Home |
Search |
Today's Posts |
#1
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Hey all,
I have an interesting question. I hope someone can help. Is there a hard ware software product out there that can help with recognizing the proper speed/pitch of a voice? I am working on a record transfer project for a client of mine. All spoken word sermons given by the same person. They have been recorded over a period of about 15 years from 1951 on. The problem is this; They are saying that the recorded speed of the original master that the record was cut from isn't quite the right pitch. Some are slower & some are faster depending on the gear used at the time. I have set up the turntable with a strobe disc to ensure correct playback speed stated on the record for transfer, but they want Her to have the same sound/pitch of her voice across all titles. Is there something I can use to detect the "correct" speed? They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. BTW there are over 900 titles to be done. Thanks to all - Tom -- Tom Jancauskas Imedia |
#2
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
I have an interesting question. I hope someone can help. Is there a hard ware software product out there that can help with recognizing the proper speed/pitch of a voice? Not really. The problem is this; They are saying that the recorded speed of the original master that the record was cut from isn't quite the right pitch. Some are slower & some are faster depending on the gear used at the time. Yes, this happens. And some of them may start out sharp and wind up flat by the end of the disc. I have set up the turntable with a strobe disc to ensure correct playback speed stated on the record for transfer, but they want Her to have the same sound/pitch of her voice across all titles. Is there something I can use to detect the "correct" speed? Not really. They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. Who are "they" and do they have perfect pitch and are they willing to help out? BTW there are over 900 titles to be done. If they are all on the same media recorded on the same equipment you may be able to save some time. You give them the choice of the rough unattended transfer or the carefully attended transfer.... needless to say there is an order of magnitude or two difference in price between them. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#3
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"Tom Jancauskas" wrote in message ... Hey all, I have an interesting question. I hope someone can help. Is there a hard ware software product out there that can help with recognizing the proper speed/pitch of a voice? I am working on a record transfer project for a client of mine. All spoken word sermons given by the same person. They have been recorded over a period of about 15 years from 1951 on. The problem is this; They are saying that the recorded speed of the original master that the record was cut from isn't quite the right pitch. Some are slower & some are faster depending on the gear used at the time. I have set up the turntable with a strobe disc to ensure correct playback speed stated on the record for transfer, but they want Her to have the same sound/pitch of her voice across all titles. Is there something I can use to detect the "correct" speed? They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. People who transcribe historical recordings do this all the time by measuring the frequency of a known note in the recording, using FFT spectrum analysis features built into most modern audio editing software. |
#4
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
#6
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
I figured as much. I wanted to make sure there wasn't something I missed. I don't have any problem with the speed variation, as it sounds like the person that I have heard all along. This subject has just come up after transferring 400 of the titles. Apparently it was a "known issue" that was never brought up until now. Think of this as a way to bill more money. They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. Who are "they" and do they have perfect pitch and are they willing to help out? "they" are the the contact person from this church who is in charge of the archive, and a board of directors that doesn't quite know anything audio, but one particular person on the board HAS to be involved even though they shouldn't be. Well, if they want to go back and redo the existing transfers and they have a few tens of thousands of dollars to go back and do it carefully and individually for hundreds of discs, give them that option. As far as perfect pitch - I don't know. The "rules of the game" have changed many times for this project. Between wanting many different output formats, changing their minds about which ones, first asking for a straight transfer, then changing to wanting ALL the noise removed, then wanting the pitch corrected.... It just goes on& on. I am getting paid - but to have to go back & redo the entire catalog each time a "decision" is made.... That's fine. Bill by the hour and this sort of customer can be your biggest friend. Pennywise and pound foolish. They want it done quickly (of course) and if this guy was involved more - it would take FOREVER! They don't even want to spend the time/money on a proper record cleaning! Good money after bad I guess. If they want a noisy transfer from a dirty record, that's fine. They'll want to go back and do it again later at $250/crack. As long as you can bill the time, do whatever they want. Anyway, I thought I'd just have to do it by ear. Which is fine. The one good thing you can do is look for 60 Hz hum on the original and lock to that. That can sometimes be a real timesaver. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#7
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
