Re: [Rosegarden-user] quantization

[Ray Klassen]

oh no doubt there. rah rah rosegarden!

On Mon, Jan 2, 2017 at 10:30 AM, D. Michael McIntyre <
rosegarden.trumpe...@gmail.com> wrote:

> On 01/01/2017 09:01 PM, Ray Klassen wrote:
>
> > Ha ha. What a relief. I've tried to use grid quantize on midi files i've
> > pulled down off the net and like users the world over I thought it must
> > be something wrong with me.
>
> When I really think about it, the grid quantizer probably works better
> than I appreciate.  It's just that the heuristic quantizer works much
> better, and I always go with the best result.  It's specifically
> designed to get the best notation on the page for a typical human
> performance, and it looks for things humans do when they perform.
>
> That is the one thing separating Rosegarden from every other application
> on the planet, and it's just a back of napkin idea Chris had one day.
> He closed one eye, made some guesses, and took a stab at coming up with
> something.  It has issues that have existed since the first day, and he
> never addressed any of them.  The whole thing is only very slightly more
> refined than the first draft, as I recall.  Even so, it's an absolutely
> brilliant first draft!  That thing is truly the only reason I stuck with
> Rosegarden all these years.  You can find everything else in several
> other packages, but only Rosegarden does this good of a job rendering
> human performances as legible notation; all while preserving the
> original human performance.
>
> It mostly fails on grace notes, slurs, and heavily legato playing.  Plus
> of course when your timing is just way to hell and gone, like mine.  I'm
> a heavy off-the-beat player, and consistently hit everything either
> ahead of or behind the mark.
>
> --
> D. Michael McIntyre
>
> 
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Rosegarden-user mailing list
> Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
> https://lists.sourceforge.net/lists/listinfo/rosegarden-user
>
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[D. Michael McIntyre]

On 01/01/2017 09:01 PM, Ray Klassen wrote:

> Ha ha. What a relief. I've tried to use grid quantize on midi files i've
> pulled down off the net and like users the world over I thought it must
> be something wrong with me.

When I really think about it, the grid quantizer probably works better 
than I appreciate.  It's just that the heuristic quantizer works much 
better, and I always go with the best result.  It's specifically 
designed to get the best notation on the page for a typical human 
performance, and it looks for things humans do when they perform.

That is the one thing separating Rosegarden from every other application 
on the planet, and it's just a back of napkin idea Chris had one day. 
He closed one eye, made some guesses, and took a stab at coming up with 
something.  It has issues that have existed since the first day, and he 
never addressed any of them.  The whole thing is only very slightly more 
refined than the first draft, as I recall.  Even so, it's an absolutely 
brilliant first draft!  That thing is truly the only reason I stuck with 
Rosegarden all these years.  You can find everything else in several 
other packages, but only Rosegarden does this good of a job rendering 
human performances as legible notation; all while preserving the 
original human performance.

It mostly fails on grace notes, slurs, and heavily legato playing.  Plus 
of course when your timing is just way to hell and gone, like mine.  I'm 
a heavy off-the-beat player, and consistently hit everything either 
ahead of or behind the mark.

-- 
D. Michael McIntyre

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[Lorenzo Sutton]

On 29/12/16 20:38, Ray Klassen wrote:
> I've used rosegarden for years now and generally love it. I use it
> mostly for composition. What I would love is to be able to enter notes
> straight from a synth keyboard and have the quantization in rosegarden
> remove all of the anomalies from my substandard playing and leave me
> with just notes as if I had entered them with the mouse. -- no staccatos
> tenutos or variations in velocity. and all starting and ending precisely
> as scored. Is this possible?

I guess it would be helpful if you could share a non-quantized piece (or 
snipplet) so as to try various quantization options and maybe provide a 
more precise suggestion.

Lorenzo.

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[david]

Well, I use a lot of MIDIs found online for preparing hymns for my 
church. I think a lot of those MIDIs weren't exactly recorded by the 
best instrument player.

