Convert mono LP to digital

[Nil[_2_]]

This was discussed here some time ago, I believe, and I thought I saved
the instructions, but I can't find them, so I hope for some advice...

I want to digitalize a mono LP. I record it to the computer as a stereo
WAV file. What's the best way to make it a mono WAV file? My usual
practice has been to pick the best-sounding/least-noisy channel and
eliminate the other. But I recall someone here recommending doing
something like inverting one channel and then summing them in order to
cancel out more noise. That's not it, but something like that.

Can someone please steer me in a good direction?

[geoff]

On 5/10/2015 3:03 p.m., Nil wrote:
> This was discussed here some time ago, I believe, and I thought I saved
> the instructions, but I can't find them, so I hope for some advice...
>
> I want to digitalize a mono LP. I record it to the computer as a stereo
> WAV file. What's the best way to make it a mono WAV file? My usual
> practice has been to pick the best-sounding/least-noisy channel and
> eliminate the other. But I recall someone here recommending doing
> something like inverting one channel and then summing them in order to
> cancel out more noise. That's not it, but something like that.
>
> Can someone please steer me in a good direction?

Do that and you will also cancel out the music.

I suggest try summing them and compare the result with each separate
channel, and choose whichever gives the best result.

Results may differ slightly for each different project , due the the
potentially random nature of the noise per channel.

geoff

[Nil[_2_]]

On 04 Oct 2015, geoff > wrote in
rec.audio.pro:

> I suggest try summing them and compare the result with each
> separate channel, and choose whichever gives the best result.

OK.

> Results may differ slightly for each different project , due the
> the potentially random nature of the noise per channel.

It has a lot to do with the condition of the LP. If it's in bad shape,
it's likely that one channel is a lot more worn and noisy than the
other, especially toward the inner grooves. One channel often sounds
significantly better than the other. The one I'm working on right now
is in pretty good shape, so summing them might work well. But I know
there was an invert/summing trick that was mentioned here, and I tried
it and it seemed to work well for the source I tried it on. If I could
only remember what it was...

[Trevor]

On 5/10/2015 2:33 PM, geoff wrote:
> On 5/10/2015 3:03 p.m., Nil wrote:
>> This was discussed here some time ago, I believe, and I thought I saved
>> the instructions, but I can't find them, so I hope for some advice...
>>
>> I want to digitalize a mono LP. I record it to the computer as a stereo
>> WAV file. What's the best way to make it a mono WAV file? My usual
>> practice has been to pick the best-sounding/least-noisy channel and
>> eliminate the other. But I recall someone here recommending doing
>> something like inverting one channel and then summing them in order to
>> cancel out more noise. That's not it, but something like that.
>>
>> Can someone please steer me in a good direction?
>
> Do that and you will also cancel out the music.

And should leave some noise/distortion not equal in both channels.
What you'd have to do then is sum the channels again in phase, then
invert polarity, and finally sum those two results to cancel the
noise/distortion of step one, from the 2 channel sum.
I've never tried it, but it might help. Of course any noise/distortion
equal in both channels will not be removed, and any noise/distortion in
phase but not equal will only be partially removed, along with some
music not in phase or not equal level in both channels. If the cartridge
is not perfectly aligned you are likely to remove more actual HF music
content than you would want IMO.

Trevor.

[Peter Larsen[_3_]]

On 05-10-2015 03:03, Nil wrote:

> This was discussed here some time ago, I believe, and I thought I saved
> the instructions, but I can't find them, so I hope for some advice...

> I want to digitalize a mono LP. I record it to the computer as a stereo
> WAV file. What's the best way to make it a mono WAV file?

Sum the channels and be happy, but declick large clicks first, leave the
crackle and the small ones, they are proof of authenticity and great for
testing AD conversion and loudspeakers and their time domain behavior.
If one is dramatically better sounding than the other because of
previous playback with misaligned crap cartridge - a lot of cheap
grammophones offered that as a design feature - then select the best
sounding channel.

> My usual
> practice has been to pick the best-sounding/least-noisy channel and
> eliminate the other. But I recall someone here recommending doing
> something like inverting one channel and then summing them in order to
> cancel out more noise. That's not it, but something like that.

If you are really in trouble with the sound - applies also to 78 rpm -
try a different tip size, I had one problem lp that needed a poorer
cartridge, otherwise it was unbearably noisy.

If all else fails, you could use something akin to the center channel
extract function of Audition, but beware, it does add artifacts similar
to noise reduction artifacts.

The analog version is to convert to sum and difference and keep only the
sum track. That too focuses on what is common and discards what is
different.

> Can someone please steer me in a good direction?

You are on the track already by making a stereo transfer! - try the MS
stereo conversion first, probably the best idea, but only the actual
transfer will tell you which of the above works best.

The record should be clean. If you wash it with dishwasher, then rinse
with slightly acetic water, it will make the water pearl off of the
vinyl and easy to dab off with a cloth and it will remove calcium
deposits if there are some from previous incorrect washing. You can then
washs remaining water off with a record cleaning implement with pure
isopropylic alcohol.

Do if avaliable spray a quality antistatic record cosmetic antistatic
spray on or you will go mad from small noises. QED made some way long
time that are in fact very good.

Kind regards

Peter Larsen

[Phil Allison[_4_]]

Nil wrote:
>
>
>
> I want to digitalize a mono LP. I record it to the computer as a stereo
> WAV file. What's the best way to make it a mono WAV file? My usual
> practice has been to pick the best-sounding/least-noisy channel and
> eliminate the other. But I recall someone here recommending doing
> something like inverting one channel and then summing them in order to
> cancel out more noise. That's not it, but something like that.
>

** Stereo LPs are cut using a 45 /45 degree system that puts Left and Right channel signals on opposite sides of the grove. Also, the phase of one is reversed so that a mono signal results in purely lateral excursions of the cutter. This ensures mono compatibility and is put right by internally reverse wiring one channel of stereo pickups.

The problem with mono cut LPs is that they have mostly *stereo* noise on them because contamination & damage to each side of the groove is not identical.

Summing reverse phase, L and R signals from a mono LP removes the sound and what remains is an out of phase mix of the previous two noises. Summing this with a L + R signal may cancel one of them, but boost the other.

Worth a try maybe.


..... Phil





Simply summing L and R signals will cancel low frequency rumble from the turntable.

[Mike Rivers[_2_]]

On 10/4/2015 10:03 PM, Nil wrote:

> I want to digitalize a mono LP. I record it to the computer as a stereo
> WAV file. What's the best way to make it a mono WAV file?

Unless you're trying to make the smallest file, you're probably better
off ending up with a stereo file with both channels identical. It's less
confusing. I take it that your goal is to clean up the sound of the
original record in the process of digitizing it.

> My usual
> practice has been to pick the best-sounding/least-noisy channel and
> eliminate the other. But I recall someone here recommending doing
> something like inverting one channel and then summing them in order to
> cancel out more noise. That's not it, but something like that.

No, you don't want to do that. It'll leave only what's different between
the channels, essentially only the errors with your phono cartridge and
the crackles on one channel that aren't simultaneously on the other
channel. But there are a number of different "restoration" programs and
plug-ins, some even free, that you can try. No doubt this thread will
grow a collection of recommendations.


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

[Scott Dorsey]

Nil > wrote:
>I want to digitalize a mono LP. I record it to the computer as a stereo
>WAV file. What's the best way to make it a mono WAV file? My usual
>practice has been to pick the best-sounding/least-noisy channel and
>eliminate the other. But I recall someone here recommending doing
>something like inverting one channel and then summing them in order to
>cancel out more noise. That's not it, but something like that.
>
>Can someone please steer me in a good direction?

Well, if you do it right, you'll play it back with a mono cartridge, then you
only have to eliminate one channel in your DAW.

If you do it with a stereo cartridge, you can sum the two channels to mono
and get a result that is pretty mono. How good this is depends a lot on
how well the channels of your preamp match.

If you are doing decrackling, you may want to do it before summing to mono.
--scott

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

[Edi Zubovic]

On Sun, 04 Oct 2015 22:03:17 -0400, Nil
> wrote:

>This was discussed here some time ago, I believe, and I thought I saved
>the instructions, but I can't find them, so I hope for some advice...
>
>I want to digitalize a mono LP. I record it to the computer as a stereo
>WAV file. What's the best way to make it a mono WAV file? My usual
>practice has been to pick the best-sounding/least-noisy channel and
>eliminate the other. But I recall someone here recommending doing
>something like inverting one channel and then summing them in order to
>cancel out more noise. That's not it, but something like that.
>
>Can someone please steer me in a good direction?

-- I'm doing it in two ways: either record in the mono LP with a
stereo cartridge or use mono cartridge (internally wired but with
maintained vertical compliance). The third way would be using a
dedicated mono cartridge like old EMT (ie. Ortofon) OFS(aphire) or
OFD(iamond) cartridges. Today, Ortofon SPU fall in this category, I
think.

The last ones are dangerous to all stereo records as they have no
vertical impedance and usually require a hefty tracking force, so they
might damage a stereo record . With mono records, a greater tracking
force is no bad thing at all, though.

The EMT and Ortofons have a L-shaped cantilever, the foot holds the
stylus, vertical part holds a single coil and ends up in a kind of
partially elastic bearing which is held by a strip of Tesa cellophane
band.


In all cases, I'm using a dynamic - moving coil - cartridge. Older
LPs play well with a 1 mil stylus or somewhat less.

More recent microgroove records, say mid-60s onwards, would require
that less-than-one-mil styli.

A simple setup is the cartridge output into step-up transformers and
then directly to the sound card inputs, recording istr best made with
a dedicatred program such as Diamond Cut, which has already many
equalization curves etc. suitable for many records.

I am using stereo information to do a total cancelling out and I never
do it other ways. All music is horizontal here, get out all vertical.

Edi Zubovic, Crikvenica, Croatia

[Edi Zubovic]

On Mon, 05 Oct 2015 16:13:47 +0200, Edi Zubovic <edi.zubovic[rem
> wrote:


>
>The last ones are dangerous to all stereo records as they have no
vertical impedance ---------compliance-------
>and usually require a hefty tracking force, so they
>might damage a stereo record . With mono records, a greater tracking
>force is no bad thing at all, though.
>
>The EMT and Ortofons have a L-shaped cantilever, the foot holds the
>stylus, vertical part holds a single coil and ends up in a kind of
>partially elastic bearing which is held by a strip of Tesa cellophane
>band.
>
Sorry, swift fingers.

Edi Zubovic, Crikvenica, Croatia

[PStamler]

On Monday, October 5, 2015 at 3:08:53 AM UTC-5, Phil Allison wrote:

> Simply summing L and R signals will cancel low frequency rumble from the turntable.

And from the cutting lathe.

If the stereo recording has equal level in the two channels, simply summing them will do the job. You'll probably need to filter out infrasonic noise; a 3rd-order filter at 16Hz will do that. The scientific filters in Audition will do the job. Record in 24 bits (or 32 bits floating); after you've done everything you want to do to the signal, convert to a6 bits for burning a CD.

What's the content on the record you want to convert?

Peace,
Paul

[Scott Dorsey]

Edi Zubovic <edi.zubovic[rem > wrote:
>
>-- I'm doing it in two ways: either record in the mono LP with a
>stereo cartridge or use mono cartridge (internally wired but with
>maintained vertical compliance). The third way would be using a
>dedicated mono cartridge like old EMT (ie. Ortofon) OFS(aphire) or
>OFD(iamond) cartridges. Today, Ortofon SPU fall in this category, I
>think.

The problem with the new Ortofon mono cartridges is that they are only
available with spherical styli, so you have all the tracking issues that
are inherent with that. But the extreme vertical stiffness is a big win
for mono records.

> In all cases, I'm using a dynamic - moving coil - cartridge. Older
>LPs play well with a 1 mil stylus or somewhat less.
>
>More recent microgroove records, say mid-60s onwards, would require
>that less-than-one-mil styli.

I'm not sure I buy that in part because it depends on how you measure it.
The original standard for mono recordings was 1.0 mil spherical, but if you
use an elliptical stylus you'll find a slightly smaller stylus will ride at
the same point in the groove.

Which is part of why I think the fineline stylus is such a win when it is
possible to use them; they can ride evenly in a wide range of groove widths.
--scott

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

[Edi Zubovic]

-----------8<--------------------
>I'm not sure I buy that in part because it depends on how you measure it.
>The original standard for mono recordings was 1.0 mil spherical, but if you
>use an elliptical stylus you'll find a slightly smaller stylus will ride at
>the same point in the groove.
>
>Which is part of why I think the fineline stylus is such a win when it is
>possible to use them; they can ride evenly in a wide range of groove widths.
>--scott

Yes, fineline styli have a profile resembling a chisel with long
contact areas and I find them superior for most of LPs especially
those thin ones, after oil crisis in the 70s . That was also the time
when improved cutting allowed more densely packed grooves ie. more
time without too much sacrifice in dynamics. You need something extra
light and a really big compliance for those. Yet I have some
mirror-like LPs with really unsatisfactorily dynamics (eg. Probe
Records, Three Dog Night - Seven Separate Fools). I'd be the 8th one
if I wouldn't treat this record with special care. It seems to me that
merely a fingerprint is a big issue there. What a difference compared
to older 180 g LPs cut to be loud enough.

I had a strange thing with a Shure 15 VXMR some ten or more years ago
when I found burnt vinyl at the stylus tip and had to carefully scrape
it off with a razor blade. I suspect a tracking issue with my Dual
1218 tonearm. I haven't seen that later.

Hovewer, for standard mono LPs, those with M in an inverted triangle,
you can use a 1 mil stylus and set the tracking force quite lavishly -
they were made for that. If there's mistracking, you'll not only hear
it, you'll actually see it. Heh.

Edi Zubovic, Crikvenica, Croatia

[Paul Babiak]

On 10/05/2015 03:14 PM, Scott Dorsey wrote:
> Edi Zubovic <edi.zubovic[rem > wrote:
>>
>> -- I'm doing it in two ways: either record in the mono LP with a
>> stereo cartridge or use mono cartridge (internally wired but with
>> maintained vertical compliance). The third way would be using a
>> dedicated mono cartridge like old EMT (ie. Ortofon) OFS(aphire) or
>> OFD(iamond) cartridges. Today, Ortofon SPU fall in this category, I
>> think.
>
> The problem with the new Ortofon mono cartridges is that they are only
> available with spherical styli, so you have all the tracking issues that
> are inherent with that. But the extreme vertical stiffness is a big win
> for mono records.
>
>> In all cases, I'm using a dynamic - moving coil - cartridge. Older
>> LPs play well with a 1 mil stylus or somewhat less.
>>
>> More recent microgroove records, say mid-60s onwards, would require
>> that less-than-one-mil styli.
>
> I'm not sure I buy that in part because it depends on how you measure it.
> The original standard for mono recordings was 1.0 mil spherical, but if you
> use an elliptical stylus you'll find a slightly smaller stylus will ride at
> the same point in the groove.
>
> Which is part of why I think the fineline stylus is such a win when it is
> possible to use them; they can ride evenly in a wide range of groove widths.
> --scott
>

I have no direct experience with the Denon DL-102 mono cartridge, but it
looks good on paper, or at least on glass...

http://www3.sympatico.ca/murraya/DenonMonoPage.htm

Quote:

Basically, the DL-102 is a high output MONO moving coil cartridge which
has incorporated both vertical compliance and a 0.7 mil radius stylus,
making it compatible with stereo LP playback and is intended for
playback of both Mono and Stereo Records.

The DL-102 is NOT a stereo cartridge strapped internally for mono, but
is specially designed to output a mono signal from a stereo LP, so there
is no danger of damaging a modern LP by its use. Denon's own
instruction sheet states "The DL-102 is a monophonic output moving coil
cartridge designed for monophonic replay as well as the monophonic
playback of stereo recordings."

Paul

[None]

"Nil" > wrote in message
...
> This was discussed here some time ago, I believe, and I thought I
> saved
> the instructions, but I can't find them, so I hope for some
> advice...
>
> I want to digitalize a mono LP. I record it to the computer as a
> stereo
> WAV file. What's the best way to make it a mono WAV file? My usual
> practice has been to pick the best-sounding/least-noisy channel and
> eliminate the other. But I recall someone here recommending doing
> something like inverting one channel and then summing them in order
> to
> cancel out more noise. That's not it, but something like that.
>
> Can someone please steer me in a good direction?

If you make sure the two tracks have the proper polarity and
time-alignment, and you sum them, you get essentially the same signal,
6 dB hotter. Noise that's not correlated between the two tracks will
typically yield an average of 3 dB hotter when summed. So by summing
the tracks, you typically get a 3 dB improvement in S/N, for those
types of noise. Noise that's the same in both tracks, such as noise
from a mono master recording, will not be improved, it will increase
by the same 6 dB as the signal. For noise that is is in one track
only, you get 6 dB improvement in S/N by summing, but if you take only
the track without the noise, even better.

It pays to do some investigation on the kinds of noise that exist, and
how they do or do not correlate between channels.

[Edi Zubovic]

On Mon, 5 Oct 2015 16:33:11 -0400, "None" > wrote:

>"Nil" > wrote in message
...
>> This was discussed here some time ago, I believe, and I thought I
>> saved
>> the instructions, but I can't find them, so I hope for some
>> advice...
>>
>> I want to digitalize a mono LP. I record it to the computer as a
>> stereo
>> WAV file. What's the best way to make it a mono WAV file? My usual
>> practice has been to pick the best-sounding/least-noisy channel and
>> eliminate the other. But I recall someone here recommending doing
>> something like inverting one channel and then summing them in order
>> to
>> cancel out more noise. That's not it, but something like that.
>>
>> Can someone please steer me in a good direction?
>
>If you make sure the two tracks have the proper polarity and
>time-alignment, and you sum them, you get essentially the same signal,
>6 dB hotter. Noise that's not correlated between the two tracks will
>typically yield an average of 3 dB hotter when summed. So by summing
>the tracks, you typically get a 3 dB improvement in S/N, for those
>types of noise. Noise that's the same in both tracks, such as noise
>from a mono master recording, will not be improved, it will increase
>by the same 6 dB as the signal. For noise that is is in one track
>only, you get 6 dB improvement in S/N by summing, but if you take only
>the track without the noise, even better.
>
>It pays to do some investigation on the kinds of noise that exist, and
>how they do or do not correlate between channels.

It's about differences, some are correlated and some not. A simple
summing up is not sufficient enough for me, that's why I prefer stereo
cartridges for mono records. Even for 78RPM, Shure has its M78S, I use
its green stylus in the body of a V15. It's a 0,75 mil stylus I think,
a good all-rounder for 78s
Record stereo, do the m/s matrix and then sum it up. There's an old
Waves tool, I think "S1 - No Shuffler", a part of their DirectX
bundle. There is also a m/s matrix tool. You can do all kinds of
summing up with it, usually with "Width" = 0 and "Asymmetry" either
center or -90, you can invert with "Input Mode" etc. Very handy for
experimienting with.

Edi Zubovic, Crikvenica, Croatia

[Scott Dorsey]

Edi Zubovic <edi.zubovic[rem > wrote:
>Yes, fineline styli have a profile resembling a chisel with long
>contact areas and I find them superior for most of LPs especially
>those thin ones, after oil crisis in the 70s . That was also the time
>when improved cutting allowed more densely packed grooves ie. more
>time without too much sacrifice in dynamics. You need something extra
>light and a really big compliance for those. Yet I have some
>mirror-like LPs with really unsatisfactorily dynamics (eg. Probe
>Records, Three Dog Night - Seven Separate Fools). I'd be the 8th one
>if I wouldn't treat this record with special care. It seems to me that
>merely a fingerprint is a big issue there. What a difference compared
>to older 180 g LPs cut to be loud enough.

People want to put a whole symphony on one record, and they have to
sacrifice something, and level is the first thing to go. The more time,
the less excursion.

>I had a strange thing with a Shure 15 VXMR some ten or more years ago
>when I found burnt vinyl at the stylus tip and had to carefully scrape
>it off with a razor blade. I suspect a tracking issue with my Dual
>1218 tonearm. I haven't seen that later.

I would be very very suspicious of those Dual things. As autochangers
went they were some of the best but you sacrifice a lot with your arm
design in order to make that mechanism work. They always seemed very
resonant to me.

>Hovewer, for standard mono LPs, those with M in an inverted triangle,
>you can use a 1 mil stylus and set the tracking force quite lavishly -
>they were made for that. If there's mistracking, you'll not only hear
>it, you'll actually see it. Heh.

We call this "Gradoing" as in "since I bought those cheap tires, I can't
keep my car from gradoing all over the road."
--scott

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

[Scott Dorsey]

Paul Babiak > wrote:
>
>I have no direct experience with the Denon DL-102 mono cartridge, but it
>looks good on paper, or at least on glass...
>
>http://www3.sympatico.ca/murraya/DenonMonoPage.htm
>
>Quote:
>
>Basically, the DL-102 is a high output MONO moving coil cartridge which
>has incorporated both vertical compliance and a 0.7 mil radius stylus,
>making it compatible with stereo LP playback and is intended for
>playback of both Mono and Stereo Records.

I don't want high vertical compliance! I want to to be hard as a rock in
the vertical plane!
--scott


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

[geoff]

On 6/10/2015 10:44 a.m., Scott Dorsey wrote:
> Paul Babiak > wrote:
>> I have no direct experience with the Denon DL-102 mono cartridge, but it
>> looks good on paper, or at least on glass...
>>
>> http://www3.sympatico.ca/murraya/DenonMonoPage.htm
>>
>> Quote:
>>
>> Basically, the DL-102 is a high output MONO moving coil cartridge which
>> has incorporated both vertical compliance and a 0.7 mil radius stylus,
>> making it compatible with stereo LP playback and is intended for
>> playback of both Mono and Stereo Records.
> I don't want high vertical compliance! I want to to be hard as a rock in
> the vertical plane!
> --scott
>
>

Not horizontal plane then 8=====o

There is a "brand new pill" for that these days, if required (courtesy
Ry Cooder 'Look At Granny Run Run' - before it's time ! )

geoff


geoff

[Trevor]

On 6/10/2015 5:29 AM, PStamler wrote:
> Record in 24 bits (or 32 bits floating); after you've done everything you want to do to the
> signal, convert to 16 bits for burning a CD.

Had to smile that you think you will need more than 96dB DNR to record
old vinyl :-)
(Most DAWS work internally at 32bit floating, or better now anyway)
Still, won't hurt at least.

Trevor.

[PStamler]

On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
> On 6/10/2015 5:29 AM, PStamler wrote:
> > Record in 24 bits (or 32 bits floating); after you've done everything you want to do to the
> > signal, convert to 16 bits for burning a CD.
>
> Had to smile that you think you will need more than 96dB DNR to record
> old vinyl :-)
> (Most DAWS work internally at 32bit floating, or better now anyway)
> Still, won't hurt at least.

That's a long and complicated discussion, which I don't have time for while the rice is on the stove. Suffice it to say that every declicking algorithm I've ever used works more effectively, with fewer artifacts, on 24-bit files (or the 32-bit version of them) than on 16. Worth using just on that basis.

Peace,
Paul

[Trevor]

On 6/10/2015 1:03 PM, PStamler wrote:
> On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>> On 6/10/2015 5:29 AM, PStamler wrote:
>>> Record in 24 bits (or 32 bits floating); after you've done
>>> everything you want to do to the signal, convert to 16 bits for
>>> burning a CD.
>>
>> Had to smile that you think you will need more than 96dB DNR to
>> record old vinyl :-) (Most DAWS work internally at 32bit floating,
>> or better now anyway) Still, won't hurt at least.
>
> That's a long and complicated discussion, which I don't have time for
> while the rice is on the stove. Suffice it to say that every
> declicking algorithm I've ever used works more effectively, with
> fewer artifacts, on 24-bit files (or the 32-bit version of them) than
> on 16. Worth using just on that basis.

OK, I'll take your word for that, but would only be because the
declicking algorithm is badly written. As I said, most DAW's and plug
ins work in 32 bit float or better now. Just because the original file
starts out with 12 bits of extra noise instead of 4 shouldn't make any
difference.

Trevor.

[Randy Yates[_2_]]

Trevor > writes:

> On 6/10/2015 1:03 PM, PStamler wrote:
>> On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>>> On 6/10/2015 5:29 AM, PStamler wrote:
>>>> Record in 24 bits (or 32 bits floating); after you've done
>>>> everything you want to do to the signal, convert to 16 bits for
>>>> burning a CD.
>>>
>>> Had to smile that you think you will need more than 96dB DNR to
>>> record old vinyl :-) (Most DAWS work internally at 32bit floating,
>>> or better now anyway) Still, won't hurt at least.
>>
>> That's a long and complicated discussion, which I don't have time for
>> while the rice is on the stove. Suffice it to say that every
>> declicking algorithm I've ever used works more effectively, with
>> fewer artifacts, on 24-bit files (or the 32-bit version of them) than
>> on 16. Worth using just on that basis.
>
> OK, I'll take your word for that, but would only be because the
> declicking algorithm is badly written.

