Chiff and Fipple Forums

At last, a system for measuring the tuning of your playing in reel time (yes, sorry, pun intended). Find out all about it, download it, and use it completely free of cost. Start at:

http://www.mcgee-flutes.com/RTTA.htm

Even if you have no intention of using the system, you'll probably find some of the information about flute tuning interesting. I've started a separate thread Historical Flute Syndromes to give us a place to discuss this sort of stuff that isn't directly related to the analysis system.

And yes, this is the same system we were talking about in our previous discussion, Bulletproofing the Messenger. I think the system is sufficiently tested now to withstand public scrutiny. But you'll be the judge of that!

If you find fault with the web page, please email me. If you find problems with the RTTA software, email Graeme.

Terry

Pretty cool stuff. I'll try to get it running tonight. I always wanted something like that...so, thanks in advance!

I downloaded everything, but can't seem to get going.

I'm running Mac OS 10.4.1. Anybody else running into trouble? I'm thinking I might need an earlier version of R. I downloaded R-2.6.2.dmg, which says it is for OS 10.4.4 or higher. Is that my problem? I guess I could update my OS also.

I've gotten as far as saving a .txt file in the same folder as my polygraph.r and polygraph.bat, but when I double click polygraph.bat, I'm lost.

Michael

I have it running on a Mac OS 10.5.2 but 10.4 should be similar.

Look on Terry's web page there are some notes from me on how to get it to run.

You do not use polygraph.bat
You use the terminal to run Polygraph.r with R in the 'command line mode'

To add to the notes on Terry's site:

Once you get it running - In the 'Shell' menu (of Terminal) choose new command and type in 'R --no-save <Polygraph.r' and check the 'run in shell' box. Now terminal will remember this command and you do not have to keep typing it in.

If you still have questions post again and I'll try to help.

Bill

mahanpots wrote:I downloaded everything, but can't seem to get going.

I'm running Mac OS 10.4.1. Anybody else running into trouble? I'm thinking I might need an earlier version of R. I downloaded R-2.6.2.dmg, which says it is for OS 10.4.4 or higher. Is that my problem? I guess I could update my OS also.

I've gotten as far as saving a .txt file in the same folder as my polygraph.r and polygraph.bat, but when I double click polygraph.bat, I'm lost.

Michael

Hi Michael
If R-2.6.2 is for version MacOS 10.4.4 and you're running MacOS 10.4.1 that could certainly cause R not to work! The Polygraph.r script I've written will very probably run on earlier versions of R. So see if you can find a version of R the runs on MacOS 10.4.1 Then check R starts before trying to run Polygraph.r. Bill has described how to run it on a Mac, and I'll edit the instructions on Terry's website to make it clearer than it is now.

Cheers
Graeme

Does anyone have a sample Tartini-generated rtest.txt file they care to post for download, for simply testing the R/Polygraph installation without first generating our own data?

Thanks!

That seems like a good idea. Graeme, what about slipping one into basic.zip? I could amend the setup page to mention it.

Terry

MTGuru wrote:Does anyone have a sample Tartini-generated rtest.txt file they care to post for download, for simply testing the R/Polygraph installation without first generating our own data?

Thanks!

Done it. You can get to it from the FAQ page here:
http://www.business.otago.ac.nz/infosci ... TA/FAQ.htm

It's the one of me playing the Monzani flute as used on the page I wrote on using it here
http://www.mcgee-flutes.com/RTTA/using.htm

Cheers
Graeme

OK, so you can save that as rtest.txt into the directory you are keeping all the other suff by right-clicking and choosing "save target as ...".

If you left click on it, you get to see the vast amount of data that Tartini is handing to R. Scary. Best viewed through a half-emptied bottle of Jamisons.

Terry

Terry McGee wrote:Best viewed through a half-emptied bottle of Jamisons.

Terry

The reason that bottle is half empty is because I'd just drunk it when Terry asked me to get into this. Now that I've sobered up I'm going to try out suggestion number 11 on possible things to do with this software, at the foot of this page:
http://www.mcgee-flutes.com/RTTA/using.htm

Just kidding folks
Graeme

groxburgh wrote:

MTGuru wrote:Does anyone have a sample Tartini-generated rtest.txt file they care to post for download, for simply testing the R/Polygraph installation without first generating our own data?

Done it. You can get to it from the FAQ page here:
http://www.business.otago.ac.nz/infosci ... TA/FAQ.htm

Thank you, gentlemen!

I'm having trouble getting the Windows installation to work properly with the batch file because of weird pathname and environment stuff.

BUT ... I can successfully generate the graph by copying Polygraph.r and rtest.txt to the bin directory, and from there running:

R CMD BATCH Polygraph.r

The result is a nice Polygraph.jpg graph, and a Polygraph.r.Rout logfile.

