Evil midi file

The following midi file is one I took off Mutopia:

http://66.98.148.43/~xenharmo/midi/Mendelssohn_Octet_-_Score-1.mid

While not the only strange midi file on this site, it seems especially
egregious. I think I could fix it by hand editing, but does anyone
know what kind of midi file it is, and whether a conversion ultility
exists to convert it to a more normal midi?

--- In tuning@yahoogroups.com, "Gene Ward Smith" <gwsmith@s...> wrote:
> While not the only strange midi file on this site, it seems especially
> egregious.

Gee, you could have given a clue as to what you thought was wrong, or
what was bothering you, but...

> I think I could fix it by hand editing, but does anyone
> know what kind of midi file it is, and whether a conversion ultility
> exists to convert it to a more normal midi

It appears to be a perfectly normal midi file, just a badly put
together one. Just like someone can format and type gibberish into
Word, there you have it. With a quick glance, the problems are:

- all data is assigned to channel 10 (go figure: drums)
- there are four blank tracks, followed by the 8 tracks of the octet
- apparantly this was all step-entered, as all the note values are
severly quantized to exact values, and all the velocity data is 127
- someone must have made a lame attempt at carving some volume
musicality, as there are volume changes (CC 7). However, somewhere
around bar 17 they get written as a value of 1, which essentially
turns the volume off.

It was a pretty quick thing to fix all this in Sonar.

The raw notes are there, though I didn't have a score to check to see
if things are correct. If you want a plain midi file with all the
notes, just send me a private email with any details (like if you want
different tracks assigned to different channels, etc). I will be away
at work on Thurs, so Friday at the earliest.

I just love what people do to music with the computer sometimes. :)

Cheers,
Jon

--- In tuning@yahoogroups.com, "Jon Szanto" <jszanto@c...> wrote:
>
> --- In tuning@yahoogroups.com, "Gene Ward Smith" <gwsmith@s...> wrote:
> > While not the only strange midi file on this site, it seems especially
> > egregious.
>
> Gee, you could have given a clue as to what you thought was wrong, or
> what was bothering you, but...

Mendelssohn's Octet is for strings only. No drums.

> It was a pretty quick thing to fix all this in Sonar.

What was the result? The Mendelssohn Octet is scored for double string
quartet--four violins, two violas, and two cellos.

> I just love what people do to music with the computer sometimes. :)

I haven't been able to find any computer-readble score on the Internet
for this aside from this one. I was thinking of trying to turn it into
something halfway decent, in extended meantone of course. It amazes me
that such a well-known piece hasn't been done to death already; this
is even more of a warhorse than the Brahms string quartet #3. Mostly
it's just a matter of finding out which instruments go where, which
could be done by inspection, but I was hoping someone knew about
whatever kind of midi file this is. The person who made it clearly was
not hearing drums on his playback, so what kind of midi file is it?

--- In tuning@yahoogroups.com, "Gene Ward Smith" <gwsmith@s...> wrote:
> Mendelssohn's Octet is for strings only. No drums.

No kidding. I've only been listening to it for about 30 years,
including rehearsals here at the house!

> What was the result? The Mendelssohn Octet is scored for double string
> quartet--four violins, two violas, and two cellos.

It's a midi file - it can be scored for anything. Anyway, I took a
look at the file info, and it turns out it is probably a file that was
auto-generated from the GNU LilyPond notation program. Why and how it
didn't assign any instruments, added blank tracks, and put everything
on channel 10 (the default drum channel for those locked into GM) is
anybodies guess. It could be pilot error, it could be bugs in the
auto-generation.

> I was thinking of trying to turn it into
> something halfway decent, in extended meantone of course.

You'll have to do quite a bit of work, as even when I send it to you
it will be pitches and rhythm only, no dynamics or phrasing.

> Mostly
> it's just a matter of finding out which instruments go where, which
> could be done by inspection, but I was hoping someone knew about
> whatever kind of midi file this is.

It is still a midi file. It loaded with no problem and allows editing.
It just wasn't done correctly. As I used in my example, a Microsoft
Word document can be a perfectly valid file, even if filled with
nonsense. The file is fine, the contents are not correct.

I'll have it ready by tomorrow, but I gotta run now...

Cheers,
Jon

Gene,

One last thing: this is just the 2nd mvt, I don't know if you were
hoping it was the complete piece. If you have no need for just that
one movement, let me know and I won't clean up the file.

Cheers,
Jon

--- In tuning@yahoogroups.com, "Jon Szanto" <jszanto@c...> wrote:
>
> Gene,
>
> One last thing: this is just the 2nd mvt, I don't know if you were
> hoping it was the complete piece. If you have no need for just that
> one movement, let me know and I won't clean up the file.

They have the other three movements on Mutopia also. I'll email the
zipped file if that is OK.

--- In tuning@yahoogroups.com, "Gene Ward Smith" <gwsmith@s...> wrote:
> They have the other three movements on Mutopia also. I'll email the
> zipped file if that is OK.

Yep, got them when I got back from LA tonight. I'll do some work
tomorrow and get back to you...

Cheers,
Jon