Tape Speed Correction Software

[mcp6453]

I discovered some extremely (personally) valuable cassette tapes from 1982. Unfortunately, the recorder on some of them
was running erratically. Is there any software available that will do anything credible towards removing the speed/pitch
variations? Adobe Audition 1.5 has some pitch correction functionality that appears to be dynamic, but the variations
are quite extreme. Thankfully not all tapes are totally bad.

[Don Pearce]

There is pitch correction software out there that can do a really good
job. It uses residual 50/60Hz hum in the recording as its reference
and applies the correction needed to bring that back to frequency.

I can't remember its name, but it is worth a search.

d

[Scott Dorsey]

What is the problem and over what range of speed changes is it? Is it a
recorder with slowly dying batteries where the speed is ramping down slowly,
or a recorder with a bad pinch roller where the tape is alternately sticking
and letting go, or a recorder where the pinch roller has not engaged at al
and so the speed is wild and random and mostly very high?

Is there by any chance a constant tone that could be used as a reference?
Like a 60 Hz hum?

Is the job to make music sound good, or to get intelligible voices?

Somehow I am thinking of Jamie Howarth at Plangent... he has never done
anything like this but maybe he'd like to try. The low running speed of
the cassette makes it much harder than with normal tape.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[hank alrich]

I recently saw mention of a plugin, from a different source, that
pruported to do this, too. Will see if I can recall where I got that
bit.

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[mcp6453]

On 10/19/2014 1:57 PM, Jeff Henig wrote:

> Is this the kind of thing to which you are referring?
>
>
> https://www.izotope.com/en/products/audio-repair/rx/features/

It looks like "Time & Pitch" may be the function that's needed, but it's part of the $1200 package.

It's all voice. The objective is to cause the voices to be as close to normal as possible. The pitch varies quite a bit.
I don't know what the problem with the deck was. It cleared up by the end of one tape. The machine was dragging while it
was recording. There's probably 60 Hz in there somewhere.

[Mike Rivers]

On 10/19/2014 9:00 AM, mcp6453 wrote:
> I discovered some extremely (personally) valuable cassette tapes from 1982. Unfortunately, the recorder on some of them
> was running erratically. Is there any software available that will do anything credible towards removing the speed/pitch
> variations?

No magic bullets for this. Since the recordings are personally valuable
to you, you should just be patient and take the time to work through
them section by section. Once you start looking and listening closely,
you'll probably get a better idea of what's happening (it may not be
completely random) and it'll go faster as you gain experience.


--
"Today's production equipment is IT based and cannot be operated without
a passing knowledge of computing, although it seems that it can be
operated without a passing knowledge of audio" - John Watkinson

Drop by http://mikeriversaudio.wordpress.com now and then

[Peter Larsen]

"mcp6453" <mcp...@gmail.com> skrev i en meddelelse
news:SMCdndlGx5GRfd7J...@giganews.com...

Hum is your saviour, Audition can probably fix it if you divide it in
segments, it can do alter the speed with a linear variation, if in segments
it is likely to be close enough and there is almost always hum enough on an
analog tape recording for the speed variations to be documented.

Use FFT analysis to determine the hum frequency on the tape and adjust speed
until it gets correct. You have to do it to apply any kind of EQ or notch
filtering anyway.

I would prefer to sample at 96 kHz for such drastic work and then probably
downsample to 48 or 44.1 for general fixitology when the speed issues are
fixed.

Kind regards

Peter Larsen

[Peter Larsen]

"Mike Rivers" <mm1...@yahoo.com> skrev i en meddelelse
news:m237o5$qtl$1...@dont-email.me...

> On 10/19/2014 9:00 AM, mcp6453 wrote:
>> I discovered some extremely (personally) valuable cassette tapes from
>> 1982. Unfortunately, the recorder on some of them
>> was running erratically. Is there any software available that will do
>> anything credible towards removing the speed/pitch
>> variations?

> No magic bullets for this.

Not correct as long as it is recorder drift we are talking of, it is trivial
but possibly tedious and AA1.5 will do all that is needed out of the
shrinkwrap.

Kind regards

Peter Larsen

[Scott Dorsey]

mcp6453 <mcp...@gmail.com> wrote:
>
>It looks like "Time & Pitch" may be the function that's needed, but it's part of the $1200 package.
>
>It's all voice. The objective is to cause the voices to be as close to normal as possible. The pitch varies quite a bit.
>I don't know what the problem with the deck was. It cleared up by the end of one tape. The machine was dragging while it
>was recording. There's probably 60 Hz in there somewhere.

If there is 60 Hz, you can lock to it with resolver software and get the
speed pretty constant. You might ask Wes Dooley; he certainly has the
software for the job as well as the skills. He'll charge something but
I doubt it'll be anything like $1200.

[flatfish+++]

The software you are looking for is called "Capstan" by Celemony.

http://www.celemony.com/en/capstan

I used it on some solo piano work that had been recorded on a Tascam
with severe wow and flutter problems and it cleaned up 90 percent or
more of it in "fix it all" mode.

The difference between source and repaired was like night and day.

Worth looking into.

It doesn't run under Linux BTW.
Just thought I would mention that :)



--
flatfish+++

Linux: The Operating System That Put The City Of Munich Out Of
Business.
Before Switching To Linux Read This:
http://linuxfonts.narod.ru/why.linux.is.not.ready.for.the.desktop.current.html

[Mike Rivers]

On 10/20/2014 8:37 AM, Peter Larsen wrote:
> Not correct as long as it is recorder drift we are talking of, it is trivial
> but possibly tedious and AA1.5 will do all that is needed out of the
> shrinkwrap.

By "no magic bullet" I meant that there wasn't a one-click solution. You
need to find a section that's off speed and put it back on speed. If
it's constantly drifting, you'll need to do it in small chunks.

Tedious indeed.

[mcp6453]

On 10/20/2014 7:54 PM, flatfish+++ wrote:
> On Sun, 19 Oct 2014 12:00:11 -0400, mcp6453 wrote:
>
>> I discovered some extremely (personally) valuable cassette tapes from 1982. Unfortunately, the recorder on some of them
>> was running erratically. Is there any software available that will do anything credible towards removing the speed/pitch
>> variations? Adobe Audition 1.5 has some pitch correction functionality that appears to be dynamic, but the variations
>> are quite extreme. Thankfully not all tapes are totally bad.
>
> The software you are looking for is called "Capstan" by Celemony.
>
> http://www.celemony.com/en/capstan
>
> I used it on some solo piano work that had been recorded on a Tascam
> with severe wow and flutter problems and it cleaned up 90 percent or
> more of it in "fix it all" mode.
>
> The difference between source and repaired was like night and day.
>
> Worth looking into.
>
> It doesn't run under Linux BTW.
> Just thought I would mention that :)

It's $4,458. That's a little much. Do you have it?

[Peter Larsen]

"Mike Rivers" <mm1...@yahoo.com> skrev i en meddelelse

news:m24ejt$df8$2...@dont-email.me...

> On 10/20/2014 8:37 AM, Peter Larsen wrote:
>> Not correct as long as it is recorder drift we are talking of, it is
>> trivial
>> but possibly tedious and AA1.5 will do all that is needed out of the
>> shrinkwrap.

> By "no magic bullet" I meant that there wasn't a one-click solution. You
> need to find a section that's off speed and put it back on speed. If it's
> constantly drifting, you'll need to do it in small chunks.

Mike I have have done this with several analog tapes, it is my default
procedure to check for speed variations, usually they are reasonably linear
over a reel and AA allows a gliding stretch. So I tend to end up with 1 to 3
chunks for a reel, often just the entire reel. AA in various incarnations
has a linearly changing stretch and the standard variation of a reel to reel
is close enough to linear to make the audio fit well into narrow notch
filters. Such tend to always be relevant.

