Avid and Apple conspire to heist 9 decibels of level
Starting out with a sine wave in Avid Pro Tools at 0 dBFS, by the time it gets to the output of Apple Final Cut Pro X it is a mere -9 dBFS in level - with no intentional level changes in between. So what happened? The video explains all...
Transcript of video...
Here on the one hand we have Avid's well-respected Pro Tools digital audio workstation software. And on the other hand Apple's equally respected Final Cut Pro X video editor.
So how come a piece of audio recorded in Pro Tools can lose 9 decibels by the time it is output by Final Cut Pro X? Here's how...
It starts with a sine wave
I'm going to create a simple sine wave in Pro Tools. To do this I'll make a blank clip, then use the signal generator to create a sine wave of 400 Hz at 0 dBFS. I've chosen 400 Hz because it isn't too hard on the ears, and 0 dBFS because it makes the level differences obvious and easy to calculate.
Don't worry - I'll lower the levels of the tones in the video soundtrack.
So here is my sine wave, on a mono track. I want to export it from Pro Tools as a mono file, then import it into Final Cut Pro.
Pro Tools has options to bounce into multiple mono files, a stereo file, or a mono file in which the output channels are summed. Here I'm going to bounce to multiple mono, and use only one of the two identical files that are created.
Import into Final Cut Pro X
So far, so good. Now I'm going to import the file into Final Cut Pro X.
I already have a new library, a new event and a new project all ready to go. So I can simply import the sine wave file that I created. Here we go...
This all seems very easy. I can play the sine wave...
But let's look in a little more detail. I'm going to open up the audio meters so that we can see the level of this sine wave.
-9 dBFS. Somewhere between where I started and where I am now I've lost 9 decibels of level. And I didn't intentionally change the level anywhere in the process. To put this in perspective, a change in level of 3 dB is important, 6 dB is significant, 9 dB is... Well that's a hell of a lot of level to lose accidentally. So let me show you just how it happened.
How the 9 dB loss in level happened
Going back to Pro Tools we can see from the meter in the track that the level of the sine wave is 0 dBFS, so there's nothing funny going on there. But if I open up the master track where the level of the output of Pro Tools can be measured, we can see something different. If you look closely, you'll see that the meters in the master track only come up to -3 dBFS.
This is because of the pan law. If I pan the sine wave left, it registers 0 dBFS in the left meter of the master track. If I pan right, then it is 0 dBFS in the right meter. If I pan centre, it is -3 dBFS in both. This is so that wherever I pan the signal, it will sound approximately the same level in the stereo sound field.
Pan law can work in different ways. It can be set so that the level drops 6 dB when panned centre. This works perfectly when a stereo mix is collapsed into mono. Wherever a signal is panned, it will come through at the same level in the mono version.
But for stereo this works better when the pan law sets the drop in the centre to 3 dB. 4.5 dB can be used as a compromise but I'm happy with 3 dB as a default.
The other 6 dB
This accounts for how 3 decibels were lost. So what about the other 6?
Well, it's down to pan law again, except it's hidden behind the scenes in Final Cut Pro. In Final Cut Pro, when a mono file is placed on the timeline, it is automatically panned centre with a drop of 6 dB in both channels of the stereo output. This complies with the 6 dB pan law and also makes absolutely sure that even the highest level mono signal can't cause clipping if the output channels are combined.
So I was able to cure the loss of 3 dB in Pro Tools by panning the signal to one channel. In Final Cut Pro I can correct the level loss simply by giving the audio clip in the timeline 6 dB of gain.
And when I do that, my output meters are back up to 0 dBFS.
Helping or hindering?
I have to ask the question whether all of this is helping or hindering the user? Well, an unexpected drop in level of 9 dB is definitely not a help. But on the other hand, having a 3 dB or 6 dB pan law is more helpful, more of the time, than having no pan law at all. (Or I should say more correctly, a pan law with a 0 dB drop in the centre.)
The important thing is to recognise what's going on here and understand it. Level changes can happen 'behind the scenes' and the user needs to be aware of the processes involved so that they can take any corrective action that might be necessary and - importantly - not be caught out by any unexpected results.