cliveb wrote:
> Pat Farrell;248484 Wrote: 
>> When you "normalize" you raise the music values, and also raise the
>> noise values. The music is louder and so is the noise.
> True, but if the noise of the source that was recorded is well above
> the digital noise floor, then it's a moot point.

I was really talking about noise added by the conversion process. In 
your cartridge, turntable rumble, your cables and preamp, in the ADC, etc.


> By increasing the recording level in order to achieve close to 0dbFS,
> you're amplifying the noise of the source along with the signal. So if
> the source's noise is vastly higher than that of the recording chain,
> it doesn't really make any difference.

It depends totally on the source of the noise. Driving a louder signal 
into your ADC will improve the SNR.


> What if you happen to choose a first-pass recording level that causes
> digital clipping? The peak level of the recording will be 0dB, and you
> have no way of knowing how far over it would have gone had there been
> more headroom.

Don't do that. With good audio tools, you can find/see the instantaneous 
peaks. I use Cakewalk Sonar, it makes it easy.

If you are just guessing, you are completely correct. I was assuming 
some science.


> (One last point: I'm pleased to note that you didn't bring up the
> rounding errors bogeyman as an argument against normalisation).

Rounding is a pure red herring.

If you are hard core, you'd record at 24/88.2, and then use proper 
dithering to truncate to 16 bits.


-- 
Pat Farrell
http://www.pfarrell.com/

_______________________________________________
ripping mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/ripping

Reply via email to