could autotune correct wow?

[J. P. Gilliver]

I watch a lot of old music clips on YouTube; one of the comments I see a
lot (I'm sure _some_ of it is just the "it was better in the old days"
mentality) is something like "they could really _sing_ in those days -
no this, no that, no autotune, just singers who could sing on pitch
without …."

I've never been quite sure what autotune is, but from the nature of such
comments, and the name itself, I presume it's some device (software, I
presume) that corrects the pitch of singers who _can't_ stay on pitch,
using more or less the same principle as YouTube or VLC when playing
fast or slow, presumably hooked to something that _is_ the correct pitch
such as a feed from one of the accompanying instruments.

However, a lot of the old clips of singers who _could_ hold a note,
aren't - if they're from tapes, sometimes records, or particularly film
- because of that phenomenon possibly not familiar to the latest
generation of sound engineers: wow. (Not sure why film seems
particularly prone to it: possibly heavy reels driven entirely by film
tension [but surely that applies to tape too]?)

I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

[Brian Gaff]

Autotune comes in several flavours. It can be used to re pitch v voices. IE
as a novelty, perhaps the best known use wads on the Sher song Believe. In
that case it had a hard switch, almost like a yodel.
However its also used in harmonising and double tracking. It nearly always
sounds artificial, since its just too precise. I notice Michael Bubble uses
it for double tracking sometimes, and he is a good singer, but I just don't
like the sound it gives. Maybe its just me.
As for using it for wow. It has been used to add a kind of modulation to a
singer who has very little vibrato or whatever, but once again its too
precise.

The problem with correcting wow though is that its often too slow and
variable. You are in effect turning the effect on its head to even out the
irregularities. It has been done in reclaiming old recordings, but once
again, it can lead to some artificialness unless you can create the actual
waveforms exactly.
It would be easier in many instances to do it via sample rate.
I first noticed it being used on a live performance on Carol Kings Living
Room Tour, to make her sound a bit smoother, but that missed the whole point
of her voice and its emotion, spoiling it somewhat to my ears. When was
that, 2007 ish?

To me the biggest issue with modern recordings is close microphone
placement and excessive compression on it. It simply does not sound real to
me.
Some of the recordings I've heard done recently in Nashville, are very very
compressed against those done say 10 years ago, which sounded more natural.
Again Michael Bubbles recordings, wherever they are done are what he calls
intimate, and I call downright un-natural.
Nashville recordings have completely ruined the Shires and many others for
me.
Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by J. P. Gilliver
I watch a lot of old music clips on YouTube; one of the comments I see a
lot (I'm sure _some_ of it is just the "it was better in the old days"
mentality) is something like "they could really _sing_ in those days - no
this, no that, no autotune, just singers who could sing on pitch
without .."
I've never been quite sure what autotune is, but from the nature of such
comments, and the name itself, I presume it's some device (software, I
presume) that corrects the pitch of singers who _can't_ stay on pitch,
using more or less the same principle as YouTube or VLC when playing fast
or slow, presumably hooked to something that _is_ the correct pitch such
as a feed from one of the accompanying instruments.
However, a lot of the old clips of singers who _could_ hold a note,
aren't - if they're from tapes, sometimes records, or particularly film -
because of that phenomenon possibly not familiar to the latest generation
of sound engineers: wow. (Not sure why film seems particularly prone to
it: possibly heavy reels driven entirely by film tension [but surely that
applies to tape too]?)
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?
--
If you carry on hating, you're the one who's damaged.
- Sir Harold Atcherley, sent to the Burma/Siam railway in April 1943

[J. P. Gilliver]

In message <u3l0g9$1gp4p$***@dont-email.me> at Fri, 12 May 2023 10:22:13,
Brian Gaff <***@gmail.com> writes
[]

Post by Brian Gaff
The problem with correcting wow though is that its often too slow and
variable. You are in effect turning the effect on its head to even out the
irregularities. It has been done in reclaiming old recordings, but once
again, it can lead to some artificialness unless you can create the actual
waveforms exactly.

I've often thought about that, and that the correction "waveform" would
actually be quite a complex mathematical entity, when it's from a
mis-centered disc record, because for a given fixed offset, the amount
of wow increases as you approach the centre. Made more complicated for
non-fixed-pitch records (most 45s and 33s).

Post by Brian Gaff
It would be easier in many instances to do it via sample rate.

Agreed.
[]

Post by Brian Gaff
To me the biggest issue with modern recordings is close microphone
placement and excessive compression on it. It simply does not sound real to
me.

Certainly, the degree of compression has increased over the last few
decades - https://en.wikipedia.org/wiki/Loudness_war suggests it really
took off after CDs became mainstream and may have reduced slightly since
about 2005, though says it was first discussed in the 1940s. Compression
is a good topic for discussion on its own. Most people _say_ they don't
like it, but it may still influence their choice of, say, radio station.
(My recent experience with YouTube material has included some that is at
far too _low_ an absolute level - in one or two cases I've had to apply
my ×4 preset _twice_ to exceed half full scale [i. e. they're less than
a sixteenth full scale]. Seems to be mainly old video material: I can
only assume that some contributors are using video digitising equipment
that doesn't allow audio level adjustment after digitisation [or they
don't know how to do it]. That's yet another different matter to dynamic
range, of course.)

