Topic: if a digital recording goes to 0db, its ok, but if its over....

[heyitsmejess]

...then you run into brickwalling, right?

please correct me if im wrong.

thanks!

[nickee]

No. Brickwalling is when the preamp of the recorder overloads. Lowering the input gain wont fix it and this can happen regardless of what levels the recording peaks at. This is analog distortion. This happens mostly when recording loud sounds using microphone inputs or recording from a soundboard that sends +4dbu. The recording will sound fine during the quiet parts but will sound bad during the louder parts.

If you record a signal and the meter goes over 0db the digital distortion that results is usually called "clipping". This is when the analog to digital converter overloads.

[ghellquist]

Agree that clipping is a better term. Brickwalling I have only seen used in describing limiters.

Regardless, what it will sound like depends on the exact equipment. Some of them sound decent despite beeing above 0dB, some sound like total disaster. The safe bet is stay well away from 0dB FS. Personally I work hard to not get peaks above -12dB FS when tracking, and then raise levels at home. But then I only record classical acoustical music.

Gunnar

[heyitsmejess]

thanks for the replies, all!

i had a recording that came out with several peaks at exactly zero, but it hasnt went over.

ive been listening for the tell tale "pops" that come with a digital peak, but cant hear them, tho soundforge picks them out as clips.

i guess i will just go with it.  thanks again!

[Ozpeter]

Well, just to confuse things, there's something called intersample peaks - see http://www.sonicstate.com/news/shownews.cfm?newsid=5526

[DSatz]

Ozpeter, this is a good piece of evidence that the conventional mental picture which most of us were taught about digital recording has a few things wrong with it. If that mental model were correct, this should not be possible, yet digital signals that will cause intersample peaks above 0 dBFS in a DAC can be calculated for any given output filter design--just take the inverse of the filter's impulse response.

But such signals have essentially no chance of occurring in live music recording; the probabilities are astronomically small. If everyone on earth made digital recordings continuously from now until the Sun burns out, it still might or might not occur even once in all those recordings.

--best regards

[itook2much]

I use Sound Forge, & on some shows I've had peaks that hit 0dB (or over) that showed as clips, but there was no audible distortion.  So an occasional clip may be OK, but you wouldn't want the whole recording that hot.

[Ozpeter]

But such signals have essentially no chance of occurring in live music recording; the probabilities are astronomically small. If everyone on earth made digital recordings continuously from now until the Sun burns out, it still might or might not occur even once in all those recordings.

I have a strange way of getting my fun - when I record classical concerts, I take a perverse delight in trying to get the highest peak exactly to zero.  Any normal person would use more sensible settings... now whether amongst my recordings I could in fact find an intersample overload I don't know, but I suspect there could be one or more in there somewhere.  But I don't plan to look!