Quote:
Looking for an online version(one that doesn't require a subscription) of that write up you did about "recorded vs the real thing"... Seems like you did two articles(in two different issues of Stereophile)

I wrote about this subject in February 2010 - http://www.stereophile.com/asweseeit/the_recording_angel/ - and November 1995 - http://www.stereophile.com/asweseeit/75/.

John Atkinson
Editor, Stereophile

Rereading JA's 1995 essay I'm struck with how it fits in with some processing I now do to really overcooked recordings.

When I make mix CD's of pop music, I normalize the overall volume to an average level of about -16dbs. This almost always means that I am reducing the levels of the track by 3-8db overall, depending on how badly the original track was mastered.

One day when I was loading and normalizing tracks for a CD I was making, I was still not satisfied with the sound of the really overcooked tracks. I tried to think of ways I could restore the dynamics. By ear, it was apparent that the tracks had no real, low bass and that the treble tended to have lots of energy in the 6-10khz region but little above that. So, I applied an eq of about +6db spanning 2.5 octaves centered at 35-40hz. I would also add from .5db to 1db of boost centered at 18khz with the same q as the bass.

The results from applying this eq almost always make these damaged tracks sound much more live and dynamic without significantly changing the tonal color of the original. Because there is little consistent information in the 35-40hz on most of these tracks and because that range is more felt than heard, the tracks sound both more relaxed and dynamic. It is also clearly visible, looking at the waveform of the track, that the dynamic range has been increased.

JA's observation regarding high pass filtering seems to corroborate with my new practice of fixing bad recordings on my computer.

Great ideas, Erick.

Have you tried using one of the compression plug-ins that will perform dynamic expansion?

We don't know what compression was used, but some gentle overall expansion might add some dynamics.

No, I have not tried using an expander.

How do they work?

It's the opposite of compression. The output has greater differences in level than the input.

A number of compressors allow expansion. Instead of setting the compression to 1:2 (for example), adjust the compression to less than 1. This is expansion and will open up the dynamic range.

It may or may not do something you find pleasing - but seriously squashed recordings can be partially brought back to life this way. It may be worth it given that you are already playing with EQ.

In my line of work I deal with systems that use much signal processing. I work on the pre-processed signal side. While I don't have a very good understanding of signal processing, I do know that once you have processed any signal you have lost the original data. Any additional processing will not be accurate because of the loss of data from the previous processing.
You can not accurately expand a compressed signal.

Quote:
In my line of work I deal with systems that use much signal processing. I work on the pre-processed signal side. While I don't have a very good understanding of signal processing, I do know that once you have processed any signal you have lost the original data. Any additional processing will not be accurate because of the loss of data from the previous processing.
You can not accurately expand a compressed signal.

That depends on how it is compressed, and how much, if any, processing is done afterwards.

It is, barring issues of noise floor (which are irreversable), possible to do a level (or loudness) compression that is reversable, or very nearly so (depending on the actual mathematics of the device). Such things as 'backward compressors' (where the level adjustment is driven from the output, rather than input, side) can, if they are not "infinite compression", be reversed, with the caveat about low-level noise floor, some of which can actually be dealt with.

This kind of compressor, however, usually doesn't provide the sound that people who use compressors prefer, so it's not that easy, and most modern compression is "look-ahead", and multiband at that. There, you are, for the most part, out of luck.

Things like Dolby A and B, when implimented and tightly calibrated, even do something reasonably appropriate and can be inverted to the level you can invert something in an analog system, however issues of phase shift and frequency response make them only "mostly invertable". They are backward compressors. Ditto dBx processing of a while ago.

The problem is that it's mostly "lookahead compressors" that get used, and these have a forward side-chain, which can remove any trace of the previous level adjustment.

As you say, if you do remove the information, it's gone for good. It's just not true that a compressor HAS to remove such information. Some do, some don't, some can but may not, depending on settings.

It's also not true that "once you have processed a signal", for instance, lossless compression is a kind of processing, and you can certainly get back the same signal.

And, just as a favor, please don't confuse "loudness" and "intensity". For that see www.aes.org/sections/pnw/ppt.htm and see the "loudness tutorial".

You are correct that not all processing is irreversable.

Quote:
And, just as a favor, please don't confuse "loudness" and "intensity". For that see www.aes.org/sections/pnw/ppt.htm and see the "loudness tutorial".

Great set of tutorials JJ!

I'm interested on your findings on how we perceive phase as it relates to low bass. Does the perception of phase in bass effect the way we hear the rest of the audio spectrum or is it a seperate issue? Do we localize the upper frequencies differently because of the interaural phase in bass or does the bass create its own sense of space?

Quote:
Various people have reported, sometimes anecdotally, that above 40Hz (and below 90Hz), although one can not localize a sound source, differences in interaural phase can create a sensation of space.

Quote:
You are correct that not all processing is irreversable.

Just regrettably most of what is considered "loud enough for release", that is irreversable, since it involves mega-clipping. (unbelievably rude comment omitted)

Quote:

I'm interested on your findings on how we perceive phase as it relates to low bass. Does the perception of phase in bass effect the way we hear the rest of the audio spectrum or is it a seperate issue?

What one seems to get is a sense of "wide" vs. "narrow", which, of course, would make sense if one considers the pattern of zeros in a wide vs. narrow space.

There is no affect on direction, nor any sense of direction, merely a sense of width or not.

The problem with one sub in a room is that you can NOT achieve this. What's more, it's more "accurate" in a frequency response sense (using pressure response, which is inappropriate for a variety of reasons, but that's another subject) to use a single sub in a corner, etc.

But that's not the whole story. The energy in a soundfield is in four variables, 3 volume velocity and 1 pressure. At the ear drum, you only sense pressure, however, anything that nears quarter wavelength in size creates pressure from velocity.

I've been having this debate with Welti and Olive for some time now. Holman and Greisinger both have said things of a similar sort to what I said above, but I won't put words in their mouth.

Note, this does not mean that the old test of directional sensitivity is wrong, it is just incomplete in that it does not account for all sensation.

Thanks JJ.

Fascinating stuff. I have been observing this in my own listening - the better I dial in the bass through speaker placement(and perhaps the better the amps are keeping bass in phase) the better sense of width, depth and ease I hear in the soundstage.

Interesting to me that sound that has no directionality effect directionality.