Re: [Finale] Tempo Detection in Audacity
On 27 Sep 2009 at 2:55, Christopher Smith wrote: > http://handyprograms.blogspot.com/2008/08/metronome.html For some reason, it's not compatible with my Java installation and won't run. I think it's likely not the fault of the app -- my Java installation has been broken through several different versions of the JRE. I keep uninstalling and re-installing, but nothing seems to fix the problem. No great loss -- I don't believe Java has any place client-side in the first place, and find Java apps to be hopelessly slow and nearly unusable. -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On 27 Sep 2009 at 10:21, Christopher Smith wrote: > I imagine (perhaps wrongly!) > that the tempo might stay fairly solid for each section, but change > from variation to variation, maybe with a connection to what kinds of > subdivisions are in the top parts. David F., I hope you will point us > to a link with your findings once you have worked it all out. I'm not very far into to the project yet, and have mostly listened carefully only to the earlier recordings (1960s, 1970s). I had thought that the idea of Baroque style was fairly inflexible back then, but what I'm finding is that this is just not true. The earliest recording I've analyzed, by Baumgartner (the second recording of the piece ever, and the one that probably brought the piece to the attention of those who made the recording that ended up as a breakout hit, c. 1970) is actually quite "interesting" from the standpoint of tempo. And I mean "interesting" in the sense of the old Chinese curse! It's average tempo is SLOW AS MOLASSES. Even with cuts of 10 bars, it is still the longest recording of any in the collection I'm looking at (I've currently collected 20 of the 21 I'm sure I'm going to evaluate, and am working on a 22nd; this is not by any means *all* of the recordings of the original, and I'm mostly ignoring arrangements, though not entirely). The average BPM is 36 (the fastest I've calculated, without listening yet, is 72), and it's just awful, awful, awful. The famous recording that launched the Pachelbel juggernaught was the EMI recording by Paillard that was released in the US by Musical Heritage Society, and that recording has an almost equally slow average tempo (37BPM), but without any of the wheezing and swaying found in the Baumgartner. Indeed, the overall sound conception and approach is a rather modern and clean, unfussy one -- dare I say it, but French? Anyway, it's been fascinating. I've made five posts on the subject so far, one an overview, and then four posts on the first four recordings I've listened to. I hope you enjoy reading. I'd certainly like feedback -- though my blog has no comments, you can email me via the link on the blog (or via this email address). -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
RE: [Finale] Tempo Detection in Audacity
On 27 Sep 2009 at 12:14, Peter Taylor wrote: > << From: David W. Fenton > > My play project for the weekend is to analyze the tempos of various > recordings of the Pachelbel Canon >> > > 8-\ Sad or what? It's a great piece of music, so it's not sad at all. And I've been quite surprised that it's quite easy to get interested in the recordings, because there are a lot more interesting things going on in them than you might expect. I'm blogging the experience, in fact, at my blog: http://dfenton.com/NoComment/ So far, I've evaluated the earliest recordings I've been able to get my hands on (the earliest is not available for MP3 download, and the CD won't arrive until next week). I now have before me to analyze a string of recordings from the 70s and 80s that have been widely anthologized, and I'm anticipating this will be the low point, before I then get to the 80s and the Early Music recordings. I've actually been quite surprised at the things I've learned, and couldn't wait to get back to it today when I got my "real" work done! -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On Sun Sep 27, at SundaySep 27 7:36 AM, dhbailey wrote: tempo analysis is very important to help figure out what one feels is the proper interpretation. Listening to works which have been widely recorded can be very helpful in that regard. Some recordings of any work seem to succeed better than others, and quite often it's the difference in tempo which makes all the difference. Actually, this sounds fascinating to me. I imagine (perhaps wrongly!) that the tempo might stay fairly solid for each section, but change from variation to variation, maybe with a connection to what kinds of subdivisions are in the top parts. David F., I hope you will point us to a link with your findings once you have worked it all out. One of my colleagues did his master's thesis in analysing the swing feel of some jazz recordings. His focus was submetric (rather than tempo, so not exactly the same thing) but he related the variations in swing feel to tempo as well. It certainly didn't take him a weekend! I can't help but think that modern digital audio tools would have helped him in speed and accuracy, though. Christopher ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
