There has been a good deal of talk about some of the OTR groups around cleaning poor sounding material.  Cleaning, where appropriate, is something that I agree completely with and do in a VERY conservative way when needed.  However, there are significant risks inherent in

doing so that can completely degrade the quality of the original material beyond repair when done poorly.

 

In my experience, and I have made every miserable mistake possible, there is often a desire to completely remove all noise, pop and

crackle from the material in order to make it as clean as possible. This must be avoided.

 

With lower generation material, I recommend avoiding reprocessing with the following exceptions: equalization to restore fidelity to the

proper curve (this takes some research), removal of any 60 Hz hum and turntable rumble where needed, and PERHAPS some very minimal removal of crackle.  Anything greater, will lead to the sound of "thumps" where the crackle and pop has been removed.  Too many folks have

ruined low generation recordings (especially from 78 rpm masters) by trying to wholesale remove the existing surface noise.  Our folks in

the 78 RPM groups have learned to leave this in as it really winds up enhancing the vintage nature of the material.  Why would we want

something from 5 years ago to sound like  a new recording.  The result of aggressive noise reduction to surface noise results generally in

the introduction of serious digital artifact sounds in and around the 7000 Hz range.  These are virtually impossible to remove without

ruining the low generation fidelity of the recording.  This is a common and often fatal mistake.

 

With higher generation recordings, we are dealing with different issues.  These include a loss of fidelity from multiple copies over

the years, as well as the introduction of multi-generational tape hiss.  This problem is really the most difficult to deal with and the

one where most of the problems with cleaning occur.  There are ways to address this that can help.  First of all, it's important to assess,

within the frequency range, where the recorded material cuts off. From there, one can use a graphic equalizer to remove hiss beyond that

point.  Also, before or after this step, it may be possible to use the equalizer to reduce the low end of the sound spectrum (which often

becomes exaggerated with multi generational recordings) as well as to carefully enhance the existing upper range to increase sound clarity.

At that point, I use a filter to reduce any 60 Hz hum (very common) and only then carefully look at noise reduction.  In extreme cases, it

is rare that all of the offending hiss can be removed.  However, in the absence of low generation masters, this may be our best

alternative.

 

 But sometimes such noise can be removed without degradation and introducing artifacts. I do this by looking for a "quiet"

interval in the recording; e.g., pauses between speaking. This interval can be used to make a noise profile. This profile is often accurate

accurate for the entire recording. This profile can be used to subtract the noise from the recording. Sometimes several profiles are needed to be applied piecemeal throughout the according. I have had some excellent results with this method. The longer the "quiet" interval the better the results.

 

But this method isn't perfect. If there are pops or clicks in the quiet interval, they should be avoided, because they can bias the measuring the

background noise floor. You still must listen to the results to determine if the sound was improved. There are instances where I had to reject this method because there was little to no improvement, or I removed too much information and degraded the recording. This is usually caused by the "quiet" interval containing useful information that was buried in the noise and too hard to discern.

 

This is exactly what I do when I use noise reduction.  Great minds think alike.  Often times, In cases such as this, I'll look for a

talk/music/audience noise free section and manually remove any pops and clicks remaining before applying a very conservative de-noising.

However, I still find the results from 78 RPM transcriptions to be generally very poor.  With these, I do my de-noising up-front and the

old fashioned way.  I clean the disc, vacuum it and sometimes wet play it with a light mix of gentle hand soap.  How's that for technology!

 

Last word, be CAREFUL about the software you use.  For PC's the only good and effective software out there is expensive.  I do not

recommend Sonic Foundry/Sony plug-ins to recording programs.  These almost always leave phasing and artifacts.  The program Raygun can be helpful but potentially dangerous for artifacts if not used with great care.  The best programs are those marketed by WAVES.  These should be seen as the standard.  They are expensive.  If you can afford them, use them.  IF not, I recommend avoiding the temptation to do any

reprocessing. 

 

There are enough of us out her that can and will assist with restoring rare material.  All you have to do is ask.  I do this regularly with

material from ET's, cassettes, and open reel tapes.  I've also been able to do some limited re mastering with botched mp3's.

 

Always feel free  to contact me at: daveinoldvicnospam@shaw.ca (remove the nospam).