I'll post more if I get the other working. Cheers!

I see I'm going to have to get my son over here to help me understand some of this.
I installed earlier version of R, and I have no idea if it's up and running. I can open up the program and I get a window starts off with:

R : Copyright 2005, The R Foundation for Statistical Computing
Version 2.1.1 (2005-06-20), ISBN 3-900051-07-0

I can type in the window. Does this mean I am in the 'command line mode?'

I opened a clip of myself playing flute in Tartini and exported it as "Rtest.txt" and placed it in the same folder as Polygraph.r and Polygraph.bat.

Highwood says that once I get R running, "In the 'Shell' menu (of Terminal) choose new command and type in 'R --no-save <Polygraph.r' and check the 'run in shell' box. Now terminal will remember this command and you do not have to keep typing it in.

So, I typed 'R --no-save <Polygraph.r' Nothing happened. I can't see anything identifying itself as a 'run in shell' box when I place my cursor on it, so I hit my return key. The following appears below what I typed:
[1] "R --no-save <Polygraph.r"
Still nothing.

Sorry about the ignorance. I'm sure my son could figure it out. He graduated from Duke in Computer Science.

Michael

From reading this post and part of your historical flute syndrome piece, am I correct in guessing that this is a piece of software into which one can load a sample of recorded music that the software will turn into a chart of all the notes it contains, correctly analysed for pitch?

Great stuff, guys! This will be tons of fun to play with... I got tired of renaming files to "rtest.txt" and then renaming the jpg, so here's a quick patch that allows you to give the batch file a file name on the command line, e.g., open up a command shell and type

polygraph my-file.txt

The jpg file is named "Polygraph-my-file.jpg."

Patch for Polygraph.bat:

***Polygraph-orig.bat Thu Apr 17 08:43:00 2008
---Polygraph.bat Fri Apr 18 22:29:16 2008
***************
*** 1,3 ****
! "C:\Program Files\R\R-2.6.2\bin\R.exe" --no-save < Polygraph.r 440 5 boxplot 10 1.5 equal
rem
rem For use with version 1.0.x
--- 1,3 ----
! "C:\Program Files\R\R-2.6.2\bin\R.exe" --no-save < Polygraph.r 440 5 boxplot 10 1.5 equal %1
rem
rem For use with version 1.0.x

Patch for Polygraph.r:

*** Polygraph-orig.r Thu Apr 17 14:23:00 2008
--- Polygraph.r Fri Apr 18 22:49:19 2008
***************
*** 21,30 ****
args[6] <- 10
args[7] <- 1.5
args[8] <- "equal"
# test to see if we've got the right number of command line Arguments and if we have use them
# we just hope they're all in the correct order
testArgs <- commandArgs()
! if (length(testArgs) == 8){args <- commandArgs()}
length (commandArgs)
# [3] is pitch shift (real),
# [4] filtering diff witdth (real),
--- 21,31 ----
args[6] <- 10
args[7] <- 1.5
args[8] <- "equal"
+ args[9] <- "rtest.txt"
# test to see if we've got the right number of command line Arguments and if we have use them
# we just hope they're all in the correct order
testArgs <- commandArgs()
! if (length(testArgs) == 9){args <- commandArgs()}
length (commandArgs)
# [3] is pitch shift (real),
# [4] filtering diff witdth (real),
***************
*** 32,37 ****
--- 33,39 ----
# [6] is min number of data points for note to be included in graph (integer)
# [7] is h in vioplot which determines the smoothing - around 1.5 looks good (real)
# [8] is temperament either equal or justD
+ # [9] is data file name
#############################
# What temperament are we plotting against
# set equal temperament as the default
***************
*** 40,46 ****

#############################
# read data from file "rtest.txt" in current working directory
! rawdata <- read.table("rtest.txt", header=TRUE)

#############################
# Trying some Filtering using value of cents from command line argument
--- 42,48 ----

#############################
# read data from file "rtest.txt" in current working directory
! rawdata <- read.table(args[9], header=TRUE)

#############################
# Trying some Filtering using value of cents from command line argument
***************
*** 129,135 ****

#############################
# output graph to JPEG file rather than within R console
! jpeg("Polygraph.jpg", width = 800, height = 600, units = "px",)


#############################
--- 131,138 ----

#############################
# output graph to JPEG file rather than within R console
! outfile <- paste("Polygraph",unlist(strsplit(args[9],"\\."))[1],".jpg",sep="")
! jpeg(outfile, width = 800, height = 600, units = "px",)


#############################

Yes thanks for this idea, I'll look into it, probably a good plan. For anyone with further ideas like this can you please email the details to me, my email is on Terry's web site. Use this forum to outline what you've tried, but maybe posting all the details here is not the best. Details can scare some people. The same might apply if you've got problems getting it running. Then again it's up to you.

Cheers
Graeme