Peter Taylor wrote: << From: David W. Fenton My play project for the weekend is to analyze the tempos of various recordings of the Pachelbel Canon >> 8-\ Sad or what? I don't think so -- tempo analysis is very important to help figure out what one feels is the proper interpretation. Listening to works which have been widely recorded can be very helpful in that regard. Some recordings of any work seem to succeed better than others, and quite often it's the difference in tempo which makes all the difference. -- David H. Bailey dhbai...@davidbaileymusicstudio.com ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
RE: [Finale] Tempo Detection in Audacity
<< From: David W. Fenton My play project for the weekend is to analyze the tempos of various recordings of the Pachelbel Canon >> 8-\ Sad or what? ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
David, It might not be as bad as you think. Try this free online metronome (and a standalone applet that runs in your browser) where you can tap along with a tune and have the tempo of ANY two beats displayed. If you are not steady enough for two beats, you can take an average over the duration of a passage. You can also note different tempos in different sections. http://handyprograms.blogspot.com/2008/08/metronome.html Christopher On Sat Sep 26, at SaturdaySep 26 6:37 PM, David W. Fenton wrote: On 26 Sep 2009 at 16:44, John Blane wrote: Don't know Audacity well enough but Pro Tools can do what you're looking for. If you want to tap along to determine tempo there is a plug-in for itunes available. But this really wouldn't solve much of the problem at all. Sure tapping is a *little* easier than using a stopwatch, but not by much. I'm not interested in investing in ProTools -- I wouldn't use it often enough for it to be justifiable. But that facility is already built into Finale. Look in the Hyperscribe tool and choose Playback or Click from the beat source menu Er, with MP3 files? Why, exactly, would I have 16 different versions of the same piece in Finale, differing only in tempo? I think you quite obviously missed the context of my question. -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
RE: [Finale] Tempo Detection in Audacity
Dear David, I do a lot of performance and tempo analysis. It's not going to work so well with the beat-finder plugin in Audacity, as you've discovered. In terms of free computerised tools, you could have a look at this site: http://www.charm.rhul.ac.uk/analysing/p9_0_1.html and download Sonic Visualiser. Another option, very similar but far more basic is available from http://www.soton.ac.uk/~musicbox/charm5.html. It's a simple command line utility which measures the space between taps, and there's also a related excel spreadsheet which will graph the tempo fluctuations. It has detailed instructions, though it's not difficult to work out. One could also download the trial of Acid Music Studio or Acid Music Pro (Sony website), which will run fairly reasonable beat detection (even though it's designed for popular musics), and allow you to make finer adjustments to the detection of onset events visually and aurally by dragging/nudging. I've also used a metronome with a tempo tap feature. It gives a very good overall opinion quickly. Cheers, Angela Angela Turner FinePrint, Music Printing & Typesetting Website: www.fine-print.com.au Email: ang...@fine-print.com.au Mobile: 0402 84 85 31 -Original Message- From: finale-boun...@shsu.edu [mailto:finale-boun...@shsu.edu] On Behalf Of David W. Fenton Sent: Sunday, 27 September 2009 6:11 AM To: Finale@shsu.edu Subject: [Finale] Tempo Detection in Audacity My play project for the weekend is to analyze the tempos of various recordings of the Pachelbel Canon, and I figured I'd be able to do it with a plugin for Audacity. I'm finding, however, that there are very few beat detection plugins that are useful. I want one that will analyze a selection of the waveform, but I can't find anything like that. All the standalone apps for this autodetect from the first few seconds and if there's too much silence at the beginning (as there is for many of MP3s involved here), the detect nothing at all. And even for the ones that "work" the numbers are just wrong. It's quite clear to me that the available tools I found are not designed to deal with Classical music. Does anyone have any ideas what to do here? I have a Beat Finder plugin in the ANALYZE menu, but when I run it, I get a track at the bottom showing tempo change boundaries, but I can't seem to get any information about the actual beats or their BPMs, or anything else. Is there a plugin that will do what I want, i.e., analyze a passage and give me an average BPM? Or am I stuck doing this manually (ugh!)? -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