I'm not sure I'd agree with that. More bandwidth means more information,
especially for a high-frequency event like a click.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Randy Yates[_2_]]

Randy Yates > writes:

> Trevor > writes:
>
>> On 6/10/2015 1:03 PM, PStamler wrote:
>>> On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>>>> On 6/10/2015 5:29 AM, PStamler wrote:
>>>>> Record in 24 bits (or 32 bits floating); after you've done
>>>>> everything you want to do to the signal, convert to 16 bits for
>>>>> burning a CD.
>>>>
>>>> Had to smile that you think you will need more than 96dB DNR to
>>>> record old vinyl :-) (Most DAWS work internally at 32bit floating,
>>>> or better now anyway) Still, won't hurt at least.
>>>
>>> That's a long and complicated discussion, which I don't have time for
>>> while the rice is on the stove. Suffice it to say that every
>>> declicking algorithm I've ever used works more effectively, with
>>> fewer artifacts, on 24-bit files (or the 32-bit version of them) than
>>> on 16. Worth using just on that basis.
>>
>> OK, I'll take your word for that, but would only be because the
>> declicking algorithm is badly written.
>
> I'm not sure I'd agree with that. More bandwidth means more information,
> especially for a high-frequency event like a click.

Whoa. Serious brain fart. (Who said anything about more bandwidth.) Excuse me.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Scott Dorsey]

PStamler > wrote:
>On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>> On 6/10/2015 5:29 AM, PStamler wrote:
>> > Record in 24 bits (or 32 bits floating); after you've done everything y=
>ou want to do to the
>> > signal, convert to 16 bits for burning a CD.
>>=20
>> Had to smile that you think you will need more than 96dB DNR to record=20
>> old vinyl :-)
>> (Most DAWS work internally at 32bit floating, or better now anyway)
>> Still, won't hurt at least.
>
>That's a long and complicated discussion, which I don't have time for while=
> the rice is on the stove. Suffice it to say that every declicking algorith=
>m I've ever used works more effectively, with fewer artifacts, on 24-bit fi=
>les (or the 32-bit version of them) than on 16. Worth using just on that ba=
>sis.

That's weird. I have noticed dramatic improvements in declicking from using
higher sampling rates and wideband preamps; even though there isn't much signal
at 30 KHz there is enough to make for a better-defined edge detection.

But I have not noticed improvements from the longer sample size.

Now... I would most definitely believe that the longer sample size would be
a huge win if you're doing the RIAA de-emphasis in software, since the needed
dynamic range in that case is pretty huge. Even with the de-emphasis, if
your preamp is very wideband you may see that half of your dynamic range is
being eaten up by rumble that you're just going to filter anyway, so there is
a need to have considerably more dynamic range than the record itself may
have usable.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Trevor]

On 7/10/2015 4:09 AM, Randy Yates wrote:
> Trevor > writes:
>
>> On 6/10/2015 1:03 PM, PStamler wrote:
>>> On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>>>> On 6/10/2015 5:29 AM, PStamler wrote:
>>>>> Record in 24 bits (or 32 bits floating); after you've done
>>>>> everything you want to do to the signal, convert to 16 bits for
>>>>> burning a CD.
>>>>
>>>> Had to smile that you think you will need more than 96dB DNR to
>>>> record old vinyl :-) (Most DAWS work internally at 32bit floating,
>>>> or better now anyway) Still, won't hurt at least.
>>>
>>> That's a long and complicated discussion, which I don't have time for
>>> while the rice is on the stove. Suffice it to say that every
>>> declicking algorithm I've ever used works more effectively, with
>>> fewer artifacts, on 24-bit files (or the 32-bit version of them) than
>>> on 16. Worth using just on that basis.
>>
>> OK, I'll take your word for that, but would only be because the
>> declicking algorithm is badly written.
>
> I'm not sure I'd agree with that. More bandwidth means more information,
> especially for a high-frequency event like a click.

More *bandwidth* perhaps, but NOT storing 12 extra bits of *noise*.
(you do realise the two are not the same right?)
That just doesn't buy you anything with modern software processing.

Trevor.

[Trevor]

On 7/10/2015 4:10 AM, Randy Yates wrote:
> Randy Yates > writes:
>
>> Trevor > writes:
>>
>>> On 6/10/2015 1:03 PM, PStamler wrote:
>>>> On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>>>>> On 6/10/2015 5:29 AM, PStamler wrote:
>>>>>> Record in 24 bits (or 32 bits floating); after you've done
>>>>>> everything you want to do to the signal, convert to 16 bits for
>>>>>> burning a CD.
>>>>>
>>>>> Had to smile that you think you will need more than 96dB DNR to
>>>>> record old vinyl :-) (Most DAWS work internally at 32bit floating,
>>>>> or better now anyway) Still, won't hurt at least.
>>>>
>>>> That's a long and complicated discussion, which I don't have time for
>>>> while the rice is on the stove. Suffice it to say that every
>>>> declicking algorithm I've ever used works more effectively, with
>>>> fewer artifacts, on 24-bit files (or the 32-bit version of them) than
>>>> on 16. Worth using just on that basis.
>>>
>>> OK, I'll take your word for that, but would only be because the
>>> declicking algorithm is badly written.
>>
>> I'm not sure I'd agree with that. More bandwidth means more information,
>> especially for a high-frequency event like a click.
>
> Whoa. Serious brain fart. (Who said anything about more bandwidth.) Excuse me.

Right, ignore my previous reply then!

Trevor.

[Randy Yates[_2_]]

(Scott Dorsey) writes:

> PStamler > wrote:
>>On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>>> On 6/10/2015 5:29 AM, PStamler wrote:
>>> > Record in 24 bits (or 32 bits floating); after you've done everything y=
>>ou want to do to the
>>> > signal, convert to 16 bits for burning a CD.
>>>=20
>>> Had to smile that you think you will need more than 96dB DNR to record=20
>>> old vinyl :-)
>>> (Most DAWS work internally at 32bit floating, or better now anyway)
>>> Still, won't hurt at least.
>>
>>That's a long and complicated discussion, which I don't have time for while=
>> the rice is on the stove. Suffice it to say that every declicking algorith=
>>m I've ever used works more effectively, with fewer artifacts, on 24-bit fi=
>>les (or the 32-bit version of them) than on 16. Worth using just on that ba=
>>sis.
>
> That's weird. I have noticed dramatic improvements in declicking from using
> higher sampling rates and wideband preamps; even though there isn't much signal
> at 30 KHz there is enough to make for a better-defined edge detection.
>
> But I have not noticed improvements from the longer sample size.
>
> Now... I would most definitely believe that the longer sample size would be
> a huge win if you're doing the RIAA de-emphasis in software, since the needed
> dynamic range in that case is pretty huge. Even with the de-emphasis, if
> your preamp is very wideband you may see that half of your dynamic range is
> being eaten up by rumble that you're just going to filter anyway, so there is
> a need to have considerably more dynamic range than the record itself may
> have usable.

With software, you can always resample internally to a higher resolution,
perform the algorithm, then (if desired) requantize back to 16 bits at
the end.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[PStamler]

On Thursday, October 8, 2015 at 4:50:35 PM UTC-5, Randy Yates wrote:
> (Scott Dorsey) writes:
>
> > PStamler > wrote:
> >>On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
> >>> On 6/10/2015 5:29 AM, PStamler wrote:
> >>> > Record in 24 bits (or 32 bits floating); after you've done everything y=
> >>ou want to do to the
> >>> > signal, convert to 16 bits for burning a CD.
> >>>=20
> >>> Had to smile that you think you will need more than 96dB DNR to record=20
> >>> old vinyl :-)
> >>> (Most DAWS work internally at 32bit floating, or better now anyway)
> >>> Still, won't hurt at least.
> >>
> >>That's a long and complicated discussion, which I don't have time for while=
> >> the rice is on the stove. Suffice it to say that every declicking algorith=
> >>m I've ever used works more effectively, with fewer artifacts, on 24-bit fi=
> >>les (or the 32-bit version of them) than on 16. Worth using just on that ba=
> >>sis.
> >
> > That's weird. I have noticed dramatic improvements in declicking from using
> > higher sampling rates and wideband preamps; even though there isn't much signal
> > at 30 KHz there is enough to make for a better-defined edge detection.
> >
> > But I have not noticed improvements from the longer sample size.
> >
> > Now... I would most definitely believe that the longer sample size would be
> > a huge win if you're doing the RIAA de-emphasis in software, since the needed
> > dynamic range in that case is pretty huge. Even with the de-emphasis, if
> > your preamp is very wideband you may see that half of your dynamic range is
> > being eaten up by rumble that you're just going to filter anyway, so there is
> > a need to have considerably more dynamic range than the record itself may
> > have usable.
>
> With software, you can always resample internally to a higher resolution,
> perform the algorithm, then (if desired) requantize back to 16 bits at
> the end.

But there's no reason not to digitize it at 24 bits in the first place, with storage so cheap now.

