I think any responsible engineer would take the same approach that you describe - maximum unwanted noise reduction with no perceptable change in the audio quality.  I use CEDAR Cambridge to remove clicks and crackle, (or clickle as they now call it), and I find it generally leaves the audio un-affected.  We have even taken a de-clickled file, reversed its phase and combined it with the original file;  what results is noise only with no program signal.  I have, however, heard the results of recordists who have used CEDAR to such an extreme level that the only virtue of the end-result is the total absence of noise;  the fact that the perfomance sounded like it was occuring under water didn't seem to phase them.

db



>________________________________
> From: Jon Samuels <[log in to unmask]>
>To: [log in to unmask] 
>Sent: Monday, October 1, 2012 9:12:56 PM
>Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue way on 70's reissue LP's
> 
>Tom,
>Different transfer engineers have different philosophies.  Some believe any noise reduction is an unacceptable compromise by definition, others denoise aggressively, in the belief that the buying public can't stand surface noise, and that the engineer can EQ the denoised results to make the final product sound as good as ever.  My approach is to remove as much noise as I can without audibly affecting the signal.  No decent engineer/producer should accept digital artifacts.  That's just a given.
>You and I are on the same page here.  Everything should be done for optimal analog playback first.  When digital deticking is required, I prefer what I call manual deticking (what you call re-writing the waveform), to real time or background deticking or decrackling, though the time commitment can be huge.  When the source material is very noisy or in poor condition and can't be upgraded, I decide the acceptable level of retained noise on a case-by-case basis, erring on the side of leaving a little too much noise, as opposed to a little too much signal loss.  Other engineers can and do differ in their approaches.
>Jon Samuels
>--- On Mon, 10/1/12, Tom Fine <[log in to unmask]> wrote:
>
>From: Tom Fine <[log in to unmask]>
>Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue way on 70's reissue LP's
>To: [log in to unmask]
>Date: Monday, October 1, 2012, 5:24 PM
>
>One plus for modern digital de-ticking software is that is has control of the time domain (ie attack 
>and release parameters) that can be better achieved digitally. I still think re-writing the waveform 
>is the best solution, but I will say that there are some complex ticks, which are essentially two 
>defects in the same place in the groove, that digital solutions can do better than only the 
>steadiest of hands trying to manually re-construct the waveform without the tick in place. To my 
>ears, de-crackling software makes the worst artifacts unless used very judiciously and 
>conservatively. I'd rather hear the crackle than the artifacts. I think there is some science 
>showing that people notice a pumping background noise more than a steady-state noise. The brain 
>quickly ignores steady-state noise and pays attention to the music if there's enough musical 
>information there and the music is compelling enough. Any loud background noise is non-ideal, but we 
>all have encountered the fact that some historical and compelling audio does not exist in a pristine 
>format. We all have varying degrees of tolerance for the imperfections. I would say I have much less 
>tolerance than, say, a fan of operatic cylinders.
>
>-- Tom Fine
>
>----- Original Message ----- 
>From: "Jon Samuels" <[log in to unmask]>
>To: <[log in to unmask]>
>Sent: Monday, October 01, 2012 5:15 PM
>Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue way on 70's reissue LP's
>
>
>Dennis and I are in complete agreement here. The most effective tools in remastering ANYTHING are 
>good ears and good judgment. Sounds much easier than it is, but it's a fact. After that, having a 
>high quality sound source, proper disc preparation and using the proper playback equipment in good 
>working order are key. As Tom wrote, it is far better to have a better sound source than to clean it 
>up after the fact. Think of it as a vaccine, rather than a cure. Far better never to have the 
>disease, than to try and cure it.
>In my experience, ALL de-noising programs or hardware (whether declicking, decracking or removing 
>steady noise) have anywhere from a little to a lot of negative effect on the signal. There is no 
>free lunch. For example, Sonic No Noise has a manual deticking series of algorithms that are very 
>time consuming to use (if removing many ticks), but work quite well. I did a test however, on a 
>extremely ticky 78 side, and removed the ticks manually using these algorithms. I found to my dismay 
>that when I A-B'ed the original to the deticked version, the signal had become somewhat duller, and 
>EQ'ing did not restore the lost "bloom". Having said that, CEDAR Cambridge, Sonic No Noise, iZotope 
>RX2, Weiss DNA-1 etc, can be highly effective in reducing noise, and the negative effects can be 
>either acceptable or compensated for to some degree if the software/hardware are used judiciously.
>Jon Samuels
>
>--- On Mon, 10/1/12, Dennis Rooney <[log in to unmask]> wrote:
>
>From: Dennis Rooney <[log in to unmask]>
>Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue way on 70's reissue LP's
>To: [log in to unmask]
>Date: Monday, October 1, 2012, 4:17 PM
>
>It's clear that there is some divergence of opinion on the efficacy of some
>of the tools mentioned. All such tools, analogue or digital, old or new,
>require for their proper use what Tom Fine decribed as "expert ears and
>good taste", supplemented by knowledge, judgment and experience. There is
>NO substitute for those qualities when doing anything with audio.
>Preparation must include an equally expert cleaning before the stylus is
>lowered into the groove. I prefer the Keith Monks for reasons too obvious
>to enumerate. Yet even after such a careful cleaning, there is no guarantee
>of an optimum playback from start to finish. Following a successful
>transfer, every noise event must be checked to ensure against artifacts,
>which is why I see red when some fatuous fool talks about "automatic"
>processes for noise reduction. They don't exist.
>
>DDR
>
>On Mon, Oct 1, 2012 at 3:18 PM, Tom Fine <[log in to unmask]>wrote:
>
>> Hi Dennis:
>>
>> YES, Absolutely! This was discussed regarding the Fletcher Henderson "A
>> Study In Frustration" set. I borrowed the 90's CD set from the library and
>> I must say that the sound quality was an improvement over the LP and some
>> songs definitely are better than the current Mosaic Coleman Hawkins set,
>> which seems to have not used metal parts where Legacy/CBS did just a few
>> years earlier.
>>
>> There are some very good de-crackle tools nowadays but I still think they
>> produce very audible artifacts unless used very conservatively with expert
>> ears and good taste. For the kind of ticks and pops that used to be
>> razor-edited, I still find the best method is to manually write them out of
>> the waveform with Soundforge's waveform editor. Want to do less of this?
>> Start with a really good-quality source and clean it carefully. That said,
>> I'll be the first to admit that some of the very recent de-crackle software
>> is WAY better than 10 years ago or even a handful of years ago.
>>
>> Washboards should be played with a jug band, not used to wash clothes!
>>
>> -- Tom Fine
>>
>> ----- Original Message ----- From: "Dennis Rooney" <
>> [log in to unmask]>
>> To: <[log in to unmask]>
>> Sent: Monday, October 01, 2012 2:50 PM
>> Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue
>> way on 70's reissue LP's
>>
>>
>> I hope this thread does not suggest any nostalgia for the technique, which
>>> was *faute de mieux* at best and invariably produced audible artifacts.
>>> Vinyl noise may have masked some of them but masters manually de-clicked
>>> proved useless for digital reissue. Digital noise reduction strategies now
>>> make the tiresome task of removing bits of tape and splicing back together
>>> seem as quaintly outmoded as the washboard and the butter churn.
>>>
>>> DDR
>>>
>>> On Mon, Oct 1, 2012 at 1:38 PM, Tom Fine <[log in to unmask]>**
>>> wrote:
>>>
>>> This was one way it was done. The other way was using excellent condition
>>>> metal parts, more of which seemed to exist back in the day. With modern
>>>> methods, when you buy something like a Mosaic box set -- which has some
>>>> metal parts and also uses mass media shellacs -- you can really hear why
>>>> it's such a good idea to track down the very best source material
>>>> possible
>>>> and not assume digital or analog "repair" can save the sound. A modern
>>>> transfer of a metal part that's in good condition can sound shockingly
>>>> realistic, very close to what the microphone was hearing except for a
>>>> sharp
>>>> rolloff of the top end. When you have to use commercially-issued shellacs
>>>> -- even those that weren't tortured with steel needles at many grams of
>>>> tracking force -- you lose so much quality because of loud surface noise,
>>>> ticks and pops and the like. Modern digital remedies, when used
>>>> conservatively, can help with this. The old-school guys had a different
>>>> idea of EQ vs. modern transfer guys. I think the older generation
>>>> concentrated more on the bass up into the midrange and figured there was
>>>> little to no high treble to mess with, so they'd roll it off to cut the
>>>> surface noise. Modern engineers, the really good ones, seem to use
>>>> spectrum-analysis tools to figure out where there's more music and where
>>>> there's more noise and then EQ accordingly. The results can sound "nasal"
>>>> to some ears (including my own), and I think some modern reissues aren't
>>>> putting in all the bass that exists on the disks, but it's good not to
>>>> have
>>>> microsecond lapses from razor edits and too-sharp cutoff of treble
>>>> information. That said, there are older reissues where metal parts
>>>> existed
>>>> and were in good shape that smoke modern attempts at the same material
>>>> without access to metal parts (lost, stolen, destroyed, etc).
>>>>
>>>> -- Tom Fine
>>>>
>>>> ----- Original Message ----- From: "James Roth" <[log in to unmask]>
>>>> To: <[log in to unmask]>
>>>> Sent: Monday, October 01, 2012 1:13 PM
>>>> Subject: Re: [ARSCLIST] Reducing crackle from 78 rpm records the analogue
>>>> way on 70's reissue LP's
>>>>
>>>>
>>>> Hello all.
>>>>
>>>> In the 70's I work at "Club 99" [operatic reissue) records.
>>>> I would use a razor blade and a splicing block to "de-click" the tape
>>>> transfers of some very scratchy 78s.
>>>> It would take about 3 hours to de-click a 4-minute aria - verrrrry
>>>> tedious.
>>>> Then, we would take the tape to our sound engineer in Manhattan, David
>>>> Hancock (RIP) and he would so some expert EQ.
>>>>
>>>> Regards,
>>>> Ben Roth
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: Association for Recorded Sound Discussion List [mailto:
>>>> [log in to unmask]**GOV <[log in to unmask]>] On Behalf Of
>>>> Jan
>>>> Myren
>>>> Sent: Monday, October 01, 2012 12:22 PM
>>>> To: [log in to unmask]
>>>> Subject: [ARSCLIST] Reducing crackle from 78 rpm records the analogue way
>>>> on 70's reissue LP's
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> HI!
>>>>
>>>>
>>>>
>>>> I spent the week-end going through my collection of 70's and 80's
>>>> compilation albums of old jazz records remastered by John R.T. Davies and
>>>> Robert Parker.
>>>>
>>>>
>>>>
>>>> I must say I am a bit impressed by the way they were able to suppress
>>>> noise and especially the crackle form the old 78's they used.
>>>>
>>>>
>>>>
>>>> May anybody remember (or know) how they did that and what kind of
>>>> machines
>>>> they used back in the analogue days?
>>>>
>>>>
>>>>
>>>> All the best
>>>>
>>>> Jan
>>>>
>>>>
>>>
>>>
>>> --
>>> Dennis D. Rooney
>>> 303 W. 66th Street, 9HE
>>> New York, NY 10023
>>> 212.874.9626
>>>
>>>
>
>
>-- 
>Dennis D. Rooney
>303 W. 66th Street, 9HE
>New York, NY 10023
>212.874.9626
>
>
>