Re: [linux-audio-user] Peak Level Monitor/Limiting[Fwd: linux-audio-user Digest, Vol 42, Issue 125]

From: Dragan Noveski <perodog@email-addr-hidden>
Date: Fri Mar 30 2007 - 01:54:15 EEST

Phil Mendelsohn wrote:
>
> Message: 5
> Date: Thu, 29 Mar 2007 23:38:30 +0200
> From: Dragan Noveski <perodog@email-addr-hidden>
> Subject: Re: [linux-audio-user] Peak Level Monitor/Limiting
>
> i think, recording at low level, you are kind of "loosing disc space".
> lets say, if you are recording at 24bit, on low level the signal
> information will take only 12 bit (example) and the rest of 12 bit will
> be only empty data information.
> don't think that one can repair that with some gain plugin or
> normalization...
>
>
> If you think about this for a couple of seconds, you may
> decide to reformulate your opinion. In a low-level signal,
> you say the "top" (Most Significant) bits will be zero. In
> a high-level signal, though, the bottom 12-bits are all ones.
> What is the difference?
>
> In the computer, a signal is a signal, regardless of the
> value at which it is represented. It is only when you try and
> look at the distance from that signal to either the noise floor
> or the max, or deal with non-linear perceptual effects (equal
> loudness curves, etc.) that level may come into play.
>
> It's not the absolute level that matters, it's the dynamic
> range.
>
> Cheers,
> Phil M
>
still not agreed.
if you use 24bit you can 'describe' the signal much more exactly, as if
you use 12bit.

try to convert any soundfile into 8bit and listen to it!
so if you record at 24bit with low level, perhaps you ll ending up
'describing' the noise of you hardware, but not a signal.
and recording at low level means less dynamic, that's right.

cheers,
doc
Received on Fri Mar 30 04:15:06 2007

This archive was generated by hypermail 2.1.8 : Fri Mar 30 2007 - 04:15:06 EEST