Post by Brian Gaff
Some of the recordings I've heard done recently in Nashville, are very very
compressed against those done say 10 years ago, which sounded more natural.

You'd agree with the writer of the Wikipedia article I've cited above,
who is fairly clearly of that opinion.

Post by Brian Gaff
Again Michael Bubbles recordings, wherever they are done are what he calls
intimate, and I call downright un-natural.

(The guy's stage name has only two Bs, by the way - the one in the
middle isn't a double. [I think I've seen the e with an acute accent on
it, too, though I know that can't reliably pass through news.])

Post by Brian Gaff
Nashville recordings have completely ruined the Shires and many others for
me.
Brian

Not Borsetshire, I hope ...

[Brian Gaff]

The British country duo called the Shires after where they come from.

Yes Mr Bubble' or whatever his name is, is in fact a very good singer. He
really does not need all that messing about with the voice, but he is not
the only one. In my view high compression levels is only for noisy
environments like cars etc. The BBC seem to do it a lot. I have recording
here made digitally from in Concerts from the 70s, and recently repeated,
and the repeats are pumping and decidedly tizzy sounding, no doubt due to
too much processing.

Even recent remastering of analogue to digital points to the fact that
many of the people doing it seem never to have heard the original Vinyls.
The Beatles ones are OK as are many of the Elton John ones but may others
like Blondie, Leo Sayer etc seem to have a decidedly obvious dynamic
restriction. A converted copy of the Endless Flight LP and the CD version
shows this up. My conversion sounds more dynamic. You almost get the
impression that some engineers these days set it up and then piss off down
the pub.

Certainly many of the in concerts since mid 2000s sound like that. The last
dynamic transmission on radio 2 was back in 2001 The Bee Gees doing their
latest album at the BBC.

Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by J. P. Gilliver
[]

Post by Brian Gaff
The problem with correcting wow though is that its often too slow and
variable. You are in effect turning the effect on its head to even out the
irregularities. It has been done in reclaiming old recordings, but once
again, it can lead to some artificialness unless you can create the actual
waveforms exactly.

I've often thought about that, and that the correction "waveform" would
actually be quite a complex mathematical entity, when it's from a
mis-centered disc record, because for a given fixed offset, the amount of
wow increases as you approach the centre. Made more complicated for
non-fixed-pitch records (most 45s and 33s).

Post by Brian Gaff
It would be easier in many instances to do it via sample rate.

Agreed.
[]

Post by Brian Gaff
To me the biggest issue with modern recordings is close microphone
placement and excessive compression on it. It simply does not sound real to
me.

Certainly, the degree of compression has increased over the last few
decades - https://en.wikipedia.org/wiki/Loudness_war suggests it really
took off after CDs became mainstream and may have reduced slightly since
about 2005, though says it was first discussed in the 1940s. Compression
is a good topic for discussion on its own. Most people _say_ they don't
like it, but it may still influence their choice of, say, radio station.
(My recent experience with YouTube material has included some that is at
far too _low_ an absolute level - in one or two cases I've had to apply my
×4 preset _twice_ to exceed half full scale [i. e. they're less than a
sixteenth full scale]. Seems to be mainly old video material: I can only
assume that some contributors are using video digitising equipment that
doesn't allow audio level adjustment after digitisation [or they don't
know how to do it]. That's yet another different matter to dynamic range,
of course.)

Post by Brian Gaff
Some of the recordings I've heard done recently in Nashville, are very very
compressed against those done say 10 years ago, which sounded more natural.

You'd agree with the writer of the Wikipedia article I've cited above, who
is fairly clearly of that opinion.

Post by Brian Gaff
Again Michael Bubbles recordings, wherever they are done are what he calls
intimate, and I call downright un-natural.

(The guy's stage name has only two Bs, by the way - the one in the middle
isn't a double. [I think I've seen the e with an acute accent on it, too,
though I know that can't reliably pass through news.])

Post by Brian Gaff
Nashville recordings have completely ruined the Shires and many others for
me.
Brian

Not Borsetshire, I hope ...
--
Less rules means fewer grammar? - Marjorie in UMRA, 2014-1-28 13:14

[Liz Tuddenham]

J. P. Gilliver <***@255soft.uk> wrote:

[...]

Post by J. P. Gilliver
However, a lot of the old clips of singers who _could_ hold a note,
aren't - if they're from tapes, sometimes records, or particularly film
- because of that phenomenon possibly not familiar to the latest
generation of sound engineers: wow. (Not sure why film seems
particularly prone to it: possibly heavy reels driven entirely by film
tension [but surely that applies to tape too]?)

Sometimes the wow comes from the original sound camera, but that is very
rare. One particularly bad example is the record of "Warsaw Concertto"
from the sound track of "Dangerous Moonlight", the pitch wavers very
noticeably.

More often it is because the transcription engineer isn't doing the job
properly.

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

[John Williamson]

Post by Liz Tuddenham

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

[J. P. Gilliver]

In message <***@mid.individual.net> at Fri, 12 May 2023
11:01:57, John Williamson <***@btinternet.com> writes
[]

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Another one might be timebase whistle; of YouTube material that actually
goes that high (which is much of it, even if the source material didn't
go anywhere near that high), I've seen it (as an about -80dB [purple in
GoldWave] very visible line on the spectrogram - I can't hear that high)
on I'd say 60-80% of recordings. Though it'd only of course be of use if
it was on the original recording, whereas I suspect in most cases I see,
it's been introduced at some subsequent stage, where it obviously
wouldn't be any use, in fact using it would just make things worse.

