Re: Features for a new release. (was Re: [Denemo-devel] Midi output faulty for whole measure rests in 4/2 time)

[Jeremiah Benham]

On Thu, 2008-10-30 at 11:35 +0100, Nils Gey wrote:
> Hi people,
> 
> a few thought about this topic.
> 
> 1)There are two ways to enter midi to get notation:
>   a)Enter them one after another, without timing information. You choose the 
> length (half, quarter...) of the Note in Denemo and then press a midi key or 
> a chord (any note that is played a few ms after the first belongs to the 
> chord) and its entered to the currently selected staff at the cursor 
> position. For me, in the past, that was the most quick and reliable way to 
> get clean notation out of midi. If you want to Notate via a Midiinstrument I 
> would suggest this method.

I could add this feature for next release. I would prefer chords to be
entered via the foot sustain pedal. When I do step entry I enter the
notes in so fast that it would think it was a chord for sure if I used a
timing method. I timing method could be used for people who don't have a
foot pedal. I think maybe triggered by overlap of more than one note
longer than a few ms. I think the foot pedal could be used for a control
mode also. Maybe if the pedal is hit alone and then released then next
key would change the rhythm a=0, b=1, c=2, d=3, e=4, this would not work
for rests or dotted notes though. The user would still have to move her
hands to the computer keyboard.    

I know how to implement this using alsa and oss methods. This won't work
in windows. I would need to learn how this is done with portaudio to get
that to work on windows. 

>   b)Real-time-recording, as you know and discussed, is to have a metronome or 
> some time-signal and then play along. I have never seen any application that 
> does good work and I tried many. You will need heavy quantisatzion(?) and 
> lots of human work after you recorded to correct everything. Drum notation is 
> even more a pain to record. Live-recording is normally for non-notation 
> people with pianoroll and all this sequencer-things. I suggest to drop that 
> feature and if you want a more "humanized" playback there are other ways (for 
> example via scripts) .

Yeah. I always found setting quantization confusing. Some people I know
at work enter notation in this way. Piano is their main instrument
though. 

> 2)To the sequencer topic. I don't know what these libs are that you 
> mentioned, but I gathered a few links regarding linuxaudio and jack-midi, 
> maybe there are helpful.

Thanks for the links,
Jeremiah

> http://www.linuxjournal.com/node/1004080 An Article about Jack-Sync/Transport
> http://0pointer.de/blog/projects/guide-to-sound-apis.html Guide to 
> Linux-sound apis
> http://pin.if.uz.zgora.pl/~trasz/jack-keyboard/ Virtual Jack-Midi-keyboard
> http://sourceforge.net/projects/jack-smf-utils/ Simple Midi player and 
> recorder via jack.
> http://apps.linuxaudio.org/ Other linuxaudio-apps.
> 
> There are a few other JACK-tutorials but there are all about Jack-Audio.
> 
> Nils
> 
> 
> 
> On Thu, 30 Oct 2008 08:28:00 +0000
> Richard Shann <address@hidden> wrote:
> 
> > On Wed, 2008-10-29 at 23:06 -0500, Jeremiah Benham wrote:
> > > > What sort of performance
> > > > can be gotten from the best stuff that there is? 
> > > 
> > > I am not sure what you are referring to here. Are you asking what kind
> > > of performance can you get out of commercial packages like finale and
> > > sibelius and many that are mentioned in the link you posted? 
> > Commercial or otherwise - you can get an idea of what can be achieved
> > with what sort of effort by seeing what others have managed to do when
> > that has been there sole goal. You wouldn't expect to do more unless you
> > were going to dedicate serious effort to it. From the rest of your email
> > I understand that just looking at those examples has encouraged you not
> > to try for an unrealistic goal in turning human midi into notation. From
> > bitter experience I know how easy it is to wander off into a swamp where
> > many others have waded in before. (It's a different matter, of course,
> > if you have a new approach to the problem).
> > Richard
> > 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > Denemo-devel mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/denemo-devel
> 
>