Cant help wondering about what recording volume (amplitude on the viewable waveform) I should ideally have when doing multitrack recording.
If I have 2 tracks both recorded with max amplitude reaching the 0.5 mark on the scale (half of max value) and playback them they will produce an added max amplitude of “1.0”, i.e. max allowed signal strength, provided I have each individual volume control at “0 db”. Correct?
But what should I aim for when multitracking? Suppose using 32 bits bit depth means quite forgiving performance even when mixing fairly weak signal.
But if I continue the reasoning above then, if I have 4 tracks with 0 db on all the volume controls, I should have no more than 0.25 amplitude on each track in order to avoid clipping. Is this correct? Ok, but if I go on add up tracks like this the amplitude values will be ridiculously small for the individual track, so it ought to be better to aim for a “reasonbly strong” amplitude (at least so large that you can assess the waveform visually) and them dim the volume with the volume controls?
But if so, what should I aim for? perhaps a max level of “0.5”, in order to have ample headroom for unexpectedly strong notes?
yes there is an optimum.
should be low enough to avoid clipping ever
and high enough to avoid thermal type noise being heard
(not external noises like horns honking chairs squeaking)
switch to db mode forget linear display
do a test recording
set the highest points to be at least -18db down (i use -24)
magnify the display to help judge this while recording. and you
use the amplify function afterward to see what the actual headroom was.
if you double the tracks then move another 6db down to avoid clipping when you combine them
listen to the recording - if no hiss during quiet passages stick with it
if you can hear noise then you may be able to move the recording level up - but be careful as hiss is much less bad than clipping - can be “fixed” somewhat (clipping never can be fixed), and is less annoying to listen to if you can’t fix it. and most people will want to compress the signal to make it listenable in a home or car so you should always be way above the noise anyway so err on the side of no clipping .
The maximum amplitude will not be greater than 1.0 in this situation, so that’s correct although the maximum amplitude may be less than 1.0. You will only get a 1.0 peak if peaks at 0.5 in both tracks coincide.
Your reasoning is correct (subject to the above proviso that the peak level in the mix may be less than 1.0)
Absolutely correct, well done Holmes.
How much “headroom” you leave depends on the type of sound you are recording and how you are recording it.
For example, if you are recording an electronic organ, you can make a pretty accurate estimate of what the peak amplitude will be - just play the loudest sound that you can and set your levels so that that is a little below full-scale. On the other hand, if you are recording an inexperienced vocalist, the peak level can be massively variable - just when you thought they couldn’t go any louder they will get right on top of the microphone and hit an extra 20dB on their “top C”.
It’s always best to err on the safe side - as already said it near to impossible to fix badly clipped audio.
So how to adjust the track levels…
You can use the “Amplify Effect” or you can use the track volume sliders. In Audacity 1.3.12 there is a “Mixer Board” (in “View” menu) that provides more precise control over the track volume sliders (it works rather like having a simple “Mixer/Console View”).
If you are working in less than 32bit, you should use the track volume sliders (or the Mixer Board, which does the same thing). Adjusting track levels this way is non-destructive, so you can twiddle them up and down to your hearts content without affecting the sound quality.
If you are working in 32bit, you can also use the Amplify effect, which will be very nearly lossless. For multi-track projects, 32 bit is definitely recommended as you are likely to be doing a lot of processing and tweaking of the tracks. These losses will be small, even with 16 bit audio, but they mount up. With 32 bit audio the precision is hugely greater and even after a great deal of processing the losses will remain extremely small.
I will often use a combination of the track volume controls and the Amplify effect - whichever is more convenient at the time. Usually it is the track volume controls, as this keeps the waveforms nice and big and easy to “read”, but sometimes you may want to just raise or drop the volume of one particular sound clip by say 1dB, in which case it may be easier to just use the Amplify effect.
Audacity 1.3.12 has many enhancements over the old 1.2.6 version that make it a lot better for multi-track projects.
Almost forgot - In Audacity 1.3 you can “grab” the Audacity meter-bar by clicking on the left hand edge of it with the mouse, hold the mouse button down and pull it out of its “docking” position. Click on the right hand edge and you can stretch the meters to the full width of your screen. Drag it back to a "docking position and it will snap back into the main Audacity window with its new larger size. It’s much easier to see the recording level if you make the meters big.