[78-L] Reducing hiss - a new approach

[DAVID BURNHAM]

That got me thinking. If you have two 78s of the same recording the 
music is the same, but the noise is different, and likely different all 
the way through.

So has someone invented a computer program that compares the two 
recordings and just saves the sound that is COMMON to both? i.e. the 
music?

Let me have a share in the new patent, please :o))


-- 
John Wright

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Well here's another idea which has been rattling around in my brain for even longer than the scanning one.

First of all, I tried your idea of two 78s of the same recording, and it's a good idea - because the only thing in common with the two records would be the music - the surface noise would have no relationship at all.  The problem with this, (as I found out), is that every disc record has some degree of eccentricity - the record isn't pressed precisely in the centre of the disc - although it can be close enough not to cause audible "wow" and it can appear to be centred.  As wow occurrs, the pitch of the music will very slightly raise and lower as the wow moves the stylus away from and towards the spindle respectively.  When you allign the two playbacks, the various wow degrees will cause an effect known as "flanging".  Perhaps a turntable, (or, I guess, two turntables), could be designed to allow fine adjustments in centricity while the records are playing so that the two could be audibly alligned.  But I've never seen such a machine.

Here is my solution using the principle you propose.

A coil offers impedance to alternating current proportional to the frequency.  So picture this:  take an iron core and wind a fine wire around it in a clockwise direction until the core is covered with the wire.  Now take a second wire and wind it counter clockwise around the same core until it covers the first wire.  Then take the first wire and wind it clockwise back down the core until the second wire is covered then take the second wire and cover the new winding of the first wire etc. etc.  Eventually you're going to have two coils which offer a high resistance to audio frequencies.  Now take your 78, play it with a stereo cartridge and feed each channel through one of the windings.  The common signal, (the music), in each channel will reduce significantly the impedance offered by the coil to the other channel, while the noise, which is random, will receive no help from the other channel and will meet high resistance by the coil.  Join the
 output ends of the coil together and you will have the music clearly with the noise significantly reduced.  Theoretically, the only elements of noise which would pass through this circuit unimpeded would be those caused by a pit on one side of the groove matched with a bump on the other side of the groove of exactly the same dimensions.  I actually built a crude version of this by taking two output transformers, feeding a channel of the stereo played 78 to each primary and connecting the secondaries together out of phase.  It wasn't perfect but it did reduce the noise perceptably.

I've been told that integrated circuitry can duplicate the behaviour of any electronic component but I've never seen one which can duplicate this.  

If you work on digital work stations and you are familiar with Waves plug-ins, they recently introduced one called "Centre".  This is used to raise or lower the centre of a stereo sound stage without affecting the left/right channels.  It just has two controls - one for centre amplitude and one for left/right amplitude.  Reducing the sides amplitude does a great job of reducing surface noise on a 78 played in stereo.

db