Distortion

[Alfred001]

As I mentioned in another thread I'm writing a paper for school about digitalization of discs and I know very little about it. My main resource is a few years old so I worry that some bits might be out of date so I'd like check with you folks.

In my source it states that harmonic and intermodulation distortion are "faults which never seem to be reverse-engineered electornicaly" and therefore must be dealt with at the source, but that the author thinks some of the problems could be dealt with easily and in the future algorithms could emerge.

What is the situation currently? I'm most curious of course about what is available to professionals working in sound archives, if there is any difference between what's available to them and amateurs.

Could anyone point me to some online source on this topic?

[mmarston]

I'm still reading your reference from the other post... quite fascinating.

[mmarston]

My gut feel is that things other than minor distortion are more significant:

"The next thing we shall discover is that we cannot know, much less reverse-engineer, the exact compression-ratio of the particular limiter. It is rather like balancing a razor-blade on its edge. Even if we knew the threshold, the attack-time, the decay-time, and the actual characteristic, any slight mismatch will have a catastrophic effect - usually the gain of the expander will increase dramatically and blow up your loudspeaker. In any case, you can see from curve 10.2(C) that there may be ambiguities. If we were trying to simulate this particular limiter and we had an output 4dB above the threshold, we would not know if this corresponded to an input of +5dB or +20dB. Usually we have to live with the fact that a limiter destroys the very information needed to expand the signal again. In this respect, it differs from a reciprocal noise reduction system, where the compressed sound is deliberately designed to be expandable.

How then do we undo a limiter at all? Here is where we make use of the other clues I listed earlier - any steady background-noise, the harmonics of musical instruments, etc., - routing the signal through an expander circuit..."

This is a really good reference and I cannot thank you enough for posting it.

Best,
Mike

[Hanuman]

What is the situation currently? I'm most curious of course about what is available to professionals working in sound archives, if there is any difference between what's available to them and amateurs.

There might be certain "black box" restoration systems that archivists use which you're not likely to see at home. That said, it's more and more evident that very advanced media software is available to virtually anyone. It's usually the case that the expertise, experience, creativity and sensitivity of the engineer/technician/designer/artist is more important than the particular tools in use.

The idea of retrieving the original media from a processed version is a case that confronts media professionals constantly. The caveat spelt out in your quote is a faily typical disclaimer. I recall Quantel describing encoded video (PAL or NTSC) as a "milkshake": you can't get the milk and chocolate seperated back out successfully no matter what. More recently I came across an "omolete" analogy, also in relation to video. It was pointed out in an article about retrieving original colour from old monochrome "Doctor Who" kine-recordings that the process was akin to making an egg out of an omolete.

I doubt that the technology has dramatically moved on from the time of your paper.

[Alfred001]

My gut feel is that things other than minor distortion are more significant:

"The next thing we shall discover is that we cannot know, much less reverse-engineer, the exact compression-ratio of the particular limiter. It is rather like balancing a razor-blade on its edge. Even if we knew the threshold, the attack-time, the decay-time, and the actual characteristic, any slight mismatch will have a catastrophic effect - usually the gain of the expander will increase dramatically and blow up your loudspeaker. In any case, you can see from curve 10.2(C) that there may be ambiguities. If we were trying to simulate this particular limiter and we had an output 4dB above the threshold, we would not know if this corresponded to an input of +5dB or +20dB. Usually we have to live with the fact that a limiter destroys the very information needed to expand the signal again. In this respect, it differs from a reciprocal noise reduction system, where the compressed sound is deliberately designed to be expandable.

How then do we undo a limiter at all? Here is where we make use of the other clues I listed earlier - any steady background-noise, the harmonics of musical instruments, etc., - routing the signal through an expander circuit..."

This is a really good reference and I cannot thank you enough for posting it.

Best,
Mike

I'm glad you're finding it useful.

What is the situation currently? I'm most curious of course about what is available to professionals working in sound archives, if there is any difference between what's available to them and amateurs.

There might be certain "black box" restoration systems that archivists use which you're not likely to see at home. That said, it's more and more evident that very advanced media software is available to virtually anyone. It's usually the case that the expertise, experience, creativity and sensitivity of the engineer/technician/designer/artist is more important than the particular tools in use.

Would you say that it is still true that harmonic and intermodulation distortions are mostly dealt with at the source rather than through processing?

[Hanuman]

Would you say that it is still true that harmonic and intermodulation distortions are mostly dealt with at the source rather than through processing?

I do digital video and cinema, not audio, by way of disclosure but I be surprised if the statement above were not accurate, even now. The problem is that once you're beyond simple, obvious corrections - I'll call them "one-dimensional", probably ill-advisedly - such as frequency and level, you're into guesswork. More complicated stuff like phase-shifts, harmonic & IM distortion, compression & limiting are all much more non-linear and inter-dependent - much more multi-dimensional. Automatic correction in, say digital signal propagation and recording, analogue television or analogue tape noise-reduction depends on pre-inserted reference signals or codes that can assist a verification process down the line and you rely on perfect alignment and calibration. After the fact, none of this is available and so, even the best specified and implemented algorithm can only second guess the original signal. So the "get it right at the source" mantra is unlikely to be diluted any time soon, no matter how good post-processing gets.

I've always lived by the creed that nothing good happens to a signal from the moment it leaves the originating device. Modern digital correction lets us cheat this to some extent and there's no doubt that processing can enhance the subjective appeal of sound and images. The basic premise still holds, nevertheless.

[mmarston]

Would you say that it is still true that harmonic and intermodulation distortions are mostly dealt with at the source rather than through processing?

The problem is, as they say, "non-trivial." It is possible to measure and duplicate the transfer characteristic of an active device if one has access to it, but even so, it is not possible to determine exactly "where" on that curve a signal may have passed through it. Harmonic distortion will occur at predictable frequencies, but those are also the overtones of musical instruments. IM distortion generates sum and different frequencies, but the number of those increases geometrically with the number of frequencies present in the original signal. Not to mention that 3rd order products increase as the cube of the original signals' powers, fifth order to the fifth power, etc. For years there have been programs that calculate IM products for radio transmitting sites, but the calculations become unwieldy for a large number of transmitters. And these programs don't even attempt to consider non-periodic signals (irrelevant to the purpose.)

Not to mention that any commercial recording may have passed through dozens if not hundreds of active devices, at various unknown points on their individual transfer characteristics, which may (likely will) change dynamically even within a single track due to compression, gain riding, EQ changes, etc. So it would be impossible to accurately determine the individual contributions of all these things.

Which leaves us at basically trying to guess. Again, non-trivial...

I think for the archivist, the best one can do is to minimize the sources if distortion over which one has control, and perhaps make some adjustments by ear to what Copeland calls the "service copy." EQ and filtering choices would seem the most amenable to this subjective manual function.

Fortunately for me, I am not trying to make archival copies for the Smithsonian, but only my own personal enjoyment, so don't have to lose sleep over their lack of perfection...

Good luck with your paper!

Best,
Mike

[Dimal]

Great and informative response Mike....

Thanks mate

Mal.

[Alfred001]

Thanks for the explanations guys!

[flavio81]

Very interesting thread. Sadly i don't have any thing to contribute, except that if you could "de-intermodulate" a record, it would be a major breakthrough.