[Top] [All Lists]

Re: Digital Distortion

Subject: Re: Digital Distortion
From: Walter Knapp <>
Date: Wed, 25 Jun 2003 00:38:55 -0400
Dan Dugan wrote:
> Klas, you wrote,
>>I long for some real digitfreak to give me a list, like 1/ First do
>>filtering, 2/ then dynamic work, 3/ then noise gates, 4/ finally
>>normalizing. Just very practical and straightforward rules! And teaching =
>>the errors with doing in another way.
>>I know people who normalize between each step, say, all 10 steps. It appe=
>>to me to be a dangerous way, but they claim it's okay.
> I have an opinion on that: never normalize. Never. No need for it. At
> the last stage of CD mastering, overall levels might be raised so the
> loudest sounds are close to max or limited close to max, but that's
> as far as you want to go. In intermediate steps, you need headroom
> for processing.
> Also, sounds that are naturally low-level, like ambiences, should not
> be boosted up just so people have to turn them down to listen to
> them. At the NSS workshop someone asked me about ambiences with peaks
> reading at -20 or -30 dBFS, should they be normalized. No!

I'm somewhere in between, I guess.

I certainly think software should not have a button that says normalize
at all. You should have to make intelligent decisions about any gain
adjusts and set the values. I do like it that my software can be asked
to tell me just what the headroom of a selection is. Then I can decide
what I want to do to it. I only adjust for a reason, not just because I can=

As I have noted, what's needed in a final output CD is that it be
reasonably playable. Want to get hated? Just put out a CD that most
listeners will be adjusting the volume for each track, or maybe several
times in each track. We listen to CD's for pleasure, and don't want to
have to do extra work.

I'm not so convinced that no gain adjusts between stages of filtering
are going to be so good. Depending on the type of filtering, you may
have a considerable drop in level going through a filtering sequence. I
expect that's why Spark XL has between each modular location of it's
filter bank both a mini level indicator and the ability to set a gain
adjust. I certainly don't adjust near all I could, but if I see it
tailing way off through the bank, I'll go in and do some adjusting. I
also sometimes see a increase in level too, and if that's not adjusted
back down it will come out distorted. I've done considerable
experimenting listening to the output, watching the sonogram. I've come
to the conclusion that many, if not all filters are designed to work
optimally at certain sound levels. There can be considerable change if
they are fed less optimally.

I should probably do more tweaking in this area than I do. But it eats
considerable time to fine tune. Eventually you have to get the work
done. One advantage with Spark XL is that you can save a entire filter
bank with all it's settings. Like I can make up one for river frogs, and
it will be pretty close for other river frog recordings, saving a lot of
time. (and my set of banks for every frog species is one thing I lost in
the disk crash)

Should a CD track that's reading -30 be adjusted? Well, suppose the next
track reads -5. You are going to get a jump in volume, and have to ask,
if it belongs, and if the listener will tolerate it. Or just get mad at
you because they have to jump for the volume control. Balance is the
key. Don't make the listener get out of his easy chair, don't break
their enjoyment.



<Prev in Thread] Current Thread [Next in Thread>

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU