alex_holden

My dad learned the same exercise as an engineering apprentice at Lucas Aerospace in the 1960s.

A rooster?

There's an article by Geoffrey Crabb somewhere that talks about stamping out the rough shapes with a fly press before hand-filing.

Or rig a simple water trap between the gauge and the 'tina (a rigid air-tight container with in and out spigots on the lid).

Thanks for that Chris, it's very useful.

That's brilliant, Chris. I don't suppose you have a higher resolution version of this graph; I can't quite make out the labels?

Certainly sounds plausible, Jim. Did Wheatstone's earliest concertinas have the samper discs or was it a later improvement, I wonder? Sorry to drift off-topic again, how come it is called a pad if it's in a concertina and a pallet if it's in an accordion?

Post deleted.

Good luck!

Just saw this one on eBay. I've not seen a concertina with ends built like this before. Anyone know any more about it? http://www.ebay.co.uk/itm/251493536854

I suspect that is much too short. In my experience Google Maps often greatly underestimates journey times in rural areas (e.g. the Highlands of Scotland). Personally I'd be allowing at least 10-12 hours.

Just tried it and I agree.

Thanks, I've added the parts to my notation above.

Hi Jody, my main Mac is running 10.7 (it's too old to run anything newer). I don't think the version I've compiled would run on 10.6, and I don't know of an easy way for me to compile a version of it for an older release of Mac OS. If anyone could use a 64 bit version of Polyphone built for Mac OS 10.7 Lion and up, let me know and I'll stick it up on my website. I think your best bet may be to download the Windows version of Polyphone and run it under WineBottler. The user interface looks and behaves a bit different to a native Mac application, though it does appear to work OK.

Given that it's April 1st, you probably should have announced something completely different had won! Hurrah! I've been looking forward to learning this one. I've no idea what the stuff below the dots means, but here is a transcription into ABC: X:1 T:Zelda C:Philippe Plard M:6/8 L:1/8 Q:1/4=160 P:A2B2 K:Am P:A |: e2 A e2 A | fed e2 e | edc B2 B |1 Bcd c2 A :|2 BcB A3 | P:B |: ABc E2 E | ABc F2 F | ABc B2 B |1 dcB c2 A :|2 dcB A3 | I've made a tiny tweak to the second bar because it was the wrong length. [Further small edits at the suggestion of Chris and Doug in the comments below.]

Here's what I've been working on for the past couple of weeks: a TV or movie tune. Can anyone identify it?

Yes, and so does the SoundFont format, and so do software synthesisers. I think it might be technically possible to do this by creating two or more almost identical instruments within the SoundFont, and configuring the buttons on the controller to play the different instruments via different channels.

I agree "antiquity" is technically wrong, but doesn't "antiquated" have negative connotations that you might not want to apply to the instruments you sell?

The way I see it, on both types of instrument you sometimes play two notes in a row at a constant bellows pressure (which would cause a MIDI controller to send two notes with roughly the same velocity), and at other times you change bellows direction change between two notes, causing the second note to start out at 0 velocity and gradually increase (using a series of aftertouch or channel pressure messages). We therefore probably want to record each sample at a minimum of two velocities: one that starts at zero bellows pressure that the synth uses for notes that start out with a very low velocity, and one that starts with the bellows already at a moderate pressure before the button is pressed. A complicating factor is that the envelope of a note following a direction change will have a faster or slower rise rate depending on how energetically you are playing. I can't think of an easy way to handle that, short of having two or more different instruments within the SoundFont with different attack characteristics on those low velocity samples. The player could choose one that has fast rise times when (s)he is playing rapid bouncy energetic tunes, or one with slower rise times for gentler tunes. As Michael said, many compromises have to be made.

Many of the instruments in the Fluid R3 sound font are sampled in stereo; on some of them the left and right samples are virtually identical and the difference between mono and stereo is barely detectable even through headphones, but on others you get an obviously wider, fuller, and more realistic sound when you enable stereo playback. It's much more subtle than recording the ends separately so that left hand notes come entirely from the left hand speaker and vice versa.

Caveat emptor, voice of inexperience talking again here... I believe the simple way to do the release phase is to set the parameter "Vol env release (s)" in the Instrument section, which tells the synth how many seconds to spend reducing the amplitude to 0 after the key is released. For example in the Violin instrument in the Fluid R3 sound font, this parameter is set to 0.4. There is a slightly complicated method described here for playing a separate sample on release: this jOrgan forum thread. I'm not sure if it is worth the effort of implementing this (perhaps to capture the click of the button releasing on a concertina with noisy action?).

+ Import duty?

I'm forking this discussion off from the Bs'ing About A Diy Concertina Midi Controller topic. No worries, I don't think anybody is in a rush to get this done. Interesting. This is related to what I was saying over on the other thread about the velocity tables. A concertina note sounds different if you press a button and then squeeze the bellows vs pressing a button with the bellows already pressurised. Also consider what happens on an English/Duet when you hold a button down and "shake" the bellows back and forth. I think perhaps the lowest velocity samples will need to recorded by starting with the bellows still. Yes, but it also makes the files smaller and reduces the amount of memory used by the synthesiser. How long is the longest note you'll ever need to play? How big are your bellows? I would be inclined to loop all the samples, probably after two or three seconds (so most notes you play in real music won't reach the loop point). Looking at some existing SoundFonts this appears to be the standard way to do it, though some of them work better than others (quite a few I've listened to make an obvious "skipping CD" type hiccup every time the sample loops around). I don't know if recording the samples in stereo is worthwhile or not. I suppose it wouldn't hurt to do so if you have the capability. I think it would be best to record the sides separately though to avoid the bouncing stereo effect you get if you closely mic the sides separately. Let us know how that works out. I can't afford to upgrade my recording setup right now but I've been looking at eventually getting something like the Alesis IO/2 Express interface and a pair of cheapish condenser mics (perhaps Behringer C2s?). BTW I realised I was being daft when I said over on the other thread that I want to tune my Lachenal before I record samples from it - I can't think of any good reason why I shouldn't just tweak the samples into tune in software. I might have a play around with this over the next few days.

Great stuff! I like it that this isn't a slick choreographed and rehearsed routine like some of the other flash mobs I have seen videos of. If by chance I had been in the square when this kicked off, I would have asked around the spectators until I found somebody willing to join the dance with me.

After two evenings spent faffing around with Qt Creator and dependencies and OS versioning problems, I eventually managed to get Polyphone to compile and run on my Mac. I've played around with it a little and it looks pretty useable. My plan at the moment is to tune up my Lachenal and then have a go at making a SoundFont from it using my low-end audio recording equipment as a practise run. BTW it's very interesting to use Polyphone to take a look inside some existing sound fonts (both good and not-so-good) to see how they are constructed. Should we split off the discussion about creating concertina SoundFonts into a separate topic?