in article , Arny Krueger at wrote on 4/19/07 12:58 PM: Is there something I can use to detect the "correct" speed? They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. People who transcribe historical recordings do this all the time by measuring the frequency of a known note in the recording, using FFT spectrum analysis features built into most modern audio editing software. I'll have to try that. I have FFT in the program I use. Thanks You might get 'lucky' and have some 60Hz (or 50Hz, depending on the part of the world you're in) hum in the recordings. That would give you a pretty air-tight reference. jak |
#8
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Scott Dorsey wrote:
snip The one good thing you can do is look for 60 Hz hum on the original and lock to that. That can sometimes be a real timesaver. --scott Oops...should have read to the end of the thread before I gave the same advice..... jak |
#9
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
#10
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
#11
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
I know to bill 'em (and can) for all of the time I am putting in , but I can't help to think they could be saving so much money to do it right the first time... All you can do is tell them that. My philosophy is to set my rates higher than any of the competition and then to make sure the customer gets what he's paying for. It sure sounds like he is getting more than they paid for in this case. Brilliant idea on the 60hz hum. There is some on the master. I'll have to check that. Thanks! There is a process from Plangent Technologies that can find all sorts of parasitic noises and lock to them to reduce wow and flutter. It is very expensive but is pretty amazing. Barring that there is always a frequency counter and a parametric eq. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#12
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"Scott Dorsey" wrote in message ... Tom Jancauskas wrote: I know to bill 'em (and can) for all of the time I am putting in , but I can't help to think they could be saving so much money to do it right the first time... All you can do is tell them that. My philosophy is to set my rates higher than any of the competition and then to make sure the customer gets what he's paying for. It sure sounds like he is getting more than they paid for in this case. Brilliant idea on the 60hz hum. There is some on the master. I'll have to check that. Thanks! There is a process from Plangent Technologies that can find all sorts of parasitic noises and lock to them to reduce wow and flutter. That only makes sense if the speed error is varying. A steady-state speed errror such as the OP seems to be describing is not. It is very expensive but is pretty amazing. Barring that there is always a frequency counter and a parametric eq. That gets rid of the hum, but not the speed error. Here's how to correct a speed error in the digital domain: Using a program that allows you to adjust the sample rate in the header of a file without resampling it, change the sample rate by the desired amount. Then resample the file back to its original sample rate. I know that Audition/CEP will do this, and I believe that Audacity will do it as well. With other software, YMMV. For example if you take a file that is sampled at 44,100 Hz and adjust the sample rate in its header to 88,200 Hz, and then resample it back to 44,100 Hz, a 1.00 KHz tone in the file will now have a frequency of 2.00 KHz. This process generally runs as fast and produces as few artifacts as any. Some programs encapsulate it in a function that changes pitch and tempo together. |
#13
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"Tom Jancauskas" wrote in message
... I am working on a record transfer project for a client of mine. All spoken word sermons given by the same person. They have been recorded over a period of about 15 years from 1951 on. The problem is this; They are saying that the recorded speed of the original master that the record was cut from isn't quite the right pitch. Some are slower & some are faster depending on the gear used at the time. I have set up the turntable with a strobe disc to ensure correct playback speed stated on the record for transfer, but they want Her to have the same sound/pitch of her voice across all titles. Is there something I can use to detect the "correct" speed? Do you have Adobe Audition or something similar? If so, do a frequency analysis of a quiet spot and look for 60Hz hum. If it's not 60Hz correct the speed until it is. That's not a perfect method by any means (it depends on where the hum came from and on the constancy of the 60Hz signal from the power company) but it might give you a starting point. It's also worth remembering, though, that voices tend to get lower pitched as we age, so maybe they *shouldn't* all sound the same. Peace, Paul |
#14
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
i myself would use native instruments traktor. its a dj program, but
the timestretching is very nice on a good computer. you can change the tempo without changing the pitch, and you can change the pitch a certain amount and lock it to that pitch and then slow it down and speed it up. there are probably other programs that have good qualities such as this, but the only one ive used is traktor. im sure someone else will respond soon with a cheaper program to do this |
#15
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"ChristopheRonald" wrote in message oups.com... i myself would use native instruments traktor. its a dj program, but the timestretching is very nice on a good computer. you can change the tempo without changing the pitch, and you can change the pitch a certain amount and lock it to that pitch and then slow it down and speed it up. This turns out to be exactly what is *not* necessary in this case. When a recording is made at an odd speed, you need to change both pitch and tempo proportionately. |
#16
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
