start working on it without seeing the exact output code, just start coding a generic converter that will convert to milisecond timing.

1 or 2 people are already working on this... and I only need one. So first one done will get the cash. And I'm paying a decent price for this, trust me on that!

-Tass

[prefx]

An outside project you say? May we ask what this project is?

[Oni-Paranoia]

Seems like its music game related since it includes a .dwi or .sm file. idk

[Tasselfoot]

prefx... it doesn't matter what it's for. it's a project that someone has hired me to work on. me, personally... not FFR. and yes, obviously it's rhythm gaming related.

they've asked me to get this converter made, along with some other things... and I'm asking you guys. and am paying nicely for it being finished. although 2 solid parties are already working on it.

[dean_machine]

Planning on making something to compete with Stepmania?

I'm not able to do this btw.

Oh, I see that you've been hired, not the person who did the planning.

[Tasselfoot]

i'm pretty much a consultant... but then i had to go and open my big mouth. so i wound up having to get a bunch of other stuff taken care of. long story. annoying story. perhaps for a rainy day.

[Prime8]

I know some C++, but probably not enough for what you are asking for Tass. Sorry.

[Tasselfoot]

as i've said, i already have 2 people who are working hard on this... and both will probably be done today. as such, if you're just seeing this, you may not want to start. probably too late.

[PsYcHoZeRoSk8eR]

This is one of those reasons why I wish I was good at coding.

I still might attempt this later for my own amusement.

[Kekeb]

Originally Posted by PsYcHoZeRoSk8eR

I still might attempt this later for my own amusement.

I agree, this seems like a relatively interesting project.

I'll be working on something for personal use, but if it's finished within the time frame then I can send it your way. At this point it seems unlikely, as you already have people working on it.

[Patashu]

I think I could do this for .sm relatively easy but I don't know C++ specifically. I know java syntax, which I hear is almost the same thing.

I'd start by making a chronological list of all the stops and bpm changes ordered by when they occur. I'd load in the start beat 0 offset and beginning bpm change and calculate at what point in what measure the next bpm change/stop has to be handled; I can determine the exact millisecond timing of every note that comes before that, so I'd parse it measure by measure determining first how many rows long a measure is then reading each note in as a fraction of the time that the measure takes + the time that the measure starts at. Once I get to the place where a bpm change or stop happens, I account for it (setting the bpm I'm using or just adding that number of seconds to the time for every note I parse hereafter) then search ahead for the next one, repeating the process. I don't need to worry about how the simfile's timing varies behind me, since I've already parsed it all.

Alternatively, I could start by taking my list of changes/stops and flattening them all out onto measure intervals, heading to each one and determining what part of what measure it falls on and using this to determine the next. This may or may not simplify calculations.

I'm not sure how much using only double floating point precision might mess up calculations, though. It's all rational numbers, so I could probably use BigDecimal or whatever the C++ equivalent is?

[argo15]

well, I failed, so close to completing this project but someone beat me to it.
btw to convert to milliseconds
Time=1000*(SecondsPerMeasure*MeasureNumber+SecondsPerMeasure(*LineInMeasure/NumLinesInMeasure)-Offset);
where SPM=60/(BPM/4);

[Tasselfoot]

fyi, have 1 working converter. and a 2nd that is like 98% done.

so i'm locking this... thanks guys.

ps - $100 went to the 1st converter i got. felt bad, so still sending $20 to the 2nd one.