Originally Posted By: Frankp

I know we used to record at real high levels to keep the tape hiss to a minimum. Does that still make sense with digital recording where there isn't any tape hiss?


In a word, no.

The two methods are entirely different.

With analog magnetic tape recording, Saturation of the magnetic media was desirable in most cases.

But the modern PCM Digital recording is recording a digital number that corresponds to the amplitude of the signal measwured for each sample in time, according to the sample rate. In this method, we are not going to gain anything by increasing those numbers, and may even get into the bad territory of raising those numbers so high as to create distortion or clipping problems at the digital level.

The advent of the 32-bit PCM digital recording "engine" has greatly alleviated the problem though, previously, when all we had for PC recording was the 16-bit engine, there was the problem of digital clipping that was caused due to forcing those numbers to the top of the available bits and thus they would suddenly wrap around to the bottom, creating a nasty sound that we called, "Digital Thwack" or just "Thwack".

All that aside, there are still places around the web to be found where some explanations of digital recording tout setting the recording level for as high as possible before reaching the 0dBFS mark on the recording VU meter. I do not recommend that procedure. It can lead to ruining an otherwise good take because a section of the music performance could easily exceed 0dBFS mark due to the dynamic abilities of the performer. I try to set my levels for recording where the VU meters in the recording program are hovering below 0dB for the peaks, somewhere around the -6dBFS mark or so, sometimes lower than that if the instrument is drums, or a very dynamic singer.

When setting levels for someone, there is also the human factor to consider. If you ask the performer to sing or play "as loud as they can" they will attempt to do so but typically will always hold back for some reason. Then, when actually performing and recording the track, the emotion of the thing may come out and they will easily exceed the amount of amplitude they generated during level-set testing. So I allow for that when setting the level. I'd much rather have to perform a Gain Change on a track, or even resort to Normalization of that particular track if it contains Strong Performance rather than risk losing the section where they got too loud to the ravages of digital overshoot or clipping, even with today's 32-bit engine, which is very hard to force into digital clipping but doing so can indeed make for some distortion at playback sometimes.


--Mac

PGmusic FAQs, Tutorials and Updates! click here

You must be Audiominds.
www.audiominds.com