On Thu, Mar 25, 2004 at 10:24:58PM +0000, Bernard Hill wrote: > In message <[EMAIL PROTECTED]>, Richard Robinson > <[EMAIL PROTECTED]> writes > > > >Is anybody else here looking much at MusicXML ? I've been having a look > >over the last few days, and I must say, I'm rather impressed. It seems > >to me that this could all be tremendously useful to us, as ABC users. > > Why? It doesn't have the ability to write your own, and isn't a format > you could play from as ABC is on both counts.
Not sure what you mean ? "Write your own" - if you mean convert from abc, no it seems we can't do much of that, yet. But write directly in a text editor, you could do that. I doubt if anyone would want to, given ABC, which is nicer. "Play from" - you mean to look at directly ? Likewise (if you were using, for example, abcMIDI, it would be possible to convert it and carry on doing that). I'm not proposing it as an alternative to ABC, but as a complement. ABC is a very good way for us, people, to interact with computers on the subject of music. The advantages of MusicXML come into play when computers interact with each other. Easy interconversion would give us the best of both worlds. So, why ? 1) the traditional advantages of a "transfer format". Export our ABC to Sibelius-users, gain access to stuff written in Finale. If/as it becomes a well-known readable format, it could become the obvious format in which to archive/distribute stuff, for greatest readability. For example, to the extent that the examples on the webpage I put up work for people here, I should perhaps already be considering making http://www.leeds.ac.uk/music/Info/RRTuneBk/ available in that format. I notice nobody's yet said whether they do or not ? 2) if easy interconversion became possible, between the various ABCs and MusicXML, that advantage of transferability could extend to ABC softwares. See the comparison on the website - I took 2 files, and put them through abcm2ps. One was Finale-generated xml, one was "native" ABC, which I knew to have been tested against an ABC-parser I don't use (ie, a different dialect of ABC). Both, it's fair to note, had an element of selection - there is much about MusicXML that couldn't have been handled by the conversions I have available, and the ABC was chosen as being likely to be correct but challenging. Both generated errors, which required editing to fix, but the first was much easier to fix up than the second - MusicXML _can_ already be more accessible to an ABC user than a different dialect of ABC. If ABC writers were able to convert ABC into MusicXML when they need to pass their work onto anyone else, and ABC users were able to convert this "back" into ABC, this would give us all a new and useful handle on The Question Of The Dialects. The ABC file in question, by the way, was Jack Campin's McLennan.abc. I await BarFly's MusicXML-export with interest ... 3) doing things to MusicXML with XSLT stylesheets is, I propose, a possibility of great interest to the coders, hackers, tinkerers and question-askers among us - there are a vast number of things that are much easier done that way to a tune than by parsing the ABC. Since XML/XSLT proposes itself as a write-once-run-anywhere, platform-independent technique, it might be possible to develop libraries of useful tricks, open to any ABC user who can convert into xml (and, indeed, maybe, non-ABC users who come to xml from elsewhere. There's potentially a much wider pool of interested people here). And even more tricks would be possible if you could convert results back to abc when appropriate. For example, Jack Campin was asking a few months back about looking through a collection of tunes and finding those with a range suitable for a specific instrument. I can't remember if it was here or elsewhere, but I remember seeing it. Did you ever get any answers, Jack ? I have ways of doing things like that. You need to parse the ABC, unroll the repeat-loops, and so on, and then identify the note-pitches. This is not trivial. I've been using James Allwright's abcMIDI parser for this (because it seemed the closest to being quickly adaptable). Having obtained a list of pitches, I pick the information I'd like out of it, using whatever hack seems convenient. Perl, for nstance. For Jack to use this, he'd need to get my C source (somewhere under my Leeds "tunebook"), compile it, install perl ... and get used to working that way, which I doubt seems very natural in his environment, if it was even possible, which I doubt - my C wasn't written with Macs in mind, and even if it compiled, it probably wouldn't accept his ABC input. Techniques and tricks for processing ABC are not nearly as transportable as the data is to other people/platforms/ways of working. Whereas, when I started looking nto MusicXML, one of the first stylesheet example I saw is "noteReport.xsl". Which works through the notes of a tune, counting lengths and pitches, and displays results as a page of html. This just fails to answer Jacks question by omitting to distinguish the octave of the notes. I'm not too familiar with XSLT, but I guess it could be fixed. And if Jack could convert his ABC into XML and then run these stylesheets on it, this could well be a much more feasible way of looking for an answer to his question than any other I can think of. And the stylesheet to do this would be usable for anybody else who wanted it, regardless of platform. The techniques and tools are as transportable as the data. So that's "why" (*grin*). For now. I may think of more later. Or discover I'm wrong, of course. > >a command-line Windows/Linux abc2xml (it's incomplete, it's buggy, it's > >closed-source and email to the author bounces, but apart from that it's > >great. It exists, for a start), what other ways are there to turn > >abc into xml ? > > > > Print it out and use SharpEye to scan it in. > > SharpEye scans and produces Midi, Niff, MusicXML. Ah, yes. Not running windows, I can't look at this, but maybe this could work (if it could run from an image, that might be more convenient than hitting the paper. Not altogether a simple transparent process, though). -- Richard Robinson "The whole plan hinged upon the natural curiosity of potatoes" - S. Lem To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
