Correction of Acoustic Recorded Music

[Marc Hildebrant]

Group,

I have been doing some work on fixing acoustic recorded music to add back in the lost low frequencies. If interested, check out the online IEEE Spectrum Magazine section "Geek Life" at : http://spectrum.ieee.org/geek-life/hist ... recordings

Marc

[Henry]

Very noticeable reduction of surface noise and opening up of the acoustic space, to be sure. But to my ear the tuba is no more audible in the second, "clean" version than in the original, which is to say, not very in either case. In fact, the tuba in the original version may have slightly more presence. Again, this is a purely subjective impression. What the second version doesn't do is extend the frequency range downward, below that of the original. And I'd be interested in how the term "subharmonic" is defined! Harmonics are overtones of the fundamental; there are no such things as undertones.

There's a common misconception that acoustic recordings cut off bass frequency response. Yet trombonist Arthur Pryor's Victor recording of the "Blue Bells of Scotland" variations descends to a clearly audible pedal F (contra F), two-and-a-half octaves below middle C. That is, 87.3074 Hz in equal temperament (a' = 440 Hz). For reference, the lowest note on the piano, sub-contra A, = 55 Hz.

[Marc Hildebrant]

Henry,

Thanks for your comments. Couple of points :

1. The files on the IEEE web site are MP3 so that some compression and removal of frequencies will occur by the nature of the MP3 process.

2. The rumble on the recordings is pretty bad. In fact, I start by removing all frequencies below 150 Hz and then apply a software tool from Diamond Cut Productions that creates frequencies that are 1\2 that of the original for frequencies below about 250 Hz. The idea is that the frequencies that the recording process removed due to the mechanical\acoustic process are "put back in".

3. Spectrum Analyzer plots show significantly new frequencies in the "corrected" version.

4. If you visit the Diamond Cut Production Web Site and look for application note 10, you can read much more.

5. I do not work for Diamond Cut Productions and I did the work on my own.

Marc

[VintageTechnologies]

The "before" recording sounds really awful. With the right cartridge and stylus, you should have better S/N ratio and fidelity to start with. That being said, the end result sounds very clean without typical distortions and artifacts heard from overprocessed recordings. I'm not sure the bass has been extended much, but overall the sound is clean and clear.

[yankmycrank]

I'd be fascinated to hear how acoustically recorded orchestral or chamber music sounds with this impressive restoration approach.

[Marco Gilardetti]

Subharmonics are indeed a common concept in sound processing. The good old DBX Subharmonic Synthesizer is the first "tool" that comes to my mind.

I agree that the most impressive achievement of this work is the noise suppression. In most cases, the crackling noise and hiss are replaced with digital artifacts that, at least in my opinion, are much more annoying than the crackle & hiss themselves. This is not the case.

Coming to the bass, I really can't decide if what you achieved here is indeed better than a simple reshape of the bass frequency response performed with a wisely set parametric equalizer.

However, I also agree that the original reading of the record can be improved to some degree.

[Marc Hildebrant]

Group,

The "before" recording is just as it came off the record. I used a Stanton Pickup with a 3.7 mil stylus from I believe Expert Stylus in England. No signal processing occurred except to correct for my 78 RPM speed verses the 80 RPM that Edison used.

I you are interested, a wrote a report on the techniques I used that is now on the Diamond Cut Productions web site under Application note 10.

I have about 200 songs that I have processed so far.

Thanks for the comments !

Marc

[Wolfe]

The "before" recording sounds really awful. With the right cartridge and stylus, you should have better S/N ratio and fidelity to start with. That being said, the end result sounds very clean without typical distortions and artifacts heard from overprocessed recordings. I'm not sure the bass has been extended much, but overall the sound is clean and clear.

The original also wasn't summed to mono. So, giving the impression of more noise.

The restoration sounded very decent, there does seem to be some artifacts but it could be the MP3 compression. I wonder what one of the good, late acoustical Okeh's might sound like with this process.

[Rastus10]

Sounds great to me!

I don't have the technical knowledge, but, unlike others, am not going to go out of my way to find fault. That original recording sounds badly, really?? Someone is on this forum saying that when we probably ALL have at least one record that is beat to crap, but is the only copy of something that we have found or perhaps ever will find?? Someone actually wrote that on here?? Tough crowd.

If it sounds closer to what was actually occurring in the studio, which of course will never be achieved perfectly for some 3-5 moment of time so long passed, and it doesn't sound like the artists were underwater socializing with octopi, I'm all for it.

[Phono-Phan]

Marc has done several of my records and they all sound GREAT!!!
Thanks Marc!