Peace,
Paul

[Randy Yates[_2_]]

PStamler > writes:

> On Thursday, October 8, 2015 at 4:50:35 PM UTC-5, Randy Yates wrote:
>> (Scott Dorsey) writes:
>>
>> > PStamler > wrote:
>> >>On Monday, October 5, 2015 at 8:20:00 PM UTC-5, Trevor wrote:
>> >>> On 6/10/2015 5:29 AM, PStamler wrote:
>> >>> > Record in 24 bits (or 32 bits floating); after you've done everything y=
>> >>ou want to do to the
>> >>> > signal, convert to 16 bits for burning a CD.
>> >>>=20
>> >>> Had to smile that you think you will need more than 96dB DNR to record=20
>> >>> old vinyl :-)
>> >>> (Most DAWS work internally at 32bit floating, or better now anyway)
>> >>> Still, won't hurt at least.
>> >>
>> >>That's a long and complicated discussion, which I don't have time for while=
>> >> the rice is on the stove. Suffice it to say that every declicking algorith=
>> >>m I've ever used works more effectively, with fewer artifacts, on 24-bit fi=
>> >>les (or the 32-bit version of them) than on 16. Worth using just on that ba=
>> >>sis.
>> >
>> > That's weird. I have noticed dramatic improvements in declicking from using
>> > higher sampling rates and wideband preamps; even though there isn't much signal
>> > at 30 KHz there is enough to make for a better-defined edge detection.
>> >
>> > But I have not noticed improvements from the longer sample size.
>> >
>> > Now... I would most definitely believe that the longer sample size would be
>> > a huge win if you're doing the RIAA de-emphasis in software, since the needed
>> > dynamic range in that case is pretty huge. Even with the de-emphasis, if
>> > your preamp is very wideband you may see that half of your dynamic range is
>> > being eaten up by rumble that you're just going to filter anyway, so there is
>> > a need to have considerably more dynamic range than the record itself may
>> > have usable.
>>
>> With software, you can always resample internally to a higher resolution,
>> perform the algorithm, then (if desired) requantize back to 16 bits at
>> the end.
>
> But there's no reason not to digitize it at 24 bits in the first
> place, with storage so cheap now.

No argument there. I'm just saying that if you did, that shouldn't cause
a problem for deemphasis software. I mean, I would hope the (plugin?)
implementer(s) would know the dynamic range requirements of the
algorithm and resample/requantize if necessary.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Scott Dorsey]

Randy Yates > wrote:
>>
>> But there's no reason not to digitize it at 24 bits in the first
>> place, with storage so cheap now.
>
>No argument there. I'm just saying that if you did, that shouldn't cause
>a problem for deemphasis software. I mean, I would hope the (plugin?)
>implementer(s) would know the dynamic range requirements of the
>algorithm and resample/requantize if necessary.

I was suggesting that there may be benefit to digitizing at 24 bits when
in fact you have far more noise than signal going into the converter. Do
a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
left if you are lucky.
--scott

>--
>Randy Yates
>Digital Signal Labs
>http://www.digitalsignallabs.com


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

[Randy Yates[_2_]]

(Scott Dorsey) writes:

> Randy Yates > wrote:
>>>
>>> But there's no reason not to digitize it at 24 bits in the first
>>> place, with storage so cheap now.
>>
>>No argument there. I'm just saying that if you did, that shouldn't cause
>>a problem for deemphasis software. I mean, I would hope the (plugin?)
>>implementer(s) would know the dynamic range requirements of the
>>algorithm and resample/requantize if necessary.
>
> I was suggesting that there may be benefit to digitizing at 24 bits when
> in fact you have far more noise than signal going into the converter. Do
> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
> left if you are lucky.

So the situation you're describing is something like this: the analog
signal being input to the ADC has a relatively low analog SNR. Correct?

In that case I really don't see why digitizing at a higher depth buys
you anything. The quantization noise (if the quantization is done
"right") should be way below the input noise.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Trevor]

On 9/10/2015 2:37 PM, Randy Yates wrote:
> (Scott Dorsey) writes:
>> I was suggesting that there may be benefit to digitizing at 24 bits when
>> in fact you have far more noise than signal going into the converter. Do
>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>> left if you are lucky.
>
> So the situation you're describing is something like this: the analog
> signal being input to the ADC has a relatively low analog SNR. Correct?
>
> In that case I really don't see why digitizing at a higher depth buys
> you anything.

It doesn't with any modern DAW software. Nothing does internal
processing at 16 bits any more, and haven't for many years. But if
recording 12 bits of data at a 24bit sample size makes people happy,
then as they say, it won't hurt at least.

Trevor.

[[email protected]]

On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
> On 9/10/2015 2:37 PM, Randy Yates wrote:
> > (Scott Dorsey) writes:
> >> I was suggesting that there may be benefit to digitizing at 24 bits when
> >> in fact you have far more noise than signal going into the converter. Do
> >> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
> >> left if you are lucky.
> >
> > So the situation you're describing is something like this: the analog
> > signal being input to the ADC has a relatively low analog SNR. Correct?
> >
> > In that case I really don't see why digitizing at a higher depth buys
> > you anything.
>
> It doesn't with any modern DAW software. Nothing does internal
> processing at 16 bits any more, and haven't for many years. But if
> recording 12 bits of data at a 24bit sample size makes people happy,
> then as they say, it won't hurt at least.
>
> Trevor.

I think the point that needs to be made explicit here is this....

when digitizing records with clicks that are going to pass through a de-click algorithm, having a wide enough dynamic range and frequency response to fully capture ____the clicks___ (not just the music) can be helpful to the de-click algorithm to recognize the clicks to remove them.

Mark

[Randy Yates[_2_]]

writes:

> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>> > (Scott Dorsey) writes:
>> >> I was suggesting that there may be benefit to digitizing at 24 bits when
>> >> in fact you have far more noise than signal going into the converter. Do
>> >> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>> >> left if you are lucky.
>> >
>> > So the situation you're describing is something like this: the analog
>> > signal being input to the ADC has a relatively low analog SNR. Correct?
>> >
>> > In that case I really don't see why digitizing at a higher depth buys
>> > you anything.
>>
>> It doesn't with any modern DAW software. Nothing does internal
>> processing at 16 bits any more, and haven't for many years. But if
>> recording 12 bits of data at a 24bit sample size makes people happy,
>> then as they say, it won't hurt at least.
>>
>> Trevor.
>
> I think the point that needs to be made explicit here is this....
>
> when digitizing records with clicks that are going to pass through a
> de-click algorithm, having a wide enough dynamic range and frequency
> response to fully capture ____the clicks___ (not just the music) can
> be helpful to the de-click algorithm to recognize the clicks to remove
> them.

Mark,

I agree; that makes sense. And if you have to drop the average music
level down 20 dB (or more) in order to prevent the clicks from
saturating, it's good to have a few more bits in the ADC.

Basically we're talking about a wider dynamic range signal that standard
audio.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Les Cargill[_4_]]

Randy Yates wrote:
> writes:
>
>> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>>> (Scott Dorsey) writes:
>>>>> I was suggesting that there may be benefit to digitizing at 24 bits when
>>>>> in fact you have far more noise than signal going into the converter. Do
>>>>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>>>>> left if you are lucky.
>>>>
>>>> So the situation you're describing is something like this: the analog
>>>> signal being input to the ADC has a relatively low analog SNR. Correct?
>>>>
>>>> In that case I really don't see why digitizing at a higher depth buys
>>>> you anything.
>>>
>>> It doesn't with any modern DAW software. Nothing does internal
>>> processing at 16 bits any more, and haven't for many years. But if
>>> recording 12 bits of data at a 24bit sample size makes people happy,
>>> then as they say, it won't hurt at least.
>>>
>>> Trevor.
>>
>> I think the point that needs to be made explicit here is this....
>>
>> when digitizing records with clicks that are going to pass through a
>> de-click algorithm, having a wide enough dynamic range and frequency
>> response to fully capture ____the clicks___ (not just the music) can
>> be helpful to the de-click algorithm to recognize the clicks to remove
>> them.
>
> Mark,
>
> I agree; that makes sense. And if you have to drop the average music
> level down 20 dB (or more) in order to prevent the clicks from
> saturating, it's good to have a few more bits in the ADC.
>
> Basically we're talking about a wider dynamic range signal that standard
> audio.
>


Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
for legacy gear. Even then...

For example. my old Fostex VF16 would track @24 bit. I don't recall if
it was true 24 bit or not when connected via Lightpipe - I used
16 bit until fairly recently ( when it died ) .

That's turn-of-the-century - 2000ish - technology.