It's interesting how we have developed notches in our hearing to our
local mains frequency and its harmonics; I remember borrowing some Bob
Newhart records from my local library in, I think, the 1980s, and being
very aware of the mains buzz on those - which I assume was because they
were 60 not 50 hertz (or harmonics thereof).

[Brian Gaff]

When I played some of my early reel to reel tapes recorded from my first
stereo fm tuner if you slowed the tape down, you could hear the whistle from
the pilot tone. That was at 19khz I think.

I notice on a lot of content that comes into the system from streams there
is a very sudden roll off as the frequency gets to about 15khz. I'm only
guessing but I'd imagine this stops any weird beats with in band signals.
There was a skill on the echo which I recorded that was supposed to test
your hearing. It was basically a sliding tone. However I have a signal
generator and could hear it or feel the pressure of it at up to 18khz, but
the one on line disappeared at just under 15. I tested it by recoding it
and playing it back on goldwave at a slower speed, it was not there, When I
told Amazon they took it away, as it was actually an advert for hearing
aids.
Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by J. P. Gilliver
[]

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level could
you pick up on? Even stray air conditioning noise might be constant
enough.

Another one might be timebase whistle; of YouTube material that actually
goes that high (which is much of it, even if the source material didn't go
anywhere near that high), I've seen it (as an about -80dB [purple in
GoldWave] very visible line on the spectrogram - I can't hear that high)
on I'd say 60-80% of recordings. Though it'd only of course be of use if
it was on the original recording, whereas I suspect in most cases I see,
it's been introduced at some subsequent stage, where it obviously wouldn't
be any use, in fact using it would just make things worse.
It's interesting how we have developed notches in our hearing to our local
mains frequency and its harmonics; I remember borrowing some Bob Newhart
records from my local library in, I think, the 1980s, and being very aware
of the mains buzz on those - which I assume was because they were 60 not
50 hertz (or harmonics thereof).
--
"quidquid latine dictum sit, altum viditur". ("Anything is more impressive if
you say it in Latin")

[Brian Gaff]

Talking of hum. Some recordings obviously have a bass roll off or notch for
rumble and hum.
On Stevie Wonders Songs in the key of life, there is one particular track
backed merely with a harp and you clearly hear the 60hz on it, With careful
notch filtering these days you can remove it almost completely.
Many recordings of Classical Music on Decca on a decent sound system give
the rumble of tube trains under the recording venue. Was that the Kingsway
Halls?

Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by John Williamson

Post by Liz Tuddenham

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level could
you pick up on? Even stray air conditioning noise might be constant
enough.
--
Tciao for Now!
John.

[Liz Tuddenham]

Brian Gaff <***@gmail.com> wrote:

[...]

Post by Brian Gaff
Many recordings of Classical Music on Decca on a decent sound system give
the rumble of tube trains under the recording venue. Was that the Kingsway
Halls?

I have heard that story too, but I haven't noticed the effect on their
Kingsway Hall records..

It seems Decca were a bit unfortunate in their choice of studios. Some
recordings from their Upper Thames Street studios have copious rumble
from the trains overhead, as the studio was located underneath Canon
Street Station.

I have a set of recordings of the Queen's Hall Orchestra under Henry
Wood, which were made in the Upper Thames Street studio; some sides of
the issued version are clearly dubbings, so it is likely they were
filtered in 'post-production' to remove the worst of the rumble. On
the non-dubbed sides, the trains are clearly audible.

[J. P. Gilliver]

Post by Brian Gaff
Talking of hum. Some recordings obviously have a bass roll off or notch for
rumble and hum.
On Stevie Wonders Songs in the key of life, there is one particular track
backed merely with a harp and you clearly hear the 60hz on it, With careful
notch filtering these days you can remove it almost completely.

It's not necessarily removable with a notch - you might need a comb,
depending on how many harmonics get through.
[]

[Liz Tuddenham]

Post by John Williamson

Post by Liz Tuddenham

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Most of the early disc-recording amplifiers were powered by batteries
and the microphone circuits were low impedance and balanced, so there
wouldn't have been any electrical hum. Mains-operated equipment, other
than lamps, was a rarity, so there also wouldn't have been much acoustic
hum.

Both HMV and Columbia always used weight-motors to drive their lathes
(well into the 1950s), but Edison Bell and Decca used electric motors
once they started recording electrically. In every case they used
extremely heavy, well-balanced, turntables.

The only record with wow which I have been able to definitely attribute
to the motor is "Motoring" (Columbia MC9) from 26 May 1919. The
recording was made in the theatre with portable equipment and someone
had wound the weight motor carelessly. Instead of lying flat on the
drum, the main weight hawser had been wound with crossed turns which
increased the effective diameter of the drun once per rev. The timing
of the recorded wow corresponded with the rotation of the weight drum
(calculated from the length of the hawser and the playing time).

Most faults show up at the time of recording, but this wouldn't have
been noticed until a pressed copy was available. The song was topical
and short-lived, so they just went ahead an issued it, wow and all.

[John Williamson]

Post by Liz Tuddenham

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Most of the early disc-recording amplifiers were powered by batteries
and the microphone circuits were low impedance and balanced, so there
wouldn't have been any electrical hum. Mains-operated equipment, other
than lamps, was a rarity, so there also wouldn't have been much acoustic
hum.

Fair enough, I was thinking more of tape recordings than disc.

[Liz Tuddenham]

Post by John Williamson

Post by Liz Tuddenham

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Most of the early disc-recording amplifiers were powered by batteries
and the microphone circuits were low impedance and balanced, so there
wouldn't have been any electrical hum. Mains-operated equipment, other
than lamps, was a rarity, so there also wouldn't have been much acoustic
hum.

Fair enough, I was thinking more of tape recordings than disc.

There is a residual bias oscillator tone detectable on some tapes (you
can often hear it if you slow the tape right down). This would be at an
indeterminate frequency, but its constancy would be a good guide for
speed stability, once the required basic speed had been established.

The detection of 50c/s hum would require a high 'Q' filter, which would
be slow to respond; a filter with the same 'Q' at the much higher bias
frequency would have a proportionally faster response time. There is
also the matter of the playback machine picking up its own 50c/s hum or,
even worse, rotor-frequency hum from the capstan motor (50c/s - slip
speed); that would cause a lot of problems..

[Theo]

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

I'm pretty sure that's been done. You can also do it forensically to
timestamp recordings:
https://robertheaton.com/enf/

Theo

[J. P. Gilliver]

Post by Theo

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

I'm pretty sure that's been done. You can also do it forensically to
https://robertheaton.com/enf/
Theo

Interesting. Bit like dendrochronology (tree-ring dating). Does suggest
someone's recording the fluctuations in the grid though - scary! (Not
the recording as such, but the amount of detail needed ...)

Bit unimpressed by the webpage though - "Power is transmitted through an
electrical grid as alternating current (AC). This means that the current
flowing through the grid’s cables constantly changes direction. In
most regions - including Great Britain, where I live - the current’s
direction switches 50 times a second, or 50Hz. In some other regions,
like the US and Canada, it switches 60 times a second, or 60Hz." No, it
switches 100 times a second. Doesn't matter for the sake of the
explanation, but this sort of sloppy explaining doesn't bode well.

Reading on, he's talking about monitoring fluctuations in the range
49.8-50.2 Hz - and suggests a sample rate of 300 Hz. Yes, I know my
Nyquist, but I'd still be wary of using such a low sample rate to
monitor such small variations.

Basically, I think the chap knows what he's on about, and has tried to
simplify it to describe it - but IMO overso.

[John Williamson]

Post by J. P. Gilliver

Post by Theo

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

I'm pretty sure that's been done. You can also do it forensically to
https://robertheaton.com/enf/
Theo

Interesting. Bit like dendrochronology (tree-ring dating). Does suggest
someone's recording the fluctuations in the grid though - scary! (Not
the recording as such, but the amount of detail needed ...)

There is a second by second record available, though I'm not sure how
long it's kept for.

https://wwwhome.ewi.utwente.nl/~ptdeboer/misc/mains.html

This graph represents data for the EU grid for a period in 2005 showing
the instantaneous frequency and the difference in time shown by a clock
locked to the mains against a stable time such as the national time
references or GPS. There is a link on the page to a real problem on the
grid in February and March 2018, when the mains clock lagged by up to
six minutes. This was apparently caused by a political dispute.

https://preview.entsoe.eu/news/2018/03/06/press-release-continuing-frequency-deviation-in-the-continental-european-power-system-originating-in-serbia-kosovo-political-solution-urgently-needed-in-addition-to-technical/

I can check the mains frequency from here on line in real time, and, if
I could be bothered, record the data.

It's only really been of much use since we switched over to digital
recording, with the sample rate crystal controlled. On a tape recorder
with a synchronous motor, the recorded wavelength would vary with the
speed of the motor with a lag due to the flywheel inertia, so if you
play back at a constant speed, the frequency variations wouldn't appear.
Some of the better battery powered portables might be stable enough for
a meaningful result.

[NY]

Post by John Williamson

Post by Liz Tuddenham

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be,  is  just part of the basic set-up when you are doing
transcription work.  I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Funny you should say that. I recorded a piece of music from the TV many
years ago by the time-honoured technique of placing a microphone next to
the TV's speaker. I recorded it on my battery-powered cassette recorder
and the batteries may have been dying because there is horrendous wow:
the pitch of the playback rises every time there is a loud part of the
music - especially noticeable since some of the instruments have very
sharp attack. Maybe the extra current drawn by the amplifier on loud
sounds caused the motor voltage to sag...

When I copied the tape to my PC (via its soundcard) much more recently,
I noticed something that I couldn't hear: the 15625 Hz line frequency.
This set me thinking: perhaps some form of servo system could be devised
(either in hardware, controlling the tape recorder's motor, or in
software, post-processing the digital copy of the tape, so as to keep
the tone at exactly 15625 Hz?

[J. P. Gilliver]

[]

Post by NY

Post by John Williamson
Just wondering. If there is enough mains hum, could you read that to
stabilise everything by varying sample rates? If so, how low a level
could you pick up on? Even stray air conditioning noise might be
constant enough.

Funny you should say that. I recorded a piece of music from the TV many
years ago by the time-honoured technique of placing a microphone next
to the TV's speaker. I recorded it on my battery-powered cassette
recorder and the batteries may have been dying because there is
horrendous wow: the pitch of the playback rises every time there is a
loud part of the music - especially noticeable since some of the
instruments have very sharp attack. Maybe the extra current drawn by
the amplifier on loud sounds caused the motor voltage to sag...

But the sound would not have been coming from the cassette recorder's
speaker, would it - or were you monitoring, and using a very long mike
lead to avoid feedback? Most portable cassette machines mute the speaker
on record, at least when recording from a microphone. So unless the TV
was also running on battery ...

Post by NY
When I copied the tape to my PC (via its soundcard) much more recently,
I noticed something that I couldn't hear: the 15625 Hz line frequency.
This set me thinking: perhaps some form of servo system could be
devised (either in hardware, controlling the tape recorder's motor, or
in software, post-processing the digital copy of the tape, so as to
keep the tone at exactly 15625 Hz?

Although Liz will say correcting at source is better, I suspect getting
a servo loop to work that fast and reliably on the tape motor would be
quite a difficult task if not impossible, so software post-processing
probably a better approach (the servo wouldn't be correcting _at source_
anyway - the "damage" is done).

Are you sure the 15625 was from the original recording session? (Can you
see it moving up and down with the wow?) I've noticed a strong line
between 15 and 16 kHz on a vast number of tracks from YouTube, including
many where no video should have been anywhere near the source (e. g.
digitised old 78s!); I can only assume that they picked up something
from, say, a computer monitor. (Though those aren't at that low a
frequency these days, are they? Maybe there is - or was until recently -
a background 15625 [or US value, which is quite close] miasma in all
built-up areas!)

Was this just an intellectual thought, or is it a recording you'd really
like to retrieve: in other words, what was/is the music? We can probably
find you a recording; virtually everything is on YouTube these days (-:!

[John Williamson]

Post by J. P. Gilliver
But the sound would not have been coming from the cassette recorder's
speaker, would it - or were you monitoring, and using a very long mike
lead to avoid feedback? Most portable cassette machines mute the speaker
on record, at least when recording from a microphone. So unless the TV
was also running on battery ...

It is marginal, but the record amplifier uses more current when there is
more signal being fed to the head. This may affect the speed slightly
depending on just how flat the batteries are. This would be more evident
when the recorder uses a permanent magnet erase head, DC bias and a
mechanical governor, as many cheap recorders used to.

[NY]

Post by John Williamson

Post by J. P. Gilliver
But the sound would not have been coming from the cassette recorder's
speaker, would it - or were you monitoring, and using a very long mike
lead to avoid feedback? Most portable cassette machines mute the speaker
on record, at least when recording from a microphone. So unless the TV
was also running on battery ...

It is marginal, but the record amplifier uses more current when there is
more signal being fed to the head. This may affect the speed slightly
depending on just how flat the batteries are. This would be more evident
when the recorder uses a permanent magnet erase head, DC bias and a
mechanical governor, as many cheap recorders used to.

And this *was* a cheap recorder - might have been my first one as a child
which was

Phillips EL3302 https://www.petervis.com/Tape Recorders/Philips Compact
Cassette Recorder EL-3302/Philips Compact Cassette Recorder EL-3302.html

<https://www.ebay.co.uk/itm/374376088111?mkevt=1&mkcid=1&mkrid=710-53481-19255-0&campid=5338364431&toolid=11000&_ul=UK&customid=CjwKCAjwvJyjBhApEiwAWz2nLaux44B-clDMgdJovpD9h4ZYsNWH7p1DDKrh3vODhPg49qPbuOkCiRoCJyMQAvD_BwE&gclid=CjwKCAjwvJyjBhApEiwAWz2nLaux44B-clDMgdJovpD9h4ZYsNWH7p1DDKrh3vODhPg49qPbuOkCiRoCJyMQAvD_BwE>

Just seeing that leather case brings back the very recognisable smell of it.

[John Williamson]

Post by NY

Post by John Williamson
It is marginal, but the record amplifier uses more current when there
is more signal being fed to the head. This may affect the speed
slightly depending on just how flat the batteries are. This would be
more evident when the recorder uses a permanent magnet erase head, DC
bias and a mechanical governor, as many cheap recorders used to.

And this *was* a cheap recorder - might have been my first one as a
child which was
Phillips EL3302

I have one of the model before that here, the EL3301. Two part
microphone (The microphone and remote control switch coould be
separated),and no interlock to prevent accidentally over recording
cassette albums. It all still works, with the original motor, which had
a mechanical governor.

At the time, they were actually not cheap. They were originally sold as
a premium tool to allow executives to dictate stuff while out and about.

The EL3301 was one of, if not *the* first portable cassette recorder.

[Liz Tuddenham]

Post by John Williamson

Post by NY

Post by John Williamson
It is marginal, but the record amplifier uses more current when there
is more signal being fed to the head. This may affect the speed
slightly depending on just how flat the batteries are. This would be
more evident when the recorder uses a permanent magnet erase head, DC
bias and a mechanical governor, as many cheap recorders used to.

And this *was* a cheap recorder - might have been my first one as a
child which was
Phillips EL3302

I have one of the model before that here, the EL3301. Two part
microphone (The microphone and remote control switch coould be
separated),and no interlock to prevent accidentally over recording
cassette albums. It all still works, with the original motor, which had
a mechanical governor.
At the time, they were actually not cheap. They were originally sold as
a premium tool to allow executives to dictate stuff while out and about.
The EL3301 was one of, if not *the* first portable cassette recorder.

The EL3301 was the first - and it used germainum transistors; the
EL3302 was mechanically almost identical, apart from the record-inhibit
latch, but used silicon transistors. The Class-B output stage did not
operate during recording and all the other electronics were Class-A, so
there shouldn't have been any change in battery current during loud
recording peaks (apart from the modulation meter, which only drew about
1 mA).

The design of those machines is described in great detail in the
"Philips Technical Review".

[NY]

Post by John Williamson

Post by NY
And this *was* a cheap recorder - might have been my first one as a
child which was
Phillips EL3302

I have one of the model before that here, the EL3301. Two part
microphone (The microphone and remote control switch could be
separated),and no interlock to prevent accidentally over recording
cassette albums. It all still works, with the original motor, which had
a mechanical governor.
At the time, they were actually not cheap. They were originally sold as
a premium tool to allow executives to dictate stuff while out and about.
The EL3301 was one of, if not *the* first portable cassette recorder.

I remember the two-part, slightly conical microphone, with a switch
capable of being separated from the mike itself.

I can remember using the EL3302 at a house which we left in 1969, so we
had it no later than that. Given that this model was launched in 1967,
we must have been early adopters.

Looking at photos, mine was the 3302 which has a rectangular rather then
circular record button.

Dad used it for dictating when we was working from home - I can remember
very clearly his "posh voice" when he was dictating for the secretary to
type ;-) When he later got a handheld dictation recorder that used
B-shaped cassettes (*), I was able to play with the Philips recorder.

Interesting that the 3301 didn't apparently have the record-inhibit
switch that used the absence of the tab to prevent overwriting.



(*) So not the tiny micro-cassettes that were found in some telephone
answering machines. May have been this format

although the recorder was a lot
smaller. I remember that the playback wasn't muted during rewind or
fast-forward so I got used to hearing the squawk every time Dad rewound
to correct the wording of something.

[NY]

Post by J. P. Gilliver

Post by NY
Funny you should say that. I recorded a piece of music from the TV many
years ago by the time-honoured technique of placing a microphone next to
the TV's speaker. I recorded it on my battery-powered cassette recorder
and the batteries may have been dying because there is horrendous wow: the
pitch of the playback rises every time there is a loud part of the music -
especially noticeable since some of the instruments have very sharp
attack. Maybe the extra current drawn by the amplifier on loud sounds
caused the motor voltage to sag...

But the sound would not have been coming from the cassette recorder's
speaker, would it - or were you monitoring, and using a very long mike
lead to avoid feedback? Most portable cassette machines mute the speaker
on record, at least when recording from a microphone. So unless the TV was
also running on battery ...

I wasn't implying that the sound was coming from the speaker *during the
recording*. I was meaning that when the tape was played back afterwards, the
pitch rose during loud sounds which equates to the motor slowing down at
those points during recording - assuming that on playback the motor is able
to run at constant speed.

Post by J. P. Gilliver

Post by NY
When I copied the tape to my PC (via its soundcard) much more recently, I
noticed something that I couldn't hear: the 15625 Hz line frequency. This
set me thinking: perhaps some form of servo system could be devised
(either in hardware, controlling the tape recorder's motor, or in
software, post-processing the digital copy of the tape, so as to keep the
tone at exactly 15625 Hz?

Although Liz will say correcting at source is better, I suspect getting a
servo loop to work that fast and reliably on the tape motor would be quite
a difficult task if not impossible, so software post-processing probably a
better approach (the servo wouldn't be correcting _at source_ anyway - the
"damage" is done).
Are you sure the 15625 was from the original recording session? (Can you
see it moving up and down with the wow?) I've noticed a strong line
between 15 and 16 kHz on a vast number of tracks from YouTube, including
many where no video should have been anywhere near the source (e. g.
digitised old 78s!); I can only assume that they picked up something from,
say, a computer monitor. (Though those aren't at that low a frequency
these days, are they? Maybe there is - or was until recently - a
background 15625 [or US value, which is quite close] miasma in all
built-up areas!)
Was this just an intellectual thought, or is it a recording you'd really
like to retrieve: in other words, what was/is the music? We can probably
find you a recording; virtually everything is on YouTube these days (-:!

It's an intellectual thought-experiment. I think the music may have been
from the testcard ;-) I'd be impressed if I ever managed to identify it on
a compilation of library music.

And now I think about it, I may have recorded from an old 405-line TV, so
the whistle would have been at 10125 rather than 15625 Hz. Either way, when
I generated a live spectrum of the digital recording made from the tape,
there was a clear peak at around one frequency or the other, and when I
generated an offline spectrum of a few seconds of *stable* recording (ie not
when it was wowing all over the place), the peak narrowed enough to be able
to say that it was close to one value or the other. That was using CoolEdit
software, though I imagine that Audacity can do something similar.

It's a long time since I noticed the problem. I don't know which tape it was
recorded on or where the digital WAV file of it is.

[J. P. Gilliver]

[]

Post by NY

Post by J. P. Gilliver
But the sound would not have been coming from the cassette recorder's
speaker, would it - or were you monitoring, and using a very long mike
lead to avoid feedback? Most portable cassette machines mute the
speaker on record, at least when recording from a microphone. So
unless the TV was also running on battery ...

I wasn't implying that the sound was coming from the speaker *during
the recording*. I was meaning that when the tape was played back
afterwards, the pitch rose during loud sounds which equates to the
motor slowing down at those points during recording - assuming that on
playback the motor is able to run at constant speed.

Hmm. Although another has said the record current rises a _little_ with
large amplitude signals, I'd be surprised if it would have a noticeable
effect. (Having said that, I can't otherwise explain what you describe.)
[]

Post by NY

Post by J. P. Gilliver
Was this just an intellectual thought, or is it a recording you'd
really like to retrieve: in other words, what was/is the music? We
can probably find you a recording; virtually everything is on YouTube
these days (-:!

It's an intellectual thought-experiment. I think the music may have
been from the testcard ;-) I'd be impressed if I ever managed to
identify it on a compilation of library music.

I suspected as much, and am enjoying the discussion. But if you do seek
it: "The test card circle" put out a CD not _too_ long ago (meaning
somewhere in the last couple of decades, I think!) of "test card music",
arguing that some of it was of unrecognised (by the general public) high
quality and should not be lost. (Technically, they had a point. Of
course, I bought the CD - but can't say I can at this point think of
anything on it I consider the musical equivalent of deathless prose.)

Post by NY
And now I think about it, I may have recorded from an old 405-line TV,
so the whistle would have been at 10125 rather than 15625 Hz. Either

And as well as being more within normal hearing range anyway, tended to
be louder for various reasons - components looser, especially when
thermal cycling has had its effect.

Post by NY
way, when I generated a live spectrum of the digital recording made
from the tape, there was a clear peak at around one frequency or the
other, and when I generated an offline spectrum of a few seconds of
*stable* recording (ie not when it was wowing all over the place), the
peak narrowed enough to be able to say that it was close to one value
or the other. That was using CoolEdit software, though I imagine that
Audacity can do something similar.

(I think almost any audio processing software has such a function; I use
GoldWave, since I bought it before Audacity came out and am used to it.)

Post by NY
It's a long time since I noticed the problem. I don't know which tape
it was recorded on or where the digital WAV file of it is.

[Theo]

Post by NY
When I copied the tape to my PC (via its soundcard) much more recently,
I noticed something that I couldn't hear: the 15625 Hz line frequency.
This set me thinking: perhaps some form of servo system could be devised
(either in hardware, controlling the tape recorder's motor, or in
software, post-processing the digital copy of the tape, so as to keep
the tone at exactly 15625 Hz?

I don't think that would be difficult to do in software. Do an FFT based on
(say) a second's worth of samples, look for a peak at 15625 +/- X (eg
X=10%). Measure the frequency of the peak. Subtract the from 15625 to get
the sample rate correction. If you slide this across (do an FFT every 50ms
of the previous second) then it would be a fairly smoothly varying
correction that follows the wow of the tape. When you have the corrected
sample rate for each period, interpolate your source samples back to a
baseline constant 44.1kHz (for example) sampling rate.

Theo

[J. P. Gilliver]

In message <1qam54f.1j66ejt1hbbzggN%***@poppyrecords.invalid.invalid> at
Fri, 12 May 2023 10:36:05, Liz Tuddenham
[]

Post by Liz Tuddenham

Post by J. P. Gilliver
generation of sound engineers: wow. (Not sure why film seems
particularly prone to it: possibly heavy reels driven entirely by film
tension [but surely that applies to tape too]?)

Sometimes the wow comes from the original sound camera, but that is very
rare. One particularly bad example is the record of "Warsaw Concertto"
from the sound track of "Dangerous Moonlight", the pitch wavers very
noticeably.

You mean even if the film is played at a dead constant speed (say,
feeding from a loose box, or I suppose a telecine with proper speed
tracking), there is wow on it, from the original sound camera?
Interesting. I must admit I've only ever had recordings of the Warsaw
played by others - I think I've only seen the actual film once (on TV).

Post by Liz Tuddenham
More often it is because the transcription engineer isn't doing the job
properly.

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting

Oh, I couldn't agree more: not introducing an error is better than
trying to correct it afterwards, especially with something as
mathematically complex as wow from a mis-centred record. But sometimes
we're stuck with only the transcription (and probably in mp3 format
too).

Post by Liz Tuddenham
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

What do you look at? The run-out final groove would I guess be the best
reference, though only on records where it is centered, which on earlier
records it often deliberately isn't (I presume to trigger the auto-stop
mechanism on players).

