[Top] [All Lists]

Re: Re: Digital Distortion / Meters

Subject: Re: Re: Digital Distortion / Meters
From: Walter Knapp <>
Date: Mon, 23 Jun 2003 15:49:19 -0400
evertveldhuis wrote:
> Walt,
> VU meters (analog needles) are average readings, but LED's on digital 
> equipment are always peak readings (unless specifically stated 
> otherwise on the meter or in the manual).
> The problem with digital meters is they read AFTER the preamp and 
> AFTER the ADC. 
> So if the signal is already too hot for the mic (high SPL) or too hot 
> for the preamp, it is already distorted BEFORE the AD converting, so 
> any digital level indication does not mean the signal is undistorted. 
> It just means the energy of the signal is at a certain level. It 
> tells nothing about the possible clipping of a signal. 
> You have to monitor / listen to the signal with good headphones to 
> judge wether a recording has clipping or not. 
> Most of the times it is analog clipping in or before the preamp 
> section I am afraid.

Certainly the analog clipping is also a problem. Particularly with 
consumer gear. I don't have much problem with the mic, except for sudden 
close impulses into the Telinga. I'm sure some of those the mic overloads.

All that the digital reading coming from the digital stream means is 
that it won't show readings over the clip level. It still shows what's 
going on when you are set anywhere near right. If it's showing you 
getting up to 0db, you are probably clipping.

If the digital meters were giving you actual peaks, they would show you 
the correct pad particularly for the A/D, even if they are after 
everything. I think part of the problem is that a lot of the digital 
meters are set up to simulate the readings of analog meters, they 
average according to the standards set for analog meters. They also rise 
a decay based on the ballistics of analog meters.

Unless they have a hold function they have to average, we cannot take in 
44,000 readings a second in a meaningful way even if the meter could 
display them. In actuality the meter is ultimately limited by the 
display refresh time, which is much slower. And our eyes are even 
slower. 30 frames a second is flicker free for us. Yet in 1/30th of a 
second there could be quite a few wavetops to clip. And clipping even 
just a few is noticeable.

Do not trust digital meters to be giving true peaks even in the digital 
stream. I know the portadisc records much higher levels than it's peak 
reading would indicate. In some cases the difference may be enough to 
clip the A/D with a 10dB pad in the indication. Though I know most such 
situations and use a larger pad.

So, with the Portadisc, the meter will not be giving you a reading 
that's direct off the digital stream, it's a average even for the peaks 
off that stream. I think the margin display does do better, but it's 
hard to watch that. Two numbers (for the two channels) constantly 
changing. (again they have to be averages) It would be nice if that had 
a longer hold. It's numbers agree with what I find after digital 
transfer into my computer. Or very close. So, it appears to be what you 
describe, calculated from the digital stream. I do check it once in a 
while to get a feel for the sharper impulses. Ideally the Portadisc's 
level display would include the margin numbers in the same display. I'm 
betting that would help a lot.

It boils down to you have only the level indicator to help you. And 
experience in what happens with certain indications. Setting it without 
reference to the meter is sure to create problems. Use the meter, but 
use it intelligently.

Your other choice is to carry a accurate sound meter with a good peak 
reading. Then develop a table for each mic as to what to set the gain. 
Not very practical. Though as a informal check it can help.

And, yes, monitoring with headphones is a must. Though it's a different 
path through the recorder than the recording. That will help to catch 
some of the front end problems. Though I find that those do make it 
through into the recording, so are reflected in the level reading.



<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