My own MIDI recordings are difficult to clean up. (Not that I'm a great, 
"currently practiced up to my old skills" keyboard player.) But I often 
think that MIDI's timing resolution is far more precise than that of any 
human. So the duration of a recorded quarter note isn't the precise 
960(?) that Rosegarden (might?) define it as. Then the quantizer gets to 
try to figure out just how to handle something simple like the end of 
the note. How far does the actual recorded note extend before the 
quanitzer decides it's supposed to be a quarter note tied to a following 
note?

Then toss in the apparent fact that Rosegarden really doesn't handle 
MIDI tracks with 2 contrapuntal melodies on a track.

Quantization does seem to help situations where the bar lines Rosegarden 
shows when importing a recorded MIDI - you know, where somehow more 
beats end up in a measure than belong there?

I don't know, but with recorded hymn MIDIs, I've often found it faster 
to just put in the notes in the score view than to try to clean things up.

One thing I find annoying about MIDI. Absolute time. Meaning a note that 
on a score should play right here on beat 2 of measure 4, somehow got an 
absolute time that puts it in some other measure entirely. I'm sure 
there was a reason for this in the MIDI standard, but it bugs me!

On 01/01/2017 04:01 PM, Ray Klassen wrote:
> Ha ha. What a relief. I've tried to use grid quantize on midi files i've
> pulled down off the net and like users the world over I thought it must
> be something wrong with me.
>
> On Jan 1, 2017 3:56 PM, "D. Michael McIntyre"
> >
> wrote:
>
> On 12/31/2016 06:43 PM, Ray Klassen wrote:
>
> PS You're right. I would have expected magic from the grid quantize.
> what is it doing if it's not doing this?
>
>
> Damn if I know.  I've never been able to figure it out from
> observing the results, and I've never actually looked into that code.
>
> --
> D. Michael McIntyre


-- 
David W. Jones
gn...@hawaii.rr.com
authenticity, honesty, community
http://dancingtreefrog.com

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[Ray Klassen]

Ha ha. What a relief. I've tried to use grid quantize on midi files i've
pulled down off the net and like users the world over I thought it must be
something wrong with me.

On Jan 1, 2017 3:56 PM, "D. Michael McIntyre" <
rosegarden.trumpe...@gmail.com> wrote:

> On 12/31/2016 06:43 PM, Ray Klassen wrote:
>
> PS You're right. I would have expected magic from the grid quantize.
>> what is it doing if it's not doing this?
>>
>
> Damn if I know.  I've never been able to figure it out from observing the
> results, and I've never actually looked into that code.
>
> --
> D. Michael McIntyre
>
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[D. Michael McIntyre]

On 12/31/2016 06:43 PM, Ray Klassen wrote:

> PS You're right. I would have expected magic from the grid quantize.
> what is it doing if it's not doing this?

Damn if I know.  I've never been able to figure it out from observing 
the results, and I've never actually looked into that code.

-- 
D. Michael McIntyre

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[Ray Klassen]

However dreadful your playing is, mine is probably worse. But the notation
quantizer/Fix notation function does most of what I want. I can reset
velocity across the board too. If I could just wholesale erase all of the
tenutos/staccatos, that would probably do the trick.

Thanks for your help.

PS You're right. I would have expected magic from the grid quantize. what
is it doing if it's not doing this?

On Thu, Dec 29, 2016 at 3:36 PM, D. Michael McIntyre <
rosegarden.trumpe...@gmail.com> wrote:

