Concertina Midi Samples

[Brian Humphrey]

In recent discussions on midi concertinas, some interest has been expressed in having midi concertina files. For example, these quotes from Jim Lucas, May 7, 2005, in "Anglo Midi Photos Added":

 

" ...what if you sometimes want to sound like a concertina, and sometimes like a trumpet, marimba, percussion ensemble, string orchestra, or "Thai bath house" ...?"

 

"Or what if you wanted to be able to select between different concertina voices? Even though you would have to record your own samples, you would presumably only need to do that once, and then you wouldn't have to carry around an amboyna Edeophone, a pinhole Æola, a metal-ended New Model, a brass-reeded baritone, a contrabass, and a Stagi miniature. Or, if you're an anglo player, C/G, G/D, Bb/F, and baritone in both metal- and wooden-ended models, both Jeffries- and Wheatstone-reeded. Or just think about taking sound samples from that Jones with the lovely tone, but sluggish action, so you get the same tone from a new instrument with Geuns-Wakker rivetted action."

 

You wrote what I was thinking, Jim.

 

As an interested "midi neophyte" at the very beginning of the learning curve, I'm wondering if we could collaborate to build a library of concertina sound samples to share. Collectively, we have a wonderful variety of concertinas that could be sampled.

 

What do you all think?

[Chris Timson]

I would be interested in contributing, if I knew how. I don't posess a sampler. It may, for all I know, be possible to use one of my computers, but I don't know how. FAQ anyone?

 

Chris

[Stephen Chambers]

FAQ anyone?

<{POST_SNAPBACK}>

MIDI 101 ?

[Henrik Müller]

--  Even though you would have to record your own samples, you would presumably only need to do that once, and then you wouldn't have to carry around an amboyna Edeophone, a pinhole Æola, a metal-ended New Model, a brass-reeded baritone, a contrabass, and a Stagi miniature. Or, if you're an anglo player, C/G, G/D, Bb/F, and baritone in both metal- and wooden-ended models, both Jeffries- and Wheatstone-reeded. --

<{POST_SNAPBACK}>

An interesting idea. The main problem would (probably) be that once all notes from an instrument are sampled, they will need to be adjusted in amplitude, so they sound even (don't have large jumps from note to note) and correct change in amplitude from low to high end, resembling the sampled instrument.

 

What about the pressure (pull/push)? I happen to play, fairly violently, a miniature Stagi. Violently, since it is my way of getting an expression out it. What I am aiming at, is that timbre will changed with the pressure, so which should be chosen when sampled?

 

Don't get me wrong here - I am not trying to kill a good idea at birth - it is a fascinating idea. And since I have one of the instruments with fewest buttons (18), I should volunteer to be the first!

 

But someone in this forum has done an instrument voice like that, I am sure. Can't remember who, but I have the voice somewhere, I downloaded it, and tried and it worked fine with the virtual synth it Apple's QuickTime. It could be valuable to get some input from there, from someone who has done it before.

 

Technically, I see (noooo, problems again!) - no, challenges, I mean: there are two important aspect of a sound: the attack part, and the steady-state part. I have no idea how it is done for samples to be used as instruments, except that the steady-state part need to be looped without glitches, so the whole thing goes like this: ATTACK-LOOP-LOOP-LOOP-LOOP - hmm, do we need a decay part as well?

 

Congratulations, Brian - I am interested! And "Thai Bath House" does sound intriguing, Jim!

 

Henrik

[Howard Mitchell]

--  Even though you would have to record your own samples, you would presumably only need to do that once, and then you wouldn't have to carry around an amboyna Edeophone, a pinhole Æola, a metal-ended New Model, a brass-reeded baritone, a contrabass, and a Stagi miniature. Or, if you're an anglo player, C/G, G/D, Bb/F, and baritone in both metal- and wooden-ended models, both Jeffries- and Wheatstone-reeded. --

<{POST_SNAPBACK}>

An interesting idea. The main problem would (probably) be that once all notes from an instrument are sampled, they will need to be adjusted in amplitude, so they sound even (don't have large jumps from note to note) and correct change in amplitude from low to high end, resembling the sampled instrument.

 

What about the pressure (pull/push)? I happen to play, fairly violently, a miniature Stagi. Violently, since it is my way of getting an expression out it. What I am aiming at, is that timbre will changed with the pressure, so which should be chosen when sampled?

 

Don't get me wrong here - I am not trying to kill a good idea at birth - it is a fascinating idea. And since I have one of the instruments with fewest buttons (18), I should volunteer to be the first!

 

But someone in this forum has done an instrument voice like that, I am sure. Can't remember who, but I have the voice somewhere, I downloaded it, and tried and it worked fine with the virtual synth it Apple's QuickTime. It could be valuable to get some input from there, from someone who has done it before.

 

Technically, I see (noooo, problems again!) - no, challenges, I mean: there are two important aspect of a sound: the attack part, and the steady-state part. I have no idea how it is done for samples to be used as instruments, except that the steady-state part need to be looped without glitches, so the whole thing goes like this: ATTACK-LOOP-LOOP-LOOP-LOOP - hmm, do we need a decay part as well?

 

Congratulations, Brian - I am interested! And "Thai Bath House" does sound intriguing, Jim!

 

Henrik

<{POST_SNAPBACK}>

 

There are a number of different concepts being discussed here.

 

MIDI (Musical Instrument Digital Interface) is strictly the physical and logical standard for transferring real-time music data from one place to another. It contains such things as Note Events (note on, note off) and Articulation (Bend, Volume etc). A MIDI signal doesn't say "play Bb softly like a piano", it says "use voice 5, start note number 34 with velocity 60, end note number 34". It's up to the sound producing device to interpret that.

 

MIDI signals from a controller (like a midi concertina) trigger a device to produce sound. This can be any sound but is often the General MIDI sound set which is what many people understand as MIDI sounds. It can however be one of many "Sound Banks" in the device.

 

A simple sampler will enable you to capture a single note from a instrument and put it in a sound bank then allow a midi signal to trigger the sound at any pitch and amplitude envelope. Attack, Deacay, Sustain, Release are the four stages usually modelled.

 

A more sophisticated sampler allows you to take a range of samples from an instrument, let's say the note C in 4 different octaves, each played softly, medium and loud. It would then allow you play the instrument using a midi signal and would select the nearest appropriate pitch and sound depending on the note number and the articulation data. This is how high end sampled instuments work without getting a "chipmunk" effect as the pitch changes.

 

There is an open standard for storing the mass of data which defines an instrument in this way. It can be simple or massively complex, using a different sample for each note and each level of articulation. The file is called a soundfont by analogy to a printed font.

 

There are programmes like Vienna Soundfont Editor for producing soundfonts and managing banks of soundfonts.

 

It would not be a trivial matter to produce a library of concertina soundfonts but it is doable.

 

If you want chapter and verse on soundfonts and understand terms like sub-chunk, hydra and big-endian go to http://www.soundfont.com/documents/sfspec21.pdf

 

 

 

Howard Mitchell

[David Barnert]

It would not be a trivial matter to produce a library of concertina soundfonts but it is doable.

<{POST_SNAPBACK}>

See my related post here.