Re: dynamicAbsoluteVolumeFunction documented

[Stephen]

After sleeping on it, I realized this is not exactly user friendly and it
still is not documented. But Han-Wen, dynamicAbsoluteVolumeFunction and
instrumentEqualizer still do nothing, zero, nada.

If my example is wrong, then midi-volume-equalizer.ly is wrong as well,
since they do exactly the same thing. They both bypass the function and 
manipulate the alist directly.

Han-Wen, there are implications in what I am doing that are implied, but not
stated that have yet to be addressed. The default Dynamic Absolute Volume
values in midi.scm are wrong and there still needs to be a way to append
instruments to the Intrument Equalizer. On many systems, midi volume 0.2 and
0.0 are nearly identical; the lower levels of midi volume are inaudible,
depending on the dynamic range of the speakers.

In my files, setting a note to piano is the same nearly as muting out the
part. The contrast between loud and soft is greater than it should be. In
midi.scm, the absolute volume alist is this:

("sf" . 1.00)
("fffff" . 0.95)
("ffff" . 0.91)
("fff" . 0.81)
("ff" . 0.71)
("f" . 0.61)
("mf" . 0.50)
("mp" . 0.40)
("p" . 0.30)
("pp" . 0.20)
("ppp" . 0.10)
("pppp" . 0.08)
("ppppp" . 0.05)


Interestingly, there is no such thing as \fffff, but nevermind that. I
propose the midi volume range be from 0.2 (same as 0.0) to 1.00. And I
propose you substitute my values in midi.scm:

("sf" . 1.00)
("fffff" . 0.95)
("ffff" . 0.92)
("fff" . 0.85)
("ff" . 0.80)
("f" . 0.75)
("mf" . 0.68)
("mp" . 0.61)
("p" . 0.55)
("pp" . 0.49)
("ppp" . 0.42)
("pppp" . 0.34)
("ppppp" . 0.25)

Attached to this message is midi.scm with these values substituted. Although
they are not perfect, the previous attachment set-midi-absolute-volume.ly
demonstrates that the midi volume values are evenly spaced. The point it not
to evenly space them numerically, but to evenly space them audibly; so you
have to test the file by listening to the midi output to see what I mean.
Actually, I changed pppp and ppppp so I will attach both files here just so 
there
is no confusion.

As you can see, it is as if I am defining 0.20 as zero volume and adding it 
to the midi volume of ppppp, arranging the dynamic midi volume values 
between 0.25 and 1.00 inclusive. Of course actual midi zero volume is 
different on every system, but I think effective midi zero volume does not 
correlate with 0.00 on any system. This is a judgement call I am making 
based on the experience of having to play the dynamic markings in the piano 
music properly for my teacher.

Actually, if my volume is set low, I can't hear any midi volume below 0.50. 
I set p so even at low volume levels I would not think it was muted out. 
Anything below p still isn't audible unless my volume is medium to loud 
which is what I expect. It is subjective, but it is not as subjective as you 
might think. The best way to establish the midi volume levels is to set sf 
to 1.00 and set the ones below it to be as high as possible while still 
maintaining an audible distinction between each one. That's basically what I 
did. Assuming we should arrange them evenly between 0 and 1 is wrong. I 
doubt that all the midi instruments even respond to the full range of midi 
volume values.

As far as '
;; this is broken: we should not ever export variables from Scheme.

' goes, I can read and understand computer languages, but the comments are 
not always so clear. Perhaps if I compared midi.scm to the other files in 
its directory it would become clear, but the code works and I am focusing on 
what it does rather than how it does it. The worst thing about the TODO and 
your response is that in neither instance was an alternative even hinted at. 
Are you saying the procedures do or should be made to work and setting the 
alist associated with each one directly is not the method you want to go 
with?

Stephen

----- Original Message ----- 
From: "Han-Wen Nienhuys" <address@hidden>
To: "Stephen" <address@hidden>
Cc: "lilypond devel" <address@hidden>
Sent: Saturday, May 21, 2005 6:05 AM
Subject: Re: dynamicAbsoluteVolumeFunction documented


> Stephen wrote:
> > dynamicAbsoluteVolumeFunction still has  [DOCUMENT-ME] as its
> > documentation.
> >
> > I cannot find any use for  instrumentEqualizer  or
> > dynamicAbsoluteVolumeFunction  other than to set the
> > instrument-equalizer-alist and absolute-volume-alist respectively which
> > it does automatically anyway and I recommend removing them from
> > perfomer-init.ly and the documentation.
> >
> > instrument-equalizer-alist and absolute-volume-alist can each be set or
> > appended to directly.
> >
> > midi-volume-equalizer.ly in input/no-notation demonstrates how to append
> > an instrument to instrument-equalizer-alist.
> >
> > I am attaching a file called set-midi-absolute-volume.ly which
> > demonstrates how to set absolute-volume-alist to new values which I
> > suggest placing in the same directory.
>
> I think you're getting it backwards. The scheme variables don't work
> together well with LilyPond property scoping, and my recommendation is
> actually the reverse: to make the alists private. midi.scm doesn't say
>
>   ;; this is broken: we should not ever export variables from Scheme.
>
> for nothing.
>
> > Perhaps it is not necessary to define default-dynamic-absolute-volume and
> > default-instrument-equalizer public in midi.scm. Do you have to define
> > the procedures public in order to define the alists public? Or are they
> > public only to make the procedures themselves accessible to the user?
>
> > Do the examples in input/no-notation show up anywhere in the
> > documentation?
>
> No. Primarily because there is no output. Perhaps we should study on how
> to get MIDI linked in from the HTML pages.
>
> --
>  Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen

set-midi-absolute-volume.ly
Description: Text Data

set-midi-absolute-volume.midi
Description: audio/mid

midi.scm
Description: Binary data