> On 12/29/2016 02:38 PM, Ray Klassen wrote:
>
> > with just notes as if I had entered them with the mouse. -- no staccatos
> > tenutos or variations in velocity. and all starting and ending precisely
> > as scored. Is this possible?
>
> I've found there is no magic fix it button.  In principle the grid
> quantizer should be your friend here.  Un-check [ ] Add articulations to
> make that stuff go away, dial up the most likely parameters, and try
> your luck.  In practice, the notation quantizer often does a better job.
>   You can use Adjust -> Quantize -> Fix notation quantization to make
> the durations you see on the page permanent.
>
> I can't find an option to turn off tenutos and such while recording.
> You can select all and use the trash can icon in the top right corner to
> throw all those marks away.
>
> All in all though, I've always found this to be a fiddly process.  You
> try something, and it works well for most of a passage, but there is one
> troublesome section where your playing was too far off.  Try a different
> quantizer, different options, or just draw a bit of notation by hand.
>
> The better you play, the better this works.  It has the most trouble
> with slurs and grace notes, which tend to come out as a mangled tangle
> of unintended chords.  In my own personal projects, I probably never get
> more than 2/3 of the notation cleaned up using automatic tools.  My
> playing is pretty dreadful.
>
> --
> D. Michael McIntyre
>
> 
> --
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, SlashDot.org! http://sdm.link/slashdot
> ___
> Rosegarden-user mailing list
> Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
> https://lists.sourceforge.net/lists/listinfo/rosegarden-user
>
--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization

[D. Michael McIntyre]

On 12/29/2016 02:38 PM, Ray Klassen wrote:

> with just notes as if I had entered them with the mouse. -- no staccatos
> tenutos or variations in velocity. and all starting and ending precisely
> as scored. Is this possible?

I've found there is no magic fix it button.  In principle the grid 
quantizer should be your friend here.  Un-check [ ] Add articulations to 
make that stuff go away, dial up the most likely parameters, and try 
your luck.  In practice, the notation quantizer often does a better job. 
  You can use Adjust -> Quantize -> Fix notation quantization to make 
the durations you see on the page permanent.

I can't find an option to turn off tenutos and such while recording. 
You can select all and use the trash can icon in the top right corner to 
throw all those marks away.

All in all though, I've always found this to be a fiddly process.  You 
try something, and it works well for most of a passage, but there is one 
troublesome section where your playing was too far off.  Try a different 
quantizer, different options, or just draw a bit of notation by hand.

The better you play, the better this works.  It has the most trouble 
with slurs and grace notes, which tend to come out as a mangled tangle 
of unintended chords.  In my own personal projects, I probably never get 
more than 2/3 of the notation cleaned up using automatic tools.  My 
playing is pretty dreadful.

-- 
D. Michael McIntyre

--
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user

Re: [Rosegarden-user] quantization to 8th note triplets

[D. Michael McIntyre]

On 03/08/2013 01:44 PM, Stefan Thomas wrote:

 how can I quantize to a passage to exact eight-note triplets, or maybee
 even sixteenth sixtuplets?

To experiment with this, I slapped together a passage consisting of a 
string of 8th note triplets entered manually in the notation editor.  I 
went into a matrix view and randomly messed up a number of the 
durations, making them longer or shorter than they were.

I went back to the notation editor, selected the new mess, and hit the Q 
button.

To clean up the notation only, the following settings worked rather well:

   quantizer:   heuristic
   base grid unit:  8th note
   tuplet level:triplet

After running that, the notation quantizer did a nice job of aligning 
the notes on the page with what they were supposed to be, and the messed 
up performance durations were untouched.  This is generally a good and 
useful thing meant to make the notation look pretty while the 
performance remains somewhat irregular and human.

If you want to do a hard quantize to physically change the performance 
duration, the usual tool would be to use the grid quantizer.  However, 
there is no suitable base grid unit to choose to align on 8th note 
triplets.  I have no idea why we have the base grid units we do, but 
that list has been unchanged for 10 years.

I started to pronounce this impossible, but then I remembered a trick. 
Select the notes, then Adjust - Quantize - Fix Notation Quantization. 
  That gets it done, and what you see on the sheet now becomes the 
performance duration as well, the same as if you'd gotten here with a 
grid quantize.

All of that being said, the quantizer can be very hit or miss, and 
sometimes there's just no substitute for a lot of laborious hand work.
-- 
D. Michael McIntyre

--
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and remains a good choice in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev
___
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user