Chiff and Fipple Forums

kenny wrote:Seamus Tansey : "Farewell To Gurteen / Morrison's".
https://youtu.be/ykzTaJ-D2mQ

You are certainly turning up some interesting material, but also challenging audio! That one is really more "Two track mono" than stereo. (Probably a separate mic each on piano and flute, plugged into a stereo recorder.) There is a bit of spill between flute and piano, but they are almost recorded individually, one extreme left, one extreme right. And a healthy dollop of background noise.

So, just for fun. compare it to this. I've cleaned the worst of the noise of each track, topped and tailed them, centered the flute in the stereo field, given the piano a little bit of stereo width and re-centered it, brightened the two instruments and given it a little bit of reverb. Sheesh, I'm exhausted!

http://mcgee-flutes.com/Sounds/Seamus%2 ... ison's.mp3

All pretty subjective stuff, so you could certainly play around with it and get a better result.

Now, here's a question. Do we think this track is too sharp, or too fast, or both?

Terry McGee wrote:Now, here's a question. Do we think this track is too sharp, or too fast, or both?

I'm sure Seamus sent the piano player outside to tune.

Heh heh, that can take a while....

My guess it that that track is either a semitone sharp or about 6% fast (which also results in it being about a semitone sharp). But which is it?

Conceivably it could be either.

1. Seamus likes his flute with the slide pushed in, and the piano player played a semitone up. My late partner Gillian used to play the old piano at the Pot Belly in Canberra a semitone up as it could no longer be maintained at concert pitch. Caused havoc with visiting piano players....

2. For some reason the recording machine was running slow, causing the playback at standard speed to be fast and higher pitch.

I like it slowed down the 6%.

Now this raises an interesting topic. What technology do we have to help us pitch a recording like this? The notes are coming pretty fast, so a regular tuner doesn't have long to grab on. We can play a drone note and tune it by ear as best we can to the recording, that that seems a bit slap dash in this day and age. We can do a Fast Fourier Transform (Spectrum analysis), but that's not going to yield much precision either. It would be nice to zero in within a few cents if we could....

Terry McGee wrote:
Now this raises an interesting topic. What technology do we have to help us pitch a recording like this? The notes are coming pretty fast, so a regular tuner doesn't have long to grab on. We can play a drone note and tune it by ear as best we can to the recording, that that seems a bit slap dash in this day and age. We can do a Fast Fourier Transform (Spectrum analysis), but that's not going to yield much precision either. It would be nice to zero in within a few cents if we could....

Check out Melodyne's pitch editing tool: https://www.celemony.com/en/melodyne/wh ... elodyne-do

There's a video demo here, which shows you that Melodyne actually identifies the pitch of each note and allows you to adjust (and of course you can repitch the entire track to a different key): https://youtu.be/J-oQ6VQt8Qw

The full Studio version of Melodyne is expensive, but there's an "essentials" version that's reasonably affordable.

Terry McGee wrote:Now this raises an interesting topic. What technology do we have to help us pitch a recording like this? The notes are coming pretty fast, so a regular tuner doesn't have long to grab on. We can play a drone note and tune it by ear as best we can to the recording, that that seems a bit slap dash in this day and age. We can do a Fast Fourier Transform (Spectrum analysis), but that's not going to yield much precision either. It would be nice to zero in within a few cents if we could....

If the notes are running too fast for a tuner to analyze, you could try slowing down the tempo without altering pitch, using a plugin or DAW function. Then isolate just a few significant notes for analysis. I know those tempo reduction algorithms introduce artifacts and it sounds gnarly, although I do it sometimes when learning tunes. The note fundamental pitch should still be there.

bradhurley wrote:

Terry McGee wrote:
Now this raises an interesting topic. What technology do we have to help us pitch a recording like this? The notes are coming pretty fast, so a regular tuner doesn't have long to grab on. We can play a drone note and tune it by ear as best we can to the recording, that that seems a bit slap dash in this day and age. We can do a Fast Fourier Transform (Spectrum analysis), but that's not going to yield much precision either. It would be nice to zero in within a few cents if we could....

Check out Melodyne's pitch editing tool: https://www.celemony.com/en/melodyne/wh ... elodyne-do

There's a video demo here, which shows you that Melodyne actually identifies the pitch of each note and allows you to adjust (and of course you can repitch the entire track to a different key): https://youtu.be/J-oQ6VQt8Qw

The full Studio version of Melodyne is expensive, but there's an "essentials" version that's reasonably affordable.

Ah, very interesting. So that's a definite way forward.

And it reminded me of what had been lurking at the back of my brain (or perhaps had fallen down behind the lathe?). Tartini, a sound investigation tool designed by the University of Otago in New Zealand offers the same style of "language lab" pitch analysis. So I dug out an old copy I had and tried it on the recording in question. It allows me to drag the music past the cursor at any speed I want while watching the pitch meter. Pretty much confirms what I had worked out by ear - that the recording is about a semitone sharp of the usual pitch of those two tunes.

Now the Tartini (or Melodyne) analysis also allows us to compare the pitch of notes played by flute and piano (providing they are in different octaves, which they are). So it could probably could yield some hints about whether the flute is a normal D flute sharpened to nearly Eb (where we would expect to see some inaccuracies in scale), or an Eb flute. I don't think it can tell us for sure about the speed of the recording though. More thought needed. I'd better go make some flutes....

Conical bore wrote:

Terry McGee wrote:Now this raises an interesting topic. What technology do we have to help us pitch a recording like this? The notes are coming pretty fast, so a regular tuner doesn't have long to grab on. We can play a drone note and tune it by ear as best we can to the recording, that that seems a bit slap dash in this day and age. We can do a Fast Fourier Transform (Spectrum analysis), but that's not going to yield much precision either. It would be nice to zero in within a few cents if we could....

If the notes are running too fast for a tuner to analyze, you could try slowing down the tempo without altering pitch, using a plugin or DAW function. Then isolate just a few significant notes for analysis. I know those tempo reduction algorithms introduce artifacts and it sounds gnarly, although I do it sometimes when learning tunes. The note fundamental pitch should still be there.

Good thinking, but it reveals another issue - the tuner gets a bit confused with multiple notes at once. So that approach would probably work fine with just the flute, but add the accompaniment and it gets messy. But the tempo change would probably assist the ear in coming to a conclusion, so it's a trick to keep in mind.

Seamus Tansey : "Miss Monaghan" - [ warning ! - this recording contains "tambourine" ! ]
https://youtu.be/ucDJ88dI4hA

Roger Sherlock : "Morrisons / The Trip To Sligo" :
https://youtu.be/NpQ8kEY9oq8

Seamus Tansey : "The Cliffs Of Moher / Paddy Fahy's Jig" : https://youtu.be/5oZT4RZ6w68

Roger Sherlock : "Fred Finn's / The Sailor On The Rock / The Old Pensioner". : https://youtu.be/yz7oD05608A

P.J. Crotty playing a relatively obscure 4-part reel "O'Connell's Farewell To Dublin" : Willie Clancy Week 1980.
The only other musician I have ever heard play this tune - in nearly 40 years now - was John Skelton, who recorded a slow-ish version on his first solo recording "One At A Time".
https://youtu.be/-crBWSvmaKg

P.J. Crotty playing a relatively obscure 4-part reel "O'Connell's Farewell To Dublin" : Willie Clancy Week 1980.

That's a lovely one Kenny. {---goes off to learn it}

As we say in Scotland, " is it no, jist" ?

PS - I did post it on "thesession.org" - [ but that would be cheating ].