Acoustic Recordings-FR Correction?

[John]

I've recently inherited a large collection of 78s- half are pop/jazz and half are classical/opera, ranging from 1906 to 1952-ish (mostly 20s-40s). I'm mostly just ripping these on an el-cheapo record player designed to look like an old suitcase phonograph that I fitted with a 78 needle. I use audacity to flip the RIAA, declick, correct speed (where necessary), and sum to mono, then I use Audition to remove noise, correct the EQ, and make any other adjustments. that might be needed. All that jazz.

When filtering the acoustics, I find that they sound much better if you cut the mid-range (~1kHz) by a decibel or so. It's not like magic, but voices and instruments tend to sound much more natural (to me, anyway). Since this is probably since they used a horn to record the sound waves, could one theoretically map out the frequency response of a recording horn and create an EQ/FFT curve that could be inverted to correct acoustic recordings? That is, assuming that at least one of them has survived and the owner(s) are willing and able to do such a test.

I know it wouldn't turn the tinny acoustic 78s into gorgeous full-range digital quality recordings (or even up to par with later electric 78s), but it would at least be interesting to hear Caruso, Jones, and others without the coloring that we hear so often on their records.

[Wolfe]

Recording horns of many different sizes and lengths were used. One surviving example wouldn't provide a very valid model.

Thomas Stockham attempted to remove the colorations in acoustic records using computer technology back in the 1970's, most famously applied the recordings of Caruso. Seems many people don't like them though.

You also have to consider that the "frequency response" / sound quality attained on the acoustic record will depend on the recording diaphragm used and other factors.

[John]

Ah, I guess that makes sense. Using the same horn for a "comic" duet and a symphony wouldn't work all that well

I find it rather interesting that Stockham could do all of that in the '70s. I thought digital-audio (let alone workable and editable digtal audio) didn't come around 'til the mid-late '80s at the earliest. Got a link to an example?

[De Soto Frank]

When transcribing historic ( pre - LP / Tape-master ) recordings, I'm a big fan of minimal processing.

Just easing-out some of the surface "hiss" and perhaps clicks / pops, but generally I prefer to hear the record pretty-much the way it sounded to the original listeners.



Maybe this is okay with me, since I grew-up listening to a Victrola VV-IX ( I wasn't allowed to touch the delicate stereo and its precious LP's ), so, the "tinny acoustic" does not seem that foreign to my ears.


I can't comment much on equipment or software to achieve this, but I know what my ear likes to hear...

Even with modern recording equipment, I imagine that Billy Murray would still sound like a bright, high, ("tinny" ?) tenor...

I remember my older brother in the early 1980's talking about the efforts of folks like Stockham, to clean-up historic (acoustical) recording of folks like Caruso, to try to reveal what they really would have sounded like; those early efforts were REALLY labor-intensive !

[Wolfe]

I find it rather interesting that Stockham could do all of that in the '70s. I thought digital-audio (let alone workable and editable digtal audio) didn't come around 'til the mid-late '80s at the earliest. Got a link to an example?

http://youtu.be/uBRyj-1h9qY

Stockham is supposed to have used computers to process the signal from the 78. The end result was still copied to analog tape.

[John]

Given that digital audio processing has made significant leaps and bounds in 35 years, it certainly wouldn't hurt to try again.

One thing that I have noticed is that on acoustic records, louder signals (a full orchestra, a soprano hitting a high note, etc.) can reach surprising levels, while softer ones get buried in noise. Here's a spectral display of an Adelina Patti record, ca. 1906 on the HMV label. For the most part, it stays between 200-2000Hz, which is probably pretty normal for such an old record. However, when she hits the high note at 2:07-ish, there's stuff up to 8kHz, which is good even for an electric 78. How the reproducers of the time could've stayed in the grooves without destroying the groove is a mystery. This would probably throw a monkey wrench into the equation

http://i.imgur.com/58Tpv89.png

(side note: I used about 7 dB's of noise reduction, a high-pass filter at 125Hz, and a -6dB/octave cutoff filter at 2121Hz to clean up the surface hiss and rumble)

http://youtu.be/uBRyj-1h9qY

Stockham is supposed to have used computers to process the signal from the 78. The end result was still copied to analog tape.

Doesn't sound really that different from another copy I found on youtube. I was expecting something something very different-sounding.

[Wolfe]

The reproducers of the time were certainly heavy enough to track the grooves, and they did wear the high frequency content (what of it there was) from the grooves, which is why when seeking the best sound, one should use mint copies of the records. Any high frequencies were the first to go when the records started getting played a lot.

And if you find a record that was recorded with a lot of volume (or a close perspective to the recording horn) the loudest notes can be stripped / greyed out as is visible on the record surface with resulting blasting / distortion. All from the action of a heavy reproducer (or reproducer, traveler arm and horn as on the early talking machines) and steel needle doing their dastardly work.

[Wolfe]

Doesn't sound really that different from another copy I found on youtube. I was expecting something something very different-sounding.

Stockham Caruso transfers are all over YouTube. After pretty much the entire Caruso Victor discography was treated to Stockham processing, it's what RCA / BMG have been issuing on vinyl and CD for years. They have a very characteristic sound. One that's very different to any other transfer I've ever heard.

[John]

Just did a comparison between the one you linked and another one on youtube. The "Stockham" one (Red) has somewhat more high-frequency (3kHz+) content than the "normal" version (from the original 78), which probably means both are from original Victor 78s and the two uploaders used different equipment/filters. I see plenty of the other "remastered" ones (the redubbed orchestra ones, because that's what people totally want to hear ), but I can't seem to find any of the Stockham ones

[Wolfe]

Another Stockham.

http://youtu.be/bzQDbl2sUqs