[Frank Stearns]

Edi Zubovic edi.zubovic[rem writes:

snips

-The second case is not good at all. It shows significant DC offset
which only steals a lot of bandwidth, ie. dynamics, among other
things. Luckily, the remedy can be done easily and is
straightforward. I wonder why is this being so often overlooked at
production of CDs. To eliminate DC offset, activate the DC offset
removal tool in your audio editor or do a high pass filtering at very
low frequerncies such as 16 Hz.


Edi is absolutely right.

A trick a multi-Grammy winning engineer taught me was to HPF every mix input channel
with a 24 dB/octave (or steeper) slope. (He uses the Oxford EQs and has 36 dB/oct
available; I'm using the regular PT EQ and only have 24 dB available.)

There's usually a subtle but noticeable increase in overall clarity to a mix when
you do this. Not only DC, but simply subsonic and LF noise can chew up overall
dynamics.

This wasn't a problem in the olden days when tape and transformers passed little
below 20 Hz. But with digital going to nearly DC, junk down there is a
consideration.

But beware DC removal tools, at least the one in Sound Forge 8.

A while back I was on this topic and discovered the following:

- the PT high pass actually is removing DC (I had my doubts for a while), but LF
noise appears to re-introduced in the mix bus -- internal dither, perhaps? Depending
on what I'm doing, I'll sometimes put a final HPF as the very last plugin on the
master fader strip.

- Sound forge 8 DC removal appears to /add/ junk at 1, 2, and 3 hz that WASN'T there
before running tool -- but not always. Odd.

On the other hand, I've seen great-sounding CDs from the Mastering lab show the same
big hunks of energy at 1, 2, and 3 hz. Are they using Sound Forge??? Is there some
common DC removal algorithm that does this for some obscure reason?

I'm more concerned about noise between 10 and 20 hz, as this might have a chance of
getting through in a better home stereo. We can probably ignore noise below 10 hz
(but not DC).

OTHO, remember the LP days, when a bad warp on a good system caused your LF drivers
to jump around A LOT at something way less than 10 Hz? That can't be good for IM
distortion performance....

Regarding asymmetries: Some instruments and voices indeed produce more air
compression and less rarefaction as they make noise -- thus the differences in
waveforms... 'Tis nature.

Frank
Mobile Audio

--