[linux-audio-dev] Safe-Ears/software volume

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: [linux-audio-dev] Safe-Ears/software volume
From: Juhana Sadeharju (kouhia_AT_nic.funet.fi)
Date: Tue May 23 2000 - 15:17:52 EEST


Hello. As it become clear a while ago I'm going to introduce some safe-ears
technology to my audio software. I feel it worth of telling out of context
of Alsa, and indeed, it has nothing to with Alsa because I were worried also
about other things in that debate.

Typically audio engineers crank up their mixer volume to hear fades.
When my friend mastered a CD for me he occasionally forgot to put
the mixer levels on the console back to normal positions and we got
a sudden loud audio next time he played the edited audio.

I have mastered some cassette copies and there I have had to crank
volumes very high levels to hear the fades which fades in to the
tape noise. I have been extra careful but it would be nice to not
worry about too high volumes at all.

The solution is a software volume. I have not seen such a software volume
in any editor software or a console yet. Tell me if I have missed them!

The software volume is used only for monitoring and therefore it doesn't
affect to the audio signal itself. The quality of original audio is
maintained.

The idea is simple: you set up the hardware volume to a reasonable level
and when you want to hear low level signal, the monitor signal is boosted
with software volume. If now you play a normal level audio, will it be
software limited and won't be too loud.

Actually a software volume can still be too high. Just first listen to
fades and then some random noise signal or audio which has been heavily
boosted and hard limited to full scale level. A help could be in using
other limiting/compression/expansion techniques so that very loud and
clipping sections of the audio would vanish entirely in monitoring fades
(which doesn't clip at all).

While the software volume saves the ears, it also improves the audio
monitoring. In hardware, the lower bits vanish and you don't hear them
even if the volume is raised up. With the software volume you can hear
them easily because the signal is played at volume levels which are
optimal for the hardware.

In a wave editor for the editing purposes, the volume levels (as well as
the waveform graph) could be enlarged automatically according to the maximum
visible peak in the current display.

As I mentioned, typical cassette/LP fades fades in to the tape/LP noise.
You cannot find the cut point by inspecting the waveform display. You have
to raise the volume. To this problem I suggest to use a spectral display
for finding the fade points. In spectral display any signals below the
waveform noise level are more better visible. (Also, dithering techniques
introduces a noise level and waveform display hides the possible signals
in the noise.)

Comments?

Juhana


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Tue May 23 2000 - 15:58:01 EEST