> Tedious indeed.

Yes, but not that bad once the workflow for a given tape has been learned.
Here is an short example from a meeting recorded in 1956, speech only, I
think this is the general type of audio asked about by the OP:

http://www.muyiovatki.dk/alarsen/mp3/sf_kbhallen_19590216_0_baand-id.mp3

It is a recording of a meeting in 1956 serially copied for listening to at
local meetings, probably of the pulpit microphone, not the original tape and
I do not know how many generations there was on the way to the digitized
copy. It also illustrates the limitations of DIY recovery ...

Kind regards

Peter Larsen

[hank alrich]

flatfish+++ <phlat...@yahoo.com> wrote:

> The software you are looking for is called "Capstan" by Celemony.

Thank you. That's the one I was tryng to remember. Saw it while looking
at other stuff at their site.

[Sean Conolly]

"mcp6453" <mcp...@gmail.com> wrote in message
news:opGdnSIZBdVcW9jJ...@giganews.com...

There is a 'rental' version - 1 week for $200. Still not cheap but for audio
with personal value (say tapes from a deceased family member) it's a
bargain.

Sean

[Mike Rivers]

On 10/22/2014 9:40 AM, Sean Conolly wrote:

>>> The software you are looking for is called "Capstan" by Celemony.

> There is a 'rental' version - 1 week for $200. Still not cheap but for audio
> with personal value (say tapes from a deceased family member) it's a
> bargain.

Better plan on two weeks, one to use it and one to figure out how to use
it.


--
For a good time, visit http://mikeriversaudio.wordpress.com

[Peter Larsen]

"Mike Rivers" <mm1...@yahoo.com> skrev i en meddelelse

news:m28gum$ch3$1...@dont-email.me...

> On 10/22/2014 9:40 AM, Sean Conolly wrote:

>>>> The software you are looking for is called "Capstan" by Celemony.

>> There is a 'rental' version - 1 week for $200. Still not cheap but for
>> audio
>> with personal value (say tapes from a deceased family member) it's a
>> bargain.

> Better plan on two weeks, one to use it and one to figure out how to use
> it.

Mike, do you recall any actual explanation of what the problem is, other
than "extreme speed variation"? - my thought when I read that was a portable
cassette recorder with low battery voltage loosing speed, I did some sonic
repair on such a tape back in the analog age using the (limited) varispeed
option of a Sony 510 and did at least get it slowed down so much that a
transcript became humanly possible, albeit at a noise-cost.

Capstan by Celemony no doubt is a fabulous product, but the point has not -
as I recall this thread - been made that it is relevant for the problem that
needs solving.

Kind regards

Peter Larsen

[mcp6453]

I actually heard back from iZotope. They don't think their software will do the job. It remains to be seen whether
Capstan might work.

[Mike Rivers]

On 10/22/2014 5:30 PM, Peter Larsen wrote:

> Mike, do you recall any actual explanation of what the problem is, other
> than "extreme speed variation"?

No, he never clarified it or posted an example. Then he (I think it was
the same "he") posted about problems with his Nakamichi cassette deck.
That got me wondering if the problem was with the original recording or
his playback.

I don't think any of us can give him a real suggestion without knowing
what the problem really is. There are plenty of pitch-time correction
tools. How difficult it will be to fix his problem really depends on how
fast the speed is varying. If there's a fairly constant change in speed
over a one minute period, he can fix it in one-minute chunks. If the
speed varies significantly over a period of a couple of seconds, then
he'll have to fix it in chunks a few seconds long.

The Plangent Process uses bias as a reference and can fix small rapid
variations (flutter) but whether it can handle a repeating 50% change in
speed over a period of a couple of seconds, like you'd get if you tapped
out a jig on the supply reel with your finger, or the pinch roller's
tension was weak and the tape kept slipping on the capstan, that's
probably going to be a very tedious process,

[mako...@yahoo.com]

If the op can post a sample of the problem and the length of the tape, ill bet someone on here will make him an offer to fix it t for a fee or maybe even for free.

