No announcement yet.

Hacking on some modern (1940s-50s) 78s -- impulse filter behavior?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Hacking on some modern (1940s-50s) 78s -- impulse filter behavior?

    Just wondering - does the impulse filter somehow turn the clicks into white noise? After a declick session the general noise level seems to be higher than with the raw file. Of course, then I run it through the CNF and take a slice off of the hiss. But just wondering.

    FWIW, my workflow usually is: record at 45 (the Dual doesn't do 78), usually with 24bit/48khz; remove RIAA (recording thru a mixer with mag phono input); remove any obvious huge clicks with interpolate; change speed to 78. Always have to do a couple of manual big-click passes first; that scaling slider in DC7 helps sometimes. Assemble into one piece with silences for later CD tracks, normalize to -2 or so and play with it. A pass or 2 of EZ Impulse filter seems to work best if it's the first filter - seems to like as much HF as possible to work with. Most of the records (especially the classical) are in good enough shape that only 1 or 2 passes are needed. Then run CNF using a sample of the lead-in or -out in spectral subtraction mode. At that point it's usually pretty clean, so convert to mono, apply suitable 78 EQ (multi-filter of paragraphics for turnover & rolloff), a little more interpolation of any remaining big clicks in quiet passages, a little 20-band EQ to taste, bandpass for CD (sometimes add a second hi-pass for rumble) and maybe a notch for unusual buzz. Normalize, convert to 16/44.1, and it's ready to burn.

    Since I'm trying to make these sound as original as I can, I usually don't go into the VVA and spice it up - keep "improvement" to a minimum. Have found that late 78s generally don't have much compression for some reason (unlike some older LPs) so the expander takes a rest. I usually try to track all of the sides in one session and edit later so levels don't accidentally change.

    DC7 works so well in this flow that I wonder what 8 has (other than the Big Clicks to reduce the need for manual interpolation) that would help. Maybe something that would correct for those occasional off-center holes? I probably wouldn't use the extra tweaks for adding highs & lows that aren't really there.

    Yes, I know, I could get another TT with 78 and a flat preamp to cut a few steps off. Not enough spare cash. This setup has done a fair job over time on my Dad's old 78 collection in which I'm now getting down to a just a few large classical albums. Would like a good record cleaner though - if I had the cash I'd get one of those first. Most of the 78s are in amazingly good shape considering that they had probably 10-15 years of use and another 60 years of storage in various garages.

    Interesting that the one I'm tracking tonite (12 sides!) doesn't seem to be available as a CD reissue at least based on some Google & Amazon searches - Koussevitsky/BSO doing Tchaikovsky's 5th on RCA Victor. Trying to get a CD together by Mom's Day for herself...

    Last edited by Craig Maier; 05-19-2019, 01:29 PM.

  • #2

    Generally the only impulse filter that can potentially turn clicks into white noise is the Narrow Crackle filter. Otherwise, they may turn a click into noise if noise is what surrounded the click (the interpolator makes that decision.).

    The only thing that I do differently is that I convert to the proper turnover via the Virtual Phono Preamp before de-clicking rather than after it.

    As for DC7 vs DC8, there are many differences; one is the replacement algorithm used for the impulse filters in DC8 which is more sophisticated than the one in DC7. You can try it as a free demo from our site at

    All the Best,

    "Who put orange juice in my orange juice?" - - - William Claude Dukenfield