--
Les Cargill

[Randy Yates[_2_]]

Les Cargill > writes:

> Randy Yates wrote:
>> writes:
>>
>>> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>>>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>>>> (Scott Dorsey) writes:
>>>>>> I was suggesting that there may be benefit to digitizing at 24 bits when
>>>>>> in fact you have far more noise than signal going into the converter. Do
>>>>>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>>>>>> left if you are lucky.
>>>>>
>>>>> So the situation you're describing is something like this: the analog
>>>>> signal being input to the ADC has a relatively low analog SNR. Correct?
>>>>>
>>>>> In that case I really don't see why digitizing at a higher depth buys
>>>>> you anything.
>>>>
>>>> It doesn't with any modern DAW software. Nothing does internal
>>>> processing at 16 bits any more, and haven't for many years. But if
>>>> recording 12 bits of data at a 24bit sample size makes people happy,
>>>> then as they say, it won't hurt at least.
>>>>
>>>> Trevor.
>>>
>>> I think the point that needs to be made explicit here is this....
>>>
>>> when digitizing records with clicks that are going to pass through a
>>> de-click algorithm, having a wide enough dynamic range and frequency
>>> response to fully capture ____the clicks___ (not just the music) can
>>> be helpful to the de-click algorithm to recognize the clicks to remove
>>> them.
>>
>> Mark,
>>
>> I agree; that makes sense. And if you have to drop the average music
>> level down 20 dB (or more) in order to prevent the clicks from
>> saturating, it's good to have a few more bits in the ADC.
>>
>> Basically we're talking about a wider dynamic range signal that standard
>> audio.
>>
>
>
> Which standard?

The "law of physics" standard...

I was not referring to any particular piece of equipment or standard,
but rather was viewing things from an analytical perspective.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Randy Yates[_2_]]

Randy Yates > writes:

> Les Cargill > writes:
>
>> Randy Yates wrote:
>>> writes:
>>>
>>>> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>>>>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>>>>> (Scott Dorsey) writes:
>>>>>>> I was suggesting that there may be benefit to digitizing at 24 bits when
>>>>>>> in fact you have far more noise than signal going into the converter. Do
>>>>>>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>>>>>>> left if you are lucky.
>>>>>>
>>>>>> So the situation you're describing is something like this: the analog
>>>>>> signal being input to the ADC has a relatively low analog SNR. Correct?
>>>>>>
>>>>>> In that case I really don't see why digitizing at a higher depth buys
>>>>>> you anything.
>>>>>
>>>>> It doesn't with any modern DAW software. Nothing does internal
>>>>> processing at 16 bits any more, and haven't for many years. But if
>>>>> recording 12 bits of data at a 24bit sample size makes people happy,
>>>>> then as they say, it won't hurt at least.
>>>>>
>>>>> Trevor.
>>>>
>>>> I think the point that needs to be made explicit here is this....
>>>>
>>>> when digitizing records with clicks that are going to pass through a
>>>> de-click algorithm, having a wide enough dynamic range and frequency
>>>> response to fully capture ____the clicks___ (not just the music) can
>>>> be helpful to the de-click algorithm to recognize the clicks to remove
>>>> them.
>>>
>>> Mark,
>>>
>>> I agree; that makes sense. And if you have to drop the average music
>>> level down 20 dB (or more) in order to prevent the clicks from
>>> saturating, it's good to have a few more bits in the ADC.
>>>
>>> Basically we're talking about a wider dynamic range signal that standard
>>> audio.
>>>
>>
>>
>> Which standard?
>
> The "law of physics" standard...
>
> I was not referring to any particular piece of equipment or standard,
> but rather was viewing things from an analytical perspective.

<Ahem..> OK, now I see my statement "...standard audio" MIGHT have had
something to do with your question, Les. Sorry!

Well, you know what I mean, don't you? OK let's say CD standard audio.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com

[Les Cargill[_4_]]

Scott Dorsey wrote:
> Les Cargill > wrote:
>>
>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>> for legacy gear. Even then...
>
> Except that many of those 24 bit converters are sufficiently noisy that
> fewer than 16 bits are actually of any use....
> --scott
>

If the converters are not that noisy, the sources are.

--
Les Cargill

[Les Cargill[_4_]]

Randy Yates wrote:
> Les Cargill > writes:
>
>> Randy Yates wrote:
>>> writes:
>>>
>>>> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>>>>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>>>>> (Scott Dorsey) writes:
>>>>>>> I was suggesting that there may be benefit to digitizing at 24 bits when
>>>>>>> in fact you have far more noise than signal going into the converter. Do
>>>>>>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>>>>>>> left if you are lucky.
>>>>>>
>>>>>> So the situation you're describing is something like this: the analog
>>>>>> signal being input to the ADC has a relatively low analog SNR. Correct?
>>>>>>
>>>>>> In that case I really don't see why digitizing at a higher depth buys
>>>>>> you anything.
>>>>>
>>>>> It doesn't with any modern DAW software. Nothing does internal
>>>>> processing at 16 bits any more, and haven't for many years. But if
>>>>> recording 12 bits of data at a 24bit sample size makes people happy,
>>>>> then as they say, it won't hurt at least.
>>>>>
>>>>> Trevor.
>>>>
>>>> I think the point that needs to be made explicit here is this....
>>>>
>>>> when digitizing records with clicks that are going to pass through a
>>>> de-click algorithm, having a wide enough dynamic range and frequency
>>>> response to fully capture ____the clicks___ (not just the music) can
>>>> be helpful to the de-click algorithm to recognize the clicks to remove
>>>> them.
>>>
>>> Mark,
>>>
>>> I agree; that makes sense. And if you have to drop the average music
>>> level down 20 dB (or more) in order to prevent the clicks from
>>> saturating, it's good to have a few more bits in the ADC.
>>>
>>> Basically we're talking about a wider dynamic range signal that standard
>>> audio.
>>>
>>
>>
>> Which standard?
>
> The "law of physics" standard...
>
> I was not referring to any particular piece of equipment or standard,
> but rather was viewing things from an analytical perspective.
>

The laws of physics are only as good as the people who measure the
experiments which verify them. You cannot get away from the
technology.

--
Les Cargill

[[email protected]]

This thread has become lugubrious. I convert mono
and stereo vinyl to digital as regularly as dropping
toast! In 24bit I just make sure I don't peak above
-6dBfs, high-pass above 30Hz, and peak limit
no more than1-2dB to get the level up a little. I
don't over-analyze it or turn it into Russia V ISIS
in Syria! gosh...!

[Les Cargill[_4_]]

Randy Yates wrote:
> Randy Yates > writes:
>
>> Les Cargill > writes:
>>
>>> Randy Yates wrote:
>>>> writes:
>>>>
>>>>> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>>>>>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>>>>>> (Scott Dorsey) writes:
>>>>>>>> I was suggesting that there may be benefit to digitizing at 24 bits when
>>>>>>>> in fact you have far more noise than signal going into the converter. Do
>>>>>>>> a sharp FIR high-pass at 18 Hz and you still have plenty of bits of signal
>>>>>>>> left if you are lucky.
>>>>>>>
>>>>>>> So the situation you're describing is something like this: the analog
>>>>>>> signal being input to the ADC has a relatively low analog SNR. Correct?
>>>>>>>
>>>>>>> In that case I really don't see why digitizing at a higher depth buys
>>>>>>> you anything.
>>>>>>
>>>>>> It doesn't with any modern DAW software. Nothing does internal
>>>>>> processing at 16 bits any more, and haven't for many years. But if
>>>>>> recording 12 bits of data at a 24bit sample size makes people happy,
>>>>>> then as they say, it won't hurt at least.
>>>>>>
>>>>>> Trevor.
>>>>>
>>>>> I think the point that needs to be made explicit here is this....
>>>>>
>>>>> when digitizing records with clicks that are going to pass through a
>>>>> de-click algorithm, having a wide enough dynamic range and frequency
>>>>> response to fully capture ____the clicks___ (not just the music) can
>>>>> be helpful to the de-click algorithm to recognize the clicks to remove
>>>>> them.
>>>>
>>>> Mark,
>>>>
>>>> I agree; that makes sense. And if you have to drop the average music
>>>> level down 20 dB (or more) in order to prevent the clicks from
>>>> saturating, it's good to have a few more bits in the ADC.
>>>>
>>>> Basically we're talking about a wider dynamic range signal that standard
>>>> audio.
>>>>
>>>
>>>
>>> Which standard?
>>
>> The "law of physics" standard...
>>
>> I was not referring to any particular piece of equipment or standard,
>> but rather was viewing things from an analytical perspective.
>
> <Ahem..> OK, now I see my statement "...standard audio" MIGHT have had
> something to do with your question, Les. Sorry!
>

Not a problem! I was not pointed at anything specific because I
wanted to see where you'd take it.

A "law of physics" standard would be one of those "quietest rooms in
the world" things, which is well above 16 bit. The silence drives
people - literally - crazy. They can't take it for long.

It's funny; on another forum where the capacitance per foot of
cables used for hi-Z interconnects was being ... optimized, I mentioned
that I'd low-passed an example of that instrument at a horrifying 3500Hz
- AM radio frequencies - and it sounded like silverware being dropped on
concrete.

Hell, I dunno - maybe the sound of silverware is where all the dinner is
on that instrument. They airbrush ice cubes, don't they?

> Well, you know what I mean, don't you? OK let's say CD standard audio.

Which isn't standard any more :) CD standard was
set before the paint on digital was dry. The silicon
people standardized on 24 bit whether we liked it or not.

You know the rest of the joke.

--
Les Cargill

[Edi Zubovic]

On Fri, 9 Oct 2015 20:10:05 -0700 (PDT), wrote:

>This thread has become lugubrious. I convert mono
>and stereo vinyl to digital as regularly as dropping
>toast! In 24bit I just make sure I don't peak above
>-6dBfs, high-pass above 30Hz, and peak limit
>no more than1-2dB to get the level up a little. I
>don't over-analyze it or turn it into Russia V ISIS
>in Syria! gosh...!


......and I use 64-bit decimal ie. float - the coprocessor's
business... A tremendeous headroom. It's a part of my algorithm, it
simply has to be 64-bit... mathematics knows more than physics.
Processing time? --It's OK, my I/O is system speed, 13 GB "Ram Drive".
But, if the system fails for whatever reason and reboots, then you
haven't spent some time with your current work, you were in the nearby
coffe bar looking at girls passing by... no backups whatever, that is.

Edi Zubovic, Crikvenica, Croatia

[geoff]

On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
> Les Cargill > wrote:
>>
>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>> for legacy gear. Even then...
>
> Except that many of those 24 bit converters are sufficiently noisy that
> fewer than 16 bits are actually of any use....
> --scott
>


Were there any that bad ?

geoff

[Trevor]

On 9/10/2015 11:23 PM, wrote:
> On Friday, October 9, 2015 at 12:03:16 AM UTC-4, Trevor wrote:
>> On 9/10/2015 2:37 PM, Randy Yates wrote:
>>> (Scott Dorsey) writes:
>>>> I was suggesting that there may be benefit to digitizing at 24
>>>> bits when in fact you have far more noise than signal going
>>>> into the converter. Do a sharp FIR high-pass at 18 Hz and you
>>>> still have plenty of bits of signal left if you are lucky.
>>>
>>> So the situation you're describing is something like this: the
>>> analog signal being input to the ADC has a relatively low analog
>>> SNR. Correct?
>>>
>>> In that case I really don't see why digitizing at a higher depth
>>> buys you anything.
>>
>> It doesn't with any modern DAW software. Nothing does internal
>> processing at 16 bits any more, and haven't for many years. But if
>> recording 12 bits of data at a 24bit sample size makes people
>> happy, then as they say, it won't hurt at least.
>
> I think the point that needs to be made explicit here is this....
>
> when digitizing records with clicks that are going to pass through a
> de-click algorithm, having a wide enough dynamic range and frequency
> response to fully capture ____the clicks___ (not just the music)
> can be helpful to the de-click algorithm to recognize the clicks to
> remove them.
>

I wouldn't bother digitising LP's that badly scratched myself, and such
clicks are FAR better removed by hand anyway. In fact I rarely use any
declicking software, I prefer to do it by hand or not at all.

Trevor.

[Trevor]

On 10/10/2015 4:24 PM, geoff wrote:
> On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
>> Les Cargill > wrote:
>>>
>>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>>> for legacy gear. Even then...
>>
>> Except that many of those 24 bit converters are sufficiently noisy that
>> fewer than 16 bits are actually of any use....
>
>
> Were there any that bad ?

Of course there were, still are. *Many* new motherboards for example can
do 24 bit audio at less than 96dB DNR. Many of the cheap PCI audio cards
too.

Trevor.

[geoff]

On 10/10/2015 7:32 p.m., Trevor wrote:
> On 10/10/2015 4:24 PM, geoff wrote:
>> On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
>>> Les Cargill > wrote:
>>>>
>>>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>>>> for legacy gear. Even then...
>>>
>>> Except that many of those 24 bit converters are sufficiently noisy that
>>> fewer than 16 bits are actually of any use....
>>
>>
>> Were there any that bad ?
>
> Of course there were, still are. *Many* new motherboards for example can
> do 24 bit audio at less than 96dB DNR. Many of the cheap PCI audio cards
> too.
>
> Trevor.

Didn't mean 'better than theoretical 16-bits. Meant 24 bit converters
better than 16-bit converters in real-life (ie more like 12-13 bits worth).

geoff

[Trevor]

On 10/10/2015 5:52 PM, geoff wrote:
> On 10/10/2015 7:32 p.m., Trevor wrote:
>> On 10/10/2015 4:24 PM, geoff wrote:
>>> On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
>>>> Les Cargill > wrote:
>>>>>
>>>>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>>>>> for legacy gear. Even then...
>>>>
>>>> Except that many of those 24 bit converters are sufficiently noisy that
>>>> fewer than 16 bits are actually of any use....
>>>
>>>
>>> Were there any that bad ?
>>
>> Of course there were, still are. *Many* new motherboards for example can
>> do 24 bit audio at less than 96dB DNR. Many of the cheap PCI audio cards
>> too.
>>
>
> Didn't mean 'better than theoretical 16-bits. Meant 24 bit converters
> better than 16-bit converters in real-life (ie more like 12-13 bits worth).

Wrong on both counts then. Many good audio cards and external converters
can do 16 bit audio at a genuine 93dB or better (15.5 bits+) in easily
measurable loop back tests. (One way i.e A-D OR D-A only, even better.)
Which is a lot better than some hardware running at 24 bits as I said.
And the better converters can get close to 20 bits of real data when
running at 24bits now. None will ever get 24 bits in the real world,
although you can simply combine converters and pads to achieve greater
range if necessary. But i can't think why off hand :-)

Trevor.

[geoff]

On 10/10/2015 10:11 p.m., Trevor wrote:
> On 10/10/2015 5:52 PM, geoff wrote:
>> On 10/10/2015 7:32 p.m., Trevor wrote:
>>> On 10/10/2015 4:24 PM, geoff wrote:
>>>> On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
>>>>> Les Cargill > wrote:
>>>>>>
>>>>>> Which standard? A/D-D/A are pretty much all 24 bit now anyway, except
>>>>>> for legacy gear. Even then...
>>>>>
>>>>> Except that many of those 24 bit converters are sufficiently noisy
>>>>> that
>>>>> fewer than 16 bits are actually of any use....
>>>>
>>>>
>>>> Were there any that bad ?
>>>
>>> Of course there were, still are. *Many* new motherboards for example can
>>> do 24 bit audio at less than 96dB DNR. Many of the cheap PCI audio cards
>>> too.
>>>
>>
>> Didn't mean 'better than theoretical 16-bits. Meant 24 bit converters
>> better than 16-bit converters in real-life (ie more like 12-13 bits
>> worth).
>
> Wrong on both counts then. Many good audio cards and external converters
> can do 16 bit audio at a genuine 93dB or better (15.5 bits+) in easily
> measurable loop back tests. (One way i.e A-D OR D-A only, even better.)
> Which is a lot better than some hardware running at 24 bits as I said.
> And the better converters can get close to 20 bits of real data when
> running at 24bits now. None will ever get 24 bits in the real world,
> although you can simply combine converters and pads to achieve greater
> range if necessary. But i can't think why off hand :-)

Again, many (any) 24-bit converters in real-life that are worse than 16
bit converters in real-life ?

Not talking motherboard type ones, talking dedicated converters for
musical purposes.

Maybe the best ever 16 bit converters could have approached theoretical
16 bits performance....(?)

geoff

[Trevor]

On 10/10/2015 9:43 PM, geoff wrote:
> On 10/10/2015 10:11 p.m., Trevor wrote:
>> On 10/10/2015 5:52 PM, geoff wrote:
>>> On 10/10/2015 7:32 p.m., Trevor wrote:
>>>> On 10/10/2015 4:24 PM, geoff wrote:
>>>>> On 10/10/2015 6:25 a.m., Scott Dorsey wrote:
>>>>>> Les Cargill > wrote:
>>>>>>>
>>>>>>> Which standard? A/D-D/A are pretty much all 24 bit now anyway,
>>>>>>> except for legacy gear. Even then...
>>>>>>
>>>>>> Except that many of those 24 bit converters are sufficiently noisy
>>>>>> that fewer than 16 bits are actually of any use....
>>>>>
>>>>>
>>>>> Were there any that bad ?
>>>>
>>>> Of course there were, still are. *Many* new motherboards for example
>>>> can do 24 bit audio at less than 96dB DNR. Many of the cheap PCI audio
>>>> cards too.
>>>>
>>>
>>> Didn't mean 'better than theoretical 16-bits. Meant 24 bit converters
>>> better than 16-bit converters in real-life (ie more like 12-13 bits
>>> worth).
>>
>> Wrong on both counts then. Many good audio cards and external converters
>> can do 16 bit audio at a genuine 93dB or better (15.5 bits+) in easily
>> measurable loop back tests. (One way i.e A-D OR D-A only, even better.)
>> Which is a lot better than some hardware running at 24 bits as I said.
>> And the better converters can get close to 20 bits of real data when
>> running at 24bits now. None will ever get 24 bits in the real world,
>> although you can simply combine converters and pads to achieve greater
>> range if necessary. But i can't think why off hand :-)
>
> Again, many (any) 24-bit converters in real-life that are worse than 16
> bit converters in real-life ?

Not Sure how much clearer I can make it for you... *YES* quite a few!!!!!

> Not talking motherboard type ones, talking dedicated converters for
> musical purposes.

Well that depends on YOUR definition of "dedicated converters for
musical purposes" doesn't it? I'm sure you can make up a definition to
prove my statement wrong if you want! Not playing THAT game! I answered
what was originally written.

> Maybe the best ever 16 bit converters could have approached theoretical
> 16 bits performance....(?)

And not all that uncommon these days in fact!!!!
Cheap IC converter chips do it easily now, but the analog interfacing
needs a little more care than the $5 cards can afford. You don't need to
spend a lot of money to get a real 15+ bits any more however.

Trevor.

[[email protected]]

On Sunday, October 4, 2015 at 10:03:24 PM UTC-4, Nil wrote:

> I want to digitalize a mono LP.

The first thing to do is a really thorough cleaning of the record which can make a major difference in surface noise. As a first layer, look into cleaning with glue. No, not kidding. Look on Youtube for demonstrations. Make sure you use TiteBond II, not regular Titebond which isn't pliable enough. The idea is it cures on the record and when peeled away takes dirt with it.

Second step is to use vacuum irrigation which involves special surfactant and distilled water and purpose-made brushes. I use Disc Doctor brushes and fluid. You can buy a pricey machine to turn the record for scrubbing ad vacuuming or make your own.

[Adrian Tuddenham[_2_]]

Edi Zubovic <edi.zubovic[rem > wrote:

> On Mon, 05 Oct 2015 16:13:47 +0200, Edi Zubovic <edi.zubovic[rem
> > wrote:
>
>
> >
> >The last ones are dangerous to all stereo records as they have no
> vertical impedance ---------compliance-------
> >and usually require a hefty tracking force, so they
> >might damage a stereo record . With mono records, a greater tracking
> >force is no bad thing at all, though.

A combination of high tracking force and low vertical compliance wrecks
the groove at high slew rates because of the pinch effect.


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

[Nil[_2_]]

On 18 Oct 2015, wrote in rec.audio.pro:

> The first thing to do is a really thorough cleaning of the record
> which can make a major difference in surface noise. As a first
> layer, look into cleaning with glue. No, not kidding. Look on
> Youtube for demonstrations. Make sure you use TiteBond II, not
> regular Titebond which isn't pliable enough. The idea is it cures
> on the record and when peeled away takes dirt with it.

Yikes! It would never have occurred to me to clean a record with glue.
I did find several videos on Youtube, and I guess the idea is plausible
- instead of wiping or scraping debris out of the grooves, the dirt
would be lifted straight out with potentially less damage to the
plastic. I'm still a little skeptical, but I may try it on an
expendable record one of these days.

> Second step is to use vacuum irrigation which involves special
> surfactant and distilled water and purpose-made brushes. I use
> Disc Doctor brushes and fluid. You can buy a pricey machine to
> turn the record for scrubbing ad vacuuming or make your own.

This isn't something I do very often, and it's not for commercial
purposes, so I wouldn't spend a lot of money on the process. But I
would like to do the best job I can with the tools I have.

[Scott Dorsey]

Nil > wrote:
>
>Yikes! It would never have occurred to me to clean a record with glue.
>I did find several videos on Youtube, and I guess the idea is plausible
>- instead of wiping or scraping debris out of the grooves, the dirt
>would be lifted straight out with potentially less damage to the
>plastic. I'm still a little skeptical, but I may try it on an
>expendable record one of these days.

It works well. You can get a higher grade PVA material that is specifically
designed for the application and stays flexible longer, but the glue will
work.

This is a thing, though, that you would want to do filthy records that you
got at a yard sale. It's something you do when you are worried about the
record possibly contaminating the cleaning machine.

You can also just do a soap and water cleaning before putting it in the
machine as well; it's not as effective as an ultrasonic machine or the glue
peel, but it's okay and it will prevent your vacuum machine from getting
gunked up.

>> Second step is to use vacuum irrigation which involves special
>> surfactant and distilled water and purpose-made brushes. I use
>> Disc Doctor brushes and fluid. You can buy a pricey machine to
>> turn the record for scrubbing ad vacuuming or make your own.
>
>This isn't something I do very often, and it's not for commercial
>purposes, so I wouldn't spend a lot of money on the process. But I
>would like to do the best job I can with the tools I have.

If you play records, you need a vacuum machine. It is the greatest sound
improvement you can get for the investment. The difference is stunning,
and you can buy an old Nitty Gritty Record Doctor for $100 or so.

The fancy machines are faster and easier to use than the Record Doctor,
but if you only have a few records now and then, the Record Doctor will
clean as well as the fancy machines.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Hank[_4_]]

In article >,
Scott Dorsey > wrote:
>Nil > wrote:
>>
>>Yikes! It would never have occurred to me to clean a record with glue.
>>I did find several videos on Youtube, and I guess the idea is plausible
>>- instead of wiping or scraping debris out of the grooves, the dirt
>>would be lifted straight out with potentially less damage to the
>>plastic. I'm still a little skeptical, but I may try it on an
>>expendable record one of these days.
>
>It works well. You can get a higher grade PVA material that is specifically
>designed for the application and stays flexible longer, but the glue will
>work.
>
>This is a thing, though, that you would want to do filthy records that you
>got at a yard sale. It's something you do when you are worried about the
>record possibly contaminating the cleaning machine.
>
>You can also just do a soap and water cleaning before putting it in the
>machine as well; it's not as effective as an ultrasonic machine or the glue
>peel, but it's okay and it will prevent your vacuum machine from getting
>gunked up.
>
I've cleaned up some pretty grungy LP's with a soap, water, and ammonia
solution. After the surface has soaked for a few minutes, I flush the
solution off with warm tap water and swabbing in the direction of the
grooves with a clean kitchen sponge. Once the surface looks relatively
clean (I've had to repeat the process on occasion), it's time to use the
vacuum washer on the disc. On a few occasions, I've had adhesive crap
on the disc surface which needed a solvent like 3M Adhesive Remover.
ALWAYS test the disk surface in something like the runout area with the
solvent on a Q-tip before using it on the recorded groove areas.

If doing 78's (shellac disks) DON'T use ammonia, as ammonia dissolves
shellac. It's OK with vinyl and styrene.
>
>If you play records, you need a vacuum machine. It is the greatest sound
>improvement you can get for the investment. The difference is stunning,
>and you can buy an old Nitty Gritty Record Doctor for $100 or so.
>
At your suggestion, Scott (actually, insistence), I bought a Nitty
Gritty vacuum machine a few years ago. I'm astounded at the results.
Almost all the LP's I've done with the Nitty Gritty play incredibly
well, considering the abuse some of them have had.

>The fancy machines are faster and easier to use than the Record Doctor,
>but if you only have a few records now and then, the Record Doctor will
>clean as well as the fancy machines.

Actually, I've felt that the base level Nitty Gritty, which you have to
rotate manually, to be superior to the fancier motor-driven ones. Some
of the LP's I've used it on have had more problems on one area than on
others, and a bit of back-and-forth attention ultimately did clean them
up pretty well.

For playback, I have a Yamaha P751 direct drive turntable, equipped with
a Stanton 681 EEE Mk III cartridge and an elliptical stylus. The
cartridge replaces a Bang & Olufsen cartridge that had a conical stylus,
and gives a much cleaner audio stream. Nothing particularly "high-end"
about this setup.

Hank