Ok, David, I won't take your bait and I'll let you have the last word - you always do. Simply this: I understood what you were asking about. Unaware that an Audacity plug-in exists that would accomplish what you needed, I was trying to be helpful by offering you some other options that perhaps would suit your needs. Instead of realizing my intention to help, whether it did or not, you choose to be your notoriously mean- spirited and unnecessarily antagonistic self. This is how you choose to respond to someone trying to help you? Isn't this list a community of musicians and Finale users and aren't we here to learn from each other and help when we can? I'm content to just letting you keep your head up your ass. For years, I've always wished that you would get married so you would have someone you could constantly berate and argue with and spare us the bandwidth. Another wish unfulfilled. sigh On Sep 26, 2009, at 6:16 PM, David W. Fenton wrote: On 26 Sep 2009 at 17:49, John Blane wrote: Um, I think you quite obviously missed the context of my trying to assist you. The context was the thread launched by my post, so that's the context I was using. You appear to have been using a different context. The fact that I didn't detect that is lamentable, but hardly my fault, given how little of what you say you intended was not explicitly stated in your response. On the other hand, I was pretty explicit about what I was looking for, a plugin for Audacity, or a standalone piece of software that would do the job. I never suggested anything about inputting them into Finale at all. You can determine the tempo with Finale while you use any other source for playback. Get it? Well, I've been doing that already, in fact. If your words had been clear that this was what you were suggesting, I'm pretty sure I would have caught it, given that I've already been doing it. Doing it that way is actually pretty unwieldy, as a matter of fact, as well as highly inaccurate, except for the average tempo of the whole piece. It's certainly easy enough to calculate the timing for, say, 6 measures of the Pachelbel in Finale, but selecting those same 6 measures in the waveform of a recording is going to be highly inaccurate. And once I know the timing on those 6 measures, why would I then need to use Finale? I already know how many beats it is, and I know how long it took, so I can use simple arithmetic to figure out what the average BPM in the passage is. And, of course, even nicer would be the ability to see the *range* of tempos within a passage, and short of doing that measure-by-measure, I don't see any worthwhile way to do that via manual methods. That's why a tool that calculates the BPM from a waveform selection would be so useful. And I wasn't suggesting that you invest in Pro Tools to achieve what you're looking for only to point out that you are not the first person that needs to determine this type of information. That, too, was not clear from what you wrote. You said this: Don't know Audacity well enough but Pro Tools can do what you're looking for. That seems to me like a recommendation of a piece of software that can do what I want, not so much as an indication that other people need this, too (though it *is* that implicitly). Had you said: You're actually not the only one who needs this, since Pro Tools can already do what you're looking for. And, of course, it would remain to be seen whether or not ProTools could actually detect BPM accurately enough from classical music textures to be useful for what I'm investigating. Apparently, you think it should be free or you'll complain about it. ??? I asked for an Audacity plugin. That pretty much implies I'm looking for something free or inexpensive, don't you think? I'm certainly not looking to acquire a completely different application that does way more than the simple task I'm asking about. Indeed, I didn't ask for free or pay, I asked for a plugin. Now, a $25 plugin might be worth it, while a $250 piece of software that I don't really need seems like overkill to me. In the end, I think your message was pretty elliptical in regard to what you say you were intending to convey. I don't know how I possibly could have interpreted it the way you had in mind -- all I had to go on was your words, and they didn't explicitly convey major parts of what you say you intended. In short, you brought a context in your head that was not the same as the context in the words of the discussion that came before your post. I'm sorry I misinterpreted your words, but taken by themselves, they really did seem to me to be pretty non-responsive to what I'd asked about. -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On 26 Sep 2009 at 17:49, John Blane wrote: > Um, I think you quite obviously missed the context of my trying to > assist you. The context was the thread launched by my post, so that's the context I was using. You appear to have been using a different context. The fact that I didn't detect that is lamentable, but hardly my fault, given how little of what you say you intended was not explicitly stated in your response. On the other hand, I was pretty explicit about what I was looking for, a plugin for Audacity, or a standalone piece of software that would do the job. > I never suggested anything about inputting them into Finale at all. > You can determine the tempo with Finale while you use any other > source for playback. Get it? Well, I've been doing that already, in fact. If your words had been clear that this was what you were suggesting, I'm pretty sure I would have caught it, given that I've already been doing it. Doing it that way is actually pretty unwieldy, as a matter of fact, as well as highly inaccurate, except for the average tempo of the whole piece. It's certainly easy enough to calculate the timing for, say, 6 measures of the Pachelbel in Finale, but selecting those same 6 measures in the waveform of a recording is going to be highly inaccurate. And once I know the timing on those 6 measures, why would I then need to use Finale? I already know how many beats it is, and I know how long it took, so I can use simple arithmetic to figure out what the average BPM in the passage is. And, of course, even nicer would be the ability to see the *range* of tempos within a passage, and short of doing that measure-by-measure, I don't see any worthwhile way to do that via manual methods. That's why a tool that calculates the BPM from a waveform selection would be so useful. > And I wasn't suggesting that you invest > in Pro Tools to achieve what you're looking for only to point out > that you are not the first person that needs to determine this type > of information. That, too, was not clear from what you wrote. You said this: > > Don't know Audacity well enough but Pro Tools can do what > > you're looking for. That seems to me like a recommendation of a piece of software that can do what I want, not so much as an indication that other people need this, too (though it *is* that implicitly). Had you said: > > You're actually not the only one who needs this, since Pro > > Tools can already do what you're looking for. And, of course, it would remain to be seen whether or not ProTools could actually detect BPM accurately enough from classical music textures to be useful for what I'm investigating. > Apparently, you think it should be free or you'll > complain about it. ??? I asked for an Audacity plugin. That pretty much implies I'm looking for something free or inexpensive, don't you think? I'm certainly not looking to acquire a completely different application that does way more than the simple task I'm asking about. Indeed, I didn't ask for free or pay, I asked for a plugin. Now, a $25 plugin might be worth it, while a $250 piece of software that I don't really need seems like overkill to me. In the end, I think your message was pretty elliptical in regard to what you say you were intending to convey. I don't know how I possibly could have interpreted it the way you had in mind -- all I had to go on was your words, and they didn't explicitly convey major parts of what you say you intended. In short, you brought a context in your head that was not the same as the context in the words of the discussion that came before your post. I'm sorry I misinterpreted your words, but taken by themselves, they really did seem to me to be pretty non-responsive to what I'd asked about. -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
Um, I think you quite obviously missed the context of my trying to assist you. I never suggested anything about inputting them into Finale at all. You can determine the tempo with Finale while you use any other source for playback. Get it? And I wasn't suggesting that you invest in Pro Tools to achieve what you're looking for only to point out that you are not the first person that needs to determine this type of information. Apparently, you think it should be free or you'll complain about it. Good luck On Sep 26, 2009, at 5:37 PM, David W. Fenton wrote: Er, with MP3 files? Why, exactly, would I have 16 different versions of the same piece in Finale, differing only in tempo? I think you quite obviously missed the context of my question. ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On 26 Sep 2009 at 16:44, John Blane wrote: > Don't know Audacity well enough but Pro Tools can do what you're > looking for. If you want to tap along to determine tempo there is a > plug-in for itunes available. But this really wouldn't solve much of the problem at all. Sure tapping is a *little* easier than using a stopwatch, but not by much. I'm not interested in investing in ProTools -- I wouldn't use it often enough for it to be justifiable. > But that facility is already built into Finale. Look in the > Hyperscribe tool and choose Playback or Click from the beat source menu Er, with MP3 files? Why, exactly, would I have 16 different versions of the same piece in Finale, differing only in tempo? I think you quite obviously missed the context of my question. -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
Don't know Audacity well enough but Pro Tools can do what you're looking for. If you want to tap along to determine tempo there is a plug-in for itunes available. But that facility is already built into Finale. Look in the Hyperscribe tool and choose Playback or Click from the beat source menu On Sep 26, 2009, at 4:15 PM, David W. Fenton wrote: On 26 Sep 2009 at 16:49, Aaron Sherber wrote: On 9/26/2009 4:11 PM, David W. Fenton wrote: Or am I stuck doing this manually (ugh!)? I know this isn't what you were asking, but why is doing it manually so tough? If you have a stopwatch and Excel, you can get the metronome timings just by counting off a few beats and plugging the duration into a formula. Have you ever tried that? On 16 separate recordings? It's not nearly as easy as you think, and the result depends on which part of the piece you listen to (since tempos are not uniform from beginning to end). Right now as a preliminary step, I'm getting the timings of the tracks with the silence left out, and then using that to calculate the average tempo (since the number of beats is the same in all performances, at least for the complete ones -- that has turned out to be an eye-opener, as some of these light classical albums available for MP3 download on Amazon have truncated versions, with fade-outs at the end, even). But since tempos vary throughout the piece, sometimes by as much as 6- 10bpm, I will still need to go back and evaluate sections. If I had an automated way to do this, it would be *much* easier. What I envision would be a tool that would analyze the entire piece and show me points where the tempo has changed sufficient to be a new tempo (a nice tool would allow me to set the sensitivity of this detection), and would indicate somehow the average tempo in each of these sections. I'm beginning to suspect this is just something nobody else is interested in doing, so no one has created a plug-in to do it, and thus I'll be left doing it manually. *sigh* -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On 26 Sep 2009 at 16:49, Aaron Sherber wrote: > On 9/26/2009 4:11 PM, David W. Fenton wrote: > > Or am I stuck doing this manually (ugh!)? > > I know this isn't what you were asking, but why is doing it manually so > tough? If you have a stopwatch and Excel, you can get the metronome > timings just by counting off a few beats and plugging the duration into > a formula. Have you ever tried that? On 16 separate recordings? It's not nearly as easy as you think, and the result depends on which part of the piece you listen to (since tempos are not uniform from beginning to end). Right now as a preliminary step, I'm getting the timings of the tracks with the silence left out, and then using that to calculate the average tempo (since the number of beats is the same in all performances, at least for the complete ones -- that has turned out to be an eye-opener, as some of these light classical albums available for MP3 download on Amazon have truncated versions, with fade-outs at the end, even). But since tempos vary throughout the piece, sometimes by as much as 6- 10bpm, I will still need to go back and evaluate sections. If I had an automated way to do this, it would be *much* easier. What I envision would be a tool that would analyze the entire piece and show me points where the tempo has changed sufficient to be a new tempo (a nice tool would allow me to set the sensitivity of this detection), and would indicate somehow the average tempo in each of these sections. I'm beginning to suspect this is just something nobody else is interested in doing, so no one has created a plug-in to do it, and thus I'll be left doing it manually. *sigh* -- David W. Fentonhttp://dfenton.com David Fenton Associates http://dfenton.com/DFA/ ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
Re: [Finale] Tempo Detection in Audacity
On 9/26/2009 4:11 PM, David W. Fenton wrote: Or am I stuck doing this manually (ugh!)? I know this isn't what you were asking, but why is doing it manually so tough? If you have a stopwatch and Excel, you can get the metronome timings just by counting off a few beats and plugging the duration into a formula. Aaron. ___ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale