In article , David Looser
wrote:
"TT" wrote in message
. au...



and then convert them to high bitrate MP3 it then removes all the
surface noise?

I can't say I'd noticed that converting to mp3 removes surface noise
(generally I don't convert to mp3), I'll give it a go.

I'd be wary of having the threashold level set high enough to do this.
Seems a bit like using a hammer in place of a screwdriver. :-)

But I'd be interested to hear what others may think of using perception
based reduction like MP3 encoding with the explicit aim of suppressing
background noise.

Slainte,

Jim

--
Change 'noise' to 'jcgl' if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

In article ,
TT wrote:
Have you noticed once you have removed the clicks and pops
(I do it physically by cutting them out) and then convert
them to high bitrate MP3 it then removes all the surface
noise? It really is marvellous way to clean up an LP.

Any time I've had a chance to compare original to a surface noise reduced
copy I prefer the original. Unless it is extremely early stuff with very
restricted bandwidth. Clicks are a different matter.

--
*I wish the buck stopped here. I could use a few.

Dave Plowman London SW
To e-mail, change noise into sound.

"Don Pearce" wrote in message
news:49c99e44.236997750@localhost...

I would have thought that reducing the level to zero would produce
almost as bad a click as letting it hit peak level.

No, not at all. That would only apply if the silent portion started abruptly
in the middle of a high-amplitude bit of the waveform, and as I said I
choose zero-crossings to start an end the silent portion. Do you remember
Garrad's "Music Recovery Module"? it was an analogue real-time click
remover. It also reduced the signal amplitude to zero for the duration of
the click because that was the technique that the designer found most
effective.

I find that using
the manual click removal in Audition (Cooledit as was) is mainly a
matter of getting the size of the chunk dead right. It appears to use
an averaging function based on what precedes and follows the
selection, followed by some sort of smooth window to make the
transitions good.

Indeed.

If the click is only on one channel, the option to
copy from the other channel seems to work well.

A useful dodge that, particularly with a mono disk (but captured in stereo).
Or average the two channels for the duration of the click when the click is
"vertical" only. Judicious use of FFT filtering for a few milliseconds
around the click can also make the click easier to remove or less audible.

If you are doing restoration on an audio file, never ever save to MP3.
It pretty much wrecks your chances of making significant improvements
later. WAV is the only way to go - disk space can hardly be an issue
any more.

I agree with that.

David.

"Dave Plowman (News)" wrote in message
...
In article ,
Iain Churches wrote:
I've often done that myself (using Cool Edit). As well as being more
convenient to listen to it also means that I can get rid of a lot of
the clicks and pops. Apart from that, though, there is no discernable
difference to the sound quality between the LP direct, and the CD copy.


That's because you have mastered it as it should be mastered - direct,
with no "post production improvemements" :-)

I doubt there are many commercial CDs which are a copy of the vinyl -
apart from rare stuff.

I don't think that is what David meant.
But there are quite a lot of CDs that were mastered
from metal matrices when analogue tapes were
not available.


But it does sort of prove how good the medium is - a properly done copy of
vinyl to CD will sound identical to the vinyl. The other way round not so.

Has this point ever been in dispute?

Iain

Jim Lesurf wrote:

In article , David Looser

[snip]

Vinyl is analogue, so any reference to "bitrates" is meaningless.

Not necessarily. As an information channel, Vinyl LP should have a Shannon
bandwidth expressible in bits per second. The difficulty is that the
channel behaviour in such a case is limited by distortion in quite a
complex manner, so determining the practical value is difficult.

A (plausible?) attempt at an answer to this is at :-

http://www.hydrogenaudio.org/forums/...hp/t35530.html

This does ignore any affects of distortion. To obtain a "ball park"
understanding of the different resolutions of vinyl and CD I would start by
assuming vinyl distortion to be small, but then it is a long time since I
last had anything to do with Shannon's law.

I did see the Gadget Show demo, the question it left me with was how
accurately they had matched the sound levels of the three samples.

--
David Pitt

"David Looser" wrote in message
...
"TT" wrote in message
. au...


Have you noticed once you have removed the clicks and pops (I do it
physically by cutting them out)

Do you mean you edit them out? ("physically" cutting them out implies
taking a razor blade to analogue tape, and I guess you don't actually mean
that!) The trouble with that is that it can produce a discontinuity in the
waveform which, itself, produces a click, and it slightly changes the
timing.

CoolEdit has a "click/pop eliminator" which works brilliantly on some
clicks and pops, but fails miserably on others. So I use a mixture of
techniques including CoolEdit's software click remover and editing out
clicks the way you do. But most commonly (assuming the software remover
fails) I reduce the signal level to zero for the duration of the click. By
going for the nearest zero-crossing I can avoid the discontinuity click,
and it preserves the timing. It can still leave an audible "hole" in the
audio, depending on the sort of programme material behind the click, but
generally I find it the least-worst option for the really difficult
clicks.

and then convert them to high bitrate MP3 it then removes all the surface
noise?

I can't say I'd noticed that converting to mp3 removes surface noise
(generally I don't convert to mp3), I'll give it a go.

David.

In Cool Edit Pro when I have a particularly bad click/pop I expand the wave
form out as far as I can an just cut the offending piece of noise. Since we
are talking about a very small period of time I have never noticed any
discontinuity to the resulting wave file. Hearing a missing
0.005sec-0.01sec piece of music missing is a bit of an ask ;-)

I find using any program that does this automatically just destroys the
music so I therefore do it manually.

Cheers TT

"Dave Plowman (News)" wrote in message
...
In article ,
TT wrote:
Have you noticed once you have removed the clicks and pops
(I do it physically by cutting them out) and then convert
them to high bitrate MP3 it then removes all the surface
noise? It really is marvellous way to clean up an LP.

Any time I've had a chance to compare original to a surface noise reduced
copy I prefer the original. Unless it is extremely early stuff with very
restricted bandwidth. Clicks are a different matter.

I totally agree. I would prefer to have a small amount of surface noise and
retain the dynamics rather than no surface noise and a dull and lifeless
recording.

But what I mentioned above about MP3 is to do with badly worn records that
have already seen better days and the surface noise is bad.

Cheers TT

"David Looser" wrote in
message

But most commonly (assuming the software
remover fails) I reduce the signal level to zero for the
duration of the click. By going for the nearest
zero-crossing I can avoid the discontinuity click, and it
preserves the timing. It can still leave an audible
"hole" in the audio, depending on the sort of programme
material behind the click, but generally I find it the

IME a far better approach with CEP/Audition is to apply an appropriate low
pass filter over the area that was afflicted with the tic, which is usually
a few milliseconds or less. I use corner frequencies on the order of a few
100 Hz in severe cases, to several kHz in mild cases. This avoids the
zero-crossing issue.

"David Pitt" wrote in message

Jim Lesurf wrote:

In article , David
Looser

[snip]

Vinyl is analogue, so any reference to "bitrates" is
meaningless.

Not necessarily. As an information channel, Vinyl LP
should have a Shannon bandwidth expressible in bits per
second.

Agreed. Both analog and digital channels have effective bandwiths, which can
be interpreted as sample rates.

The difficulty is that the channel behaviour in
such a case is limited by distortion in quite a complex
manner, so determining the practical value is difficult.

Expanding on that a bit...

Vinyl is inherently distorted at high frequencies and high amplitudes. There
is inherent geomtric distortion due to the difference between the shape of
the cutting stylus and the playback stylus. There is additional deformation
of the groove wall due to high inertial forces. The playback device itself
has trackability problems which generally increase with decreasing price.
We're not talking about 0.01% distoriton, the nonlinear distortion is up in
the 3-10% or higher range. The harmonics that are created by the
nonlinearity are usually in the ultrasonic range, but the IM products
splatter all over the audio band.

Contrast this with CD's linear PCM which is inherently distortion free at
all levels right up to 0.001 dB below clipping.

Therefore, the bandwidth of vinyl is very dependent on amplitude. High
amplitude signals have lower effective bandwidth.

So, comparing the bandwidth of LPs to CDs will be dependent on the criteria
set for reproducing high frequencies with low distortion and good sound
quality.



A (plausible?) attempt at an answer to this is at :-

http://www.hydrogenaudio.org/forums/...hp/t35530.html

This does ignore any affects of distortion.

It understates the fact that tracing distortion inherent in the LP format is
a big issue when you go much above 5-8 Khz with good quality playback
equipment. The performance of mainstream vinyl players in the days of was
well short of that.

To obtain a
"ball park" understanding of the different resolutions of
vinyl and CD I would start by assuming vinyl distortion
to be small, but then it is a long time since I last had
anything to do with Shannon's law.

The LP advocates who put the bandwidh of vinyl above 20 KHz are poorly
informed. 10 KHz might me a more reasoanble number.


I did see the Gadget Show demo, the question it left me
with was how accurately they had matched the sound levels
of the three samples.

Matching levels and time synching vinyl/digital comparisons is not trivial
and is rarely done well. I've seen it done well and the effort and skill
levels required are well beyond TV journalists and even the more technical
of audiophiles. We used audio professsionals for much of our work of that
nature.

"Dave Plowman (News)" wrote in message
...
In article ,
TT wrote:
Have you noticed once you have removed the clicks and pops
(I do it physically by cutting them out) and then convert
them to high bitrate MP3 it then removes all the surface
noise? It really is marvellous way to clean up an LP.

Any time I've had a chance to compare original to a surface noise reduced
copy I prefer the original.

Cedar makes a very good job of this. A client has compared it
to Windolene:-)

People ofte get the impression that the HF has been reduced when
the surface noise is taken away. There are many early recordings
too in which you can subsequently hear instruments you didn't
know where the-)

Iain