Mark

[mcp6453]

On 10/22/2014 10:09 PM, mako...@yahoo.com wrote:
> If the op can post a sample of the problem and the length of the tape, ill bet someone on here will make him an offer to fix it t for a fee or maybe even for free.
>
> Mark
>

I'll post a sample this weekend. The files are on a different computer. The recorder was running slowly causing the
recording to appear fast. It is not the same problem that I'm having with the Dragon. The speed ups and slow downs in
the problem recording are in the same spots.

[Adrian Tuddenham]

Could you not just play it back in sections on a variable-speed player
and sort it out with a bit of editing?

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk

[geoff]

On 23/10/2014 4:03 a.m., Mike Rivers wrote:
> On 10/22/2014 9:40 AM, Sean Conolly wrote:
>
>>>> The software you are looking for is called "Capstan" by Celemony.
>
>> There is a 'rental' version - 1 week for $200. Still not cheap but for
>> audio
>> with personal value (say tapes from a deceased family member) it's a
>> bargain.
>
> Better plan on two weeks, one to use it and one to figure out how to use
> it.
>
>

.... hopefully not in that order .

geoff

[Scott Dorsey]

In article <m29bh1$i7c$1...@dont-email.me>, Mike Rivers <mm1...@yahoo.com> wrote:
>
>I don't think any of us can give him a real suggestion without knowing
>what the problem really is. There are plenty of pitch-time correction
>tools. How difficult it will be to fix his problem really depends on how
>fast the speed is varying. If there's a fairly constant change in speed
>over a one minute period, he can fix it in one-minute chunks. If the
>speed varies significantly over a period of a couple of seconds, then
>he'll have to fix it in chunks a few seconds long.

We can give him the suggestion of sending it to Wes Dooley and having him
at least figure out what the original problem might have been.

Cheaper to call someone who has the software than to buy a lot of software
and try it all.

>The Plangent Process uses bias as a reference and can fix small rapid
>variations (flutter) but whether it can handle a repeating 50% change in
>speed over a period of a couple of seconds, like you'd get if you tapped
>out a jig on the supply reel with your finger, or the pinch roller's
>tension was weak and the tape kept slipping on the capstan, that's
>probably going to be a very tedious process,

The Plangent system can use all kinds of references, including 60 Hz leakage.
Although there are cheaper ways to resolve to 60 Hz leakage. I don't think
anyone would be able to recover bias from a cassette but they might have some
other noise that the Plangent software could be keyed on. I know they have
keyed it on repeating motor noise from a cutting lathe.

[Peter Larsen]

"Adrian Tuddenham" <adr...@poppyrecords.invalid.invalid> skrev i en
meddelelse
news:1ltypmd.1wbsvi31lqz7raN%adr...@poppyrecords.invalid.invalid...

> mcp6453 <mcp...@gmail.com> wrote:

>> On 10/22/2014 10:09 PM, mako...@yahoo.com wrote: > If the op can post a
>> sample of the problem and the length of the tape, ill bet someone on here
>> will make him an offer to fix it t for a fee or maybe even for free. > >
>> Mark > I'll post a sample this weekend. The files are on a different
>> computer. The recorder was running slowly causing the recording to appear
>> fast. It is not the same problem that I'm having with the Dragon. The
>> speed ups and slow downs in the problem recording are in the same spots.

> Could you not just play it back in sections on a variable-speed player
> and sort it out with a bit of editing?

As I recall this Adobe Audition 1.5 is available for Mike. Its gliding
stretch is the perfect treatment for analog tape with incorrect speed
because speed variations over a reel tend to vary in a reasonably linear
way.

> ~ Adrian Tuddenham ~

Kind regards

Peter Larsen

[Mike Rivers]

On 10/23/2014 10:42 PM, Peter Larsen wrote:
> As I recall this Adobe Audition 1.5 is available for Mike. Its gliding
> stretch is the perfect treatment for analog tape with incorrect speed
> because speed variations over a reel tend to vary in a reasonably linear
> way.

My interpretation of his description of the speed problem was that it
was erratic, not a change over a long enough period of time to be
considered a steady (for that time) rate of change. But it could just be
my vivid imagination.

[Peter Larsen]

"Mike Rivers" <mm1...@yahoo.com> skrev i en meddelelse
news:m2cejs$qrn$1...@dont-email.me...

> On 10/23/2014 10:42 PM, Peter Larsen wrote:

>> As I recall this Adobe Audition 1.5 is available for Mike. Its gliding
>> stretch is the perfect treatment for analog tape with incorrect speed
>> because speed variations over a reel tend to vary in a reasonably linear
>> way.

> My interpretation of his description of the speed problem was that it was
> erratic, not a change over a long enough period of time to be considered a
> steady (for that time) rate of change. But it could just be my vivid
> imagination.

We are still waiting for a good first question from the OP ... O;-)

> For a good time, visit http://mikeriversaudio.wordpress.com

Kind regards

Peter Larsen

[mcp6453]

The tape speed is erratic. I will post a segment tomorrow. The computer with the file on it is not here at the moment.

[Peter Larsen]

"mcp6453" <mcp...@gmail.com> skrev i en meddelelse
news:I-edneA1Ra81RdTJ...@giganews.com...

> The tape speed is erratic.

Is the recording the problem or the playback?

By logic such a variation could be caused by the delivery reel brake or a
braking of it by some other cause. That would result in 4 types of segments:
ok, braking, braked and speeding up. Considering the fairly slow rotation of
a casette it might not be hopeless, albeit quite tedius to address those
manually or batched.

> I will post a segment tomorrow. The computer with the file on it
> is not here at the moment.

Have you asked the guy(s) Scott suggested what a probable price range is in
case their (assumed)licensed Capstan software will do it? - even if still
unheard it sounds very much like something I'd love to outsource ...

Kind regards

Peter Larsen

[Don Pearce]

Capstan is pretty nicely demoed on Youtube. It certainly seems to do
much more than chopping the piece into chunks and re-clocking it. Have
a look

https://www.youtube.com/watch?v=QGzsnpfhIW8

d

[Peter Larsen]

"Don Pearce" <sp...@spam.com> skrev i en meddelelse
news:544ccf44...@news.eternal-september.org...

> Capstan is pretty nicely demoed on Youtube. It certainly seems to do
> much more than chopping the piece into chunks and re-clocking it. Have
> a look

> https://www.youtube.com/watch?v=QGzsnpfhIW8

Very interesting, thank you!

> d

Kind regards

Peter Larsen

[Don Pearce]

The problem is, I can't imagine wanting this for regular use. I'd
almost want it on rental for a day.

d

[Les Cargill]

I'm thinking sending it out to somebody that does it more than I do
would be even better.

--
Les Cargill

[mcp6453]

On 10/24/2014 12:42 AM, mcp6453 wrote:
>
> The tape speed is erratic. I will post a segment tomorrow. The computer with the file on it is not here at the moment.

Better late than never. Here's a one minute clip from the cassette tape mentioned previously. The problem is in the
recording, not in the playback. Not only is the speed off, it is inconsistent.

https://soundcloud.com/mcp6453/20th-anniversary

It looks like Capstan could be useful, but I don't see it happening.

[hank alrich]

Peter Larsen <dig...@hotmail.com> wrote:

> "mcp6453" <mcp...@gmail.com> skrev i en meddelelse
> news:I-edneA1Ra81RdTJ...@giganews.com...
>
> > The tape speed is erratic.
>
> Is the recording the problem or the playback?

Does he mean that the tapes speed varies over the length of a reel, or
that the tape speed varies due to tape path motion instability over the
short run, i.e., like wow and flutter, but perhaps wow-ish over an
unusually long span?

In the first case, would pitch serve as a reference to restablish
timebase?

In the second case, that fancy software that reads bias and other signal
data would seem necessary, in part because there might be no point
anywhere in the reel that the motion is stable. If the tape is
constantly accelerating adn decelerating, that would seem a genuine
nightmare to attempt to repair manually.

> By logic such a variation could be caused by the delivery reel brake or a
> braking of it by some other cause. That would result in 4 types of segments:
> ok, braking, braked and speeding up. Considering the fairly slow rotation of
> a casette it might not be hopeless, albeit quite tedius to address those
> manually or batched.
>
> > I will post a segment tomorrow. The computer with the file on it
> > is not here at the moment.
>
> Have you asked the guy(s) Scott suggested what a probable price range is in
> case their (assumed)licensed Capstan software will do it? - even if still
> unheard it sounds very much like something I'd love to outsource ...
>
> Kind regards
>
> Peter Larsen

[Mike Rivers]

On 10/26/2014 9:30 PM, mcp6453 wrote:

> Better late than never. Here's a one minute clip from the cassette tape mentioned previously.

This is something that you can fix manually with a lot of patience. It
sounds like a news broadcast and it's possible that it was broadcast
that way. Of course that's no excuse for not trying to make it better
than original if you care enough, or are willing to spend enough money
to hire someone with experience and better tools than you have.

Have you tried contacting NBC News to see if they have a copy in their
archive of the original broadcast?


--

[Scott Dorsey]

It's both speeding up and slowing down. All kinds of things can slow the
tape down, but as long as the pinch roller is locked in place and the motor
running right, it won't speed up.

This sounds very much to me like a bad pinch roller or the pinch roller
having not locked down properly. If it's on some tapes and not others,
that would make sense.

I'd talk to Wes Dooley about it. He might be too expensive to do the work
but he'll probably be happy to talk about it and suggest possibilities.
I do think this is the sort of thing the capstan software might do well.

[Peter Larsen]

"Scott Dorsey" <klu...@panix.com> skrev i en meddelelse
news:m2lj92$ob6$1...@panix2.panix.com...

> It's both speeding up and slowing down. All kinds of things can slow the
> tape down, but as long as the pinch roller is locked in place and the
> motor
> running right, it won't speed up.

The _recording_ slows down more or less or something happens in transit,
sounds like a compact cassette or perhaps a dictaphone recording on
mini-casette or even a recording answering machine recording of an incoming
call that also was mangled in transit.

> This sounds very much to me like a bad pinch roller or the pinch roller
> having not locked down properly. If it's on some tapes and not others,
> that would make sense.

> I'd talk to Wes Dooley about it. He might be too expensive to do the work
> but he'll probably be happy to talk about it and suggest possibilities.
> I do think this is the sort of thing the capstan software might do well.

NSA or the Brits or the Germans or "all of the above" probably have a better
recording of it if the journalist called home from Russia ... O;-)

Notice how clean the ending "more news in a minute" sounds, the audio was
ailing when transmitted, something that suggests that it is as good as the
radio station has it.

> --scott

Kind regards

Peter Larsen

[thekma...@gmail.com]

Mike Rivers:

Have you ever tried debating with a Vulcan?

;)

[Mike Rivers]

On 10/27/2014 6:53 PM, thekma...@gmail.com wrote:
> Mike Rivers:
> Have you ever tried debating with a Vulcan?

Sure. I always win.

[thekma...@gmail.com]

Mike Rivers wrote: "On 10/27/2014 6:53 PM, thekma...@gmail.com wrote:
> Mike Rivers:
> Have you ever tried debating with a Vulcan?

Sure. I always win.

- show quoted text -"

Kind regards!

LOLOL..