in article , Arny Krueger at
wrote on 4/19/07 6:12 PM: There is a process from Plangent Technologies that can find all sorts of parasitic noises and lock to them to reduce wow and flutter. That only makes sense if the speed error is varying. A steady-state speed errror such as the OP seems to be describing is not. It is very expensive but is pretty amazing. Barring that there is always a frequency counter and a parametric eq. That gets rid of the hum, but not the speed error. Here's how to correct a speed error in the digital domain: Using a program that allows you to adjust the sample rate in the header of a file without resampling it, change the sample rate by the desired amount. Then resample the file back to its original sample rate. I know that Audition/CEP will do this, and I believe that Audacity will do it as well. With other software, YMMV. For example if you take a file that is sampled at 44,100 Hz and adjust the sample rate in its header to 88,200 Hz, and then resample it back to 44,100 Hz, a 1.00 KHz tone in the file will now have a frequency of 2.00 KHz. This process generally runs as fast and produces as few artifacts as any. Some programs encapsulate it in a function that changes pitch and tempo together. I have something to that effect in Samplitude. I just didn't have a way to find a reference pitch. Now with the idea of looking at the 60hz noise, I MAY have a shot at this without spending a ton of time. I currently have a test disc for them to approve that I did to see if I am on the right track. We'll see what they say about it & go from there. I'll post an update when I know more. -- Tom Jancauskas Imedia |
#17
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
#18
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"Tom Jancauskas" wrote in message ... in article , Arny Krueger at wrote on 4/19/07 6:12 PM: There is a process from Plangent Technologies that can find all sorts of parasitic noises and lock to them to reduce wow and flutter. That only makes sense if the speed error is varying. A steady-state speed errror such as the OP seems to be describing is not. It is very expensive but is pretty amazing. Barring that there is always a frequency counter and a parametric eq. That gets rid of the hum, but not the speed error. Here's how to correct a speed error in the digital domain: Using a program that allows you to adjust the sample rate in the header of a file without resampling it, change the sample rate by the desired amount. Then resample the file back to its original sample rate. I know that Audition/CEP will do this, and I believe that Audacity will do it as well. With other software, YMMV. For example if you take a file that is sampled at 44,100 Hz and adjust the sample rate in its header to 88,200 Hz, and then resample it back to 44,100 Hz, a 1.00 KHz tone in the file will now have a frequency of 2.00 KHz. This process generally runs as fast and produces as few artifacts as any. Some programs encapsulate it in a function that changes pitch and tempo together. I have something to that effect in Samplitude. I just didn't have a way to find a reference pitch. Now with the idea of looking at the 60hz noise, I MAY have a shot at this without spending a ton of time. Yes, getting something that started out as the power line frequency, back to being the power line frequency, would be good. That was a good suggestion from the people who made it. With a narrow-band FFT you can almost always find traces of the power line in a recording to work with. Acoustic recordings and recordings made on battery-powered equipment might be exceptions. |
#19
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
in article , Arny Krueger at wrote on 4/19/07 6:12 PM: There is a process from Plangent Technologies that can find all sorts of parasitic noises and lock to them to reduce wow and flutter. That only makes sense if the speed error is varying. A steady-state speed errror such as the OP seems to be describing is not. There is always some varying speed error, though. But yes, fixing the wow and flutter is much more expensive than fixing a static speed change. It is very expensive but is pretty amazing. Barring that there is always a frequency counter and a parametric eq. That gets rid of the hum, but not the speed error. No, it lets you know what the speed error is. You peak the parametric around 60 Hz, then read the output with the counter. Then you adjust the speed of the table until the number on the counter falls right on 60 Hz. I have something to that effect in Samplitude. I just didn't have a way to find a reference pitch. Now with the idea of looking at the 60hz noise, I MAY have a shot at this without spending a ton of time. You can do it after the transcription, but the problem is that you then wind up having to do sample rate conversion between fairly close sample rates, which is a difficult thing to do well. Also, of course, it's hard to deal with error that varies from the beginning to the end of the disc. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#20
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
"Tom Jancauskas" wrote in message ... in article , Arny Krueger at wrote on 4/19/07 6:12 PM: Here's how to correct a speed error in the digital domain: Using a program that allows you to adjust the sample rate in the header of a file without resampling it, change the sample rate by the desired amount. Then resample the file back to its original sample rate. I know that Audition/CEP will do this, and I believe that Audacity will do it as well. With other software, YMMV. For example if you take a file that is sampled at 44,100 Hz and adjust the sample rate in its header to 88,200 Hz, and then resample it back to 44,100 Hz, a 1.00 KHz tone in the file will now have a frequency of 2.00 KHz. This process generally runs as fast and produces as few artifacts as any. Some programs encapsulate it in a function that changes pitch and tempo together. I have something to that effect in Samplitude. If you can't get it to work there, I just found out that you can do this sort of thing in the freeware program Audacity. I just didn't have a way to find a reference pitch. Audacity's built-in FFT analyzer can be used to determine frequency. However you have to set the FFT size to max (16k) and log frequency plot to get a useful plot. Click the audio track's property pull-down for "set rate" and use the "other" option to set the exact new sample frequency. When you export the file at the same sample rate as you imported it, Audacity automagically resamples the file, and you end up with a frequency-corrected file. |
#21
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
#22
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
On Thu, 19 Apr 2007 17:41:14 GMT, Tom Jancauskas
wrote: Hey all, I have an interesting question. I hope someone can help. Is there a hard ware software product out there that can help with recognizing the proper speed/pitch of a voice? I am working on a record transfer project for a client of mine. All spoken word sermons given by the same person. They have been recorded over a period of about 15 years from 1951 on. The problem is this; They are saying that the recorded speed of the original master that the record was cut from isn't quite the right pitch. Some are slower & some are faster depending on the gear used at the time. I have set up the turntable with a strobe disc to ensure correct playback speed stated on the record for transfer, but they want Her to have the same sound/pitch of her voice across all titles. Is there something I can use to detect the "correct" speed? They have given me a reference to follow, but I'm not able to tell for sure by ear what it should be. There seems to be minute differences that they are detecting. BTW there are over 900 titles to be done. Thanks to all - Tom After reading the whole thread, I applaud you on your patience. My wife is a minister and I know how church committees can be. I understand that this is a big chunk of change, but if it were me I'd be tempted to tell them that there is only one sermon here they need to hear, and it is on one of the Buddha's noble truths Impermanence. The recording devices have change constantly as well as the the medium degrading undoubtedly the minister voice has altered with the years and clearly the committee keeps changing their minds. All things are constantly changing! I think I'd get the person complaining to pick one recording that sound in pitch and get a sense of the proper pitch, then tune them by ear to as close to the mark as possible. If the pitch is altered then it is undoubtedly because the speed is off so adjusting it should put it all right. Then you can clean up the noise. But That's just me I figure if it's clearly audible and sounds like the right person, it should be enough to get the important thing which is the sermon. |
#23
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
in article , at
wrote on 4/20/07 2:15 PM: After reading the whole thread, I applaud you on your patience. My wife is a minister and I know how church committees can be. I understand that this is a big chunk of change, but if it were me I'd be tempted to tell them that there is only one sermon here they need to hear, and it is on one of the Buddha's noble truths Impermanence. The recording devices have change constantly as well as the the medium degrading undoubtedly the minister voice has altered with the years and clearly the committee keeps changing their minds. All things are constantly changing! I was actually pretty impressed with what the recordings sounded like, especially for the time they were recorded. I forgot to mention that these recordings were done all over the country in all different churches. So there is another variable. The gear seems to be consistent over the years though, with about 3 variations in gear as far as I can tell. I think I'd get the person complaining to pick one recording that sound in pitch and get a sense of the proper pitch, then tune them by ear to as close to the mark as possible. If the pitch is altered then it is undoubtedly because the speed is off so adjusting it should put it all right. This is what they are currently asking for. They have already given me a ref of what they want her to sound like EQ level & pitch wise, but the ref sounds atrocious & doesn't have all of the "fire & brimstone" of some of her sermons so the voice changes timber on the others & is interesting to try & match the ref. Then you can clean up the noise. But That's just me I figure if it's clearly audible and sounds like the right person, it should be enough to get the important thing which is the sermon. The records vary in quality from pretty good to " I think someone played this on a close & play & used a toenail as a stylus!" So there are lots o' factors in play. Thanks for your input. -- Tom Jancauskas Imedia |
#24
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
Tom Jancauskas wrote:
The gear seems to be consistent over the years though, with about 3 variations in gear as far as I can tell. This is what they are currently asking for. They have already given me a ref of what they want her to sound like EQ level & pitch wise, but the ref sounds atrocious & doesn't have all of the "fire & brimstone" of some of her sermons so the voice changes timber on the others & is interesting to try & match the ref. Aimee Semple MacPherson? --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#25
Posted to rec.audio.pro
|
|||
|
|||
Pitch/speed recognition hardware/software
|
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Pitch/Speed change question... | Pro Audio | |||
Gradual speed/pitch correction??? | Pro Audio | |||
Speed/pitch adjustment in Nakamichi 582Z cassette deck | Pro Audio | |||
Pitch-recognition software? | Pro Audio | |||
digital studio "pitch" or "tape speed" control like on a 4-track ? | Pro Audio |