[Brian Gaff]

Re centring of records. I'd guess that you could do it easily on records
where whole areas had the same groove pitch, but this varies on some between
tracks and at the start and end of course. I think you could get close
enough with a constant pitch of groove measured in several places, but
nudging the thing into exactly the right place is often long winded. I tried
it once or twice on a disc with the middle removed, and once you get it,
record it quick.
Another dodge for records that crackle a lot due to the muck that comes in
due to static, is to use one of those knowin record cleaning baths with
water and fairy liquid, Play the tracks you want while wet and the muck is
in suspension.
Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by J. P. Gilliver
Fri, 12 May 2023 10:36:05, Liz Tuddenham
[]

Post by Liz Tuddenham

Post by J. P. Gilliver
generation of sound engineers: wow. (Not sure why film seems
particularly prone to it: possibly heavy reels driven entirely by film
tension [but surely that applies to tape too]?)

Sometimes the wow comes from the original sound camera, but that is very
rare. One particularly bad example is the record of "Warsaw Concertto"
from the sound track of "Dangerous Moonlight", the pitch wavers very
noticeably.

You mean even if the film is played at a dead constant speed (say, feeding
from a loose box, or I suppose a telecine with proper speed tracking),
there is wow on it, from the original sound camera? Interesting. I must
admit I've only ever had recordings of the Warsaw played by others - I
think I've only seen the actual film once (on TV).

Post by Liz Tuddenham
More often it is because the transcription engineer isn't doing the job
properly.

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting

Oh, I couldn't agree more: not introducing an error is better than trying
to correct it afterwards, especially with something as mathematically
complex as wow from a mis-centred record. But sometimes we're stuck with
only the transcription (and probably in mp3 format too).

Post by Liz Tuddenham
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.

What do you look at? The run-out final groove would I guess be the best
reference, though only on records where it is centered, which on earlier
records it often deliberately isn't (I presume to trigger the auto-stop
mechanism on players).
--
"quidquid latine dictum sit, altum viditur". ("Anything is more impressive if
you say it in Latin")

[NY]

Post by Brian Gaff
Another dodge for records that crackle a lot due to the muck that comes
in due to static, is to use one of those knowin record cleaning baths with
water and fairy liquid, Play the tracks you want while wet and the muck is
in suspension.

Been there, done that. Need to clean the crap off the needle after copying
each track.

I remember adverts for record-cleaning devices which rotated the disc inside
a small hub which contained a soap solution and which partially enclosed a
sector of the disc which was rotated to clean the whole disc.

When I was copying some of my dad's old LPs, I found that some of them were
so bad that I needed to drip water onto each track in turn so there was a
"lake" covering the grooves, and playing with the needle immersed in that
water.

[Liz Tuddenham]

Post by NY

Post by Brian Gaff
Another dodge for records that crackle a lot due to the muck that comes
in due to static, is to use one of those knowin record cleaning baths with
water and fairy liquid, Play the tracks you want while wet and the muck is
in suspension.

Been there, done that. Need to clean the crap off the needle after copying
each track.
I remember adverts for record-cleaning devices which rotated the disc inside
a small hub which contained a soap solution and which partially enclosed a
sector of the disc which was rotated to clean the whole disc.
When I was copying some of my dad's old LPs, I found that some of them were
so bad that I needed to drip water onto each track in turn so there was a
"lake" covering the grooves, and playing with the needle immersed in that
water.

That is the best way; the surface noise becomes even worse if you allow
the record to dry out after wetting it.

Obviously, that is also the safest way of playing partially-decomposed
nitrate 'direct cut' discs - scraping a potentially explosive material
with a diamond is best done under water.

[Brian Gaff]

Young Gifted and Black was one single which had every copy off centre. Was
that Trojan? Anyway, to me it was part of the charm of the record. It was
also a bit overloaded.
I noticed when they started to make those albums with hits recorded by
others, they actually put a wow on the recording to make its sound like the
original. Of course the male lead on the Wollies copy would never have got
away with it nowadays, it was Elton John, billed as Reg on the record.
I recently heard an album cut of the original and there was no wow, and it
sounded wrong to me.
Another improved recording is Back in the USSR by the beatles. The new
version of the whit Album remixed by George Martins some has had the jet
plane sounds loop redone so it does not slip and slow down as it does on the
original. I also detected a better recording of a guitar on the start of
bungalow bill apparently re recorded by Paul on his melotron. However those
particular devices even when good were hardly pitch stable.
Brian
--
--:
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
***@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!

Post by Liz Tuddenham
[...]

Post by J. P. Gilliver
However, a lot of the old clips of singers who _could_ hold a note,
aren't - if they're from tapes, sometimes records, or particularly film
- because of that phenomenon possibly not familiar to the latest
generation of sound engineers: wow. (Not sure why film seems
particularly prone to it: possibly heavy reels driven entirely by film
tension [but surely that applies to tape too]?)

Sometimes the wow comes from the original sound camera, but that is very
rare. One particularly bad example is the record of "Warsaw Concertto"
from the sound track of "Dangerous Moonlight", the pitch wavers very
noticeably.
More often it is because the transcription engineer isn't doing the job
properly.

Post by J. P. Gilliver
I got to wondering, could autotune be used to rescue some of these old
clips from the worst excesses of wow?

The general rule is to tackle each type of distortion at source; getting
a disc so the grooves are central, regardless of wherethe centre hole
might be, is just part of the basic set-up when you are doing
transcription work. I have fitted my pick-up with an electronic
micrometer for that purpose, a groove microscope couldequally well be
used.
--
~ Liz Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk