[Dixielandjazz] Transfering 78s >> T'aint Simple But A Fun Challenge

[BudTuba at aol.com]

In a message dated 4/7/2006 10:32:00 P.M. Eastern Standard Time,  
anton.crouch at optusnet.com.au writes:

The  procedure is to record the 78 disc to the computer at 45 rpm and  then
correct the pitch using an audio editor. I don't know if CoolEdit  and/or
Audacity have this capacity but some editors certainly do. The  process
assumes that the signal is coming to the computer pre-amplified and  
equalised.

If you are (dare I use the word?) an audiophile, you may  also want to
correct the equalisation error introduced by the 45 to 78  process.




Even IF you have a turntable with 78 RPM speed, recording the old  78's at 45 
RPM is has advantages because the dynamics of turntable  buffeting the 
tonehead back and forth and the intensities of clicks increases  with speed.  The 
frequencies of noise due to scratches and worn grooves is  also relatively 
constant regardless of speed so when the frequency is restored  back up by 
converting back by a 78/45 ratio, many of those sounds  disappear in the upper end and 
can be further reduced by equalization  filters.  (Anyone wanting to know the 
exact steps I take with Cool Edit can  contact me offline for instructions 
and samples.) 
 
One might argue that playing the 78's at 45 RPM makes the low end suffer,  
but I have not found that to be the case as long as you have decent cartridge on 
 the tonearm.  The 78's were limited at the low end to about 150  
cycles/second and that becomes (45/78)*150 = 86 cps which is well above the  response 
range of hi-fi cartridges.  
 
Another factor in transferring 78's is how to best remove scratches and  
digs.  I have found that recording in stereo is anther prudent choice  because 
even though 78's were monophonic, some defects will affect one side  of the 
groove more than the other.  Therefore in some cases, just copying  the clearer 
side of the stereo tracks to both tracks can make garbled passages  sound better. 
 Cool Edit has a good scratch filter, but the time to apply  it can be fairly 
extensive.  Consider that a 3 minute recording becomes  5.20 minutes long as 
recorded at 45 RPM, the removal of scratches and digs  (which is done 
mathematically by the program inchworming only and looking at the  waveform), can take 
several minutes longer than the 5 minutes depending on how  high you set the 
discrimination level.  Of course, this depends on the  speed at which your 
computer can operate, but as a rule of thumb, if you record  at 44,100 cps and 
your computer operates at 500 mHz and 10 instructions of code  are required by 
the computer to assess one sample of the groove and more  instructions to 
remove one scratch once encountered (my guess) then your  computer is inchworming 
along the groove at 50 mHz or approximately at same rate  that your recording 
was laid down. (50,000 cps ~ 44,100 cps).  A 78  with a lot of noise takes 
longer than one that is quiet. 
 
Consequently, I looked around for other programs for removing  scratches and 
clicks and found Ray Gun Pro.  This works on a different  engineering 
principle called Fourier transform applicationand finds the defects  faster than 
inchworming.  Again, one needs to experiment to determine the  best settings, but I 
have found that by setting the scratch removal part of Ray  Gun at 50 percent 
maximum, that the proof sound of the desired recording sounds  minimally 
changed by the Ray Gun algorithms.  This is conveniently done by  pressing the 
PROOF button and listening back and forth with and without the  algorithm 
applied.  Used in this fashion, Ray Gun leaves a few scratches  that still need to be 
removed by Cool Edit, but many of those are visually  apparent as thin spikes 
when you look at the waveform and can be removed  manually or again 
automatically by Cool Edit, this time working on a cleaner  recording.
 
Finally, my take on EQUALIZATION:  I tend to use equalization  with the goal 
of making the recording as pleasing to me as possible rather  than a strict 
adherence to restoration.  Especially in the early days of  recording, people 
expected a rather thin sound from their records made and  played acoustically.  
So I suppose that trying to create a resultant that  duplicates that total 
aural response may be technically correct.  However,  there is sonic signal  that 
can be differentially amplified by  equalization curves to sound more like 
the sound in the recording room must  have been.  Of course, one man's pleasure 
can be another's poison, but I  find that raising the low frequencies from 100 
cps to about 400 and then  flattening off the response to about 10000 cps 
will enhance the final product  for me then rapidly lowering the response over 
10000 cps.  This also takes  experimentation and even asking unbiased listeners 
to pick which resultant they  like best (if you are attempting to produce CDs 
for further distribution to  others).  I suspect Dick Broadie spent many an 
hour tweaking knobs for the  excellent restorations he has made with reliving 
the experience of listening to  the early bands had they been recorded with 
modern equipment.  

 
Roy (Bud) Taylor
Smugtown Stompers Jazz Band
'we ain't just  whistlin' dixie!"
585-415-3985  Cell