Audio engineers and musicians have a habit of collecting things. Not only do they collect things, they collect things that are similar in nature and believe they are distinctly different. Furthermore, they collect them without actually testing to see if they are indeed different.
For a field that is supposed to pride itself on being both artistic and scientific this leaves a lot to be desired on the scientific end.
There are plenty of ways, however, in which you can test collectibles and see, definitively, what is different and what is the same. When trying to test out preamps, outboard racks, plugins, and more one of the best tests you can do is the null test.
The Null Test Explained
Simply put, the null test is perfect phase cancellation.
Assuming you have two in-phase sine waves, if you were to invert the polarity of one sine wave you would be left with perfect phase cancellation. That means no sound whatsoever.
Even if you were to run the sine waves through the same compressor, EQ, gate, etc., before you invert one sine wave, the result would be the same. No sound.
The null test, however, becomes an actual test once you run the sine waves through two different effects, preamps, and so forth, and try to match them.
Take an example whereby you run the first sine wave through EQ-1. In EQ-1, say, you cut 200Hz by 3dB with a Q of 0.5. Then run the second sine wave through EQ-2. In the same way, set EQ-2 to have a cut at 200Hz, by 3dB with a Q of 0.5.
Assuming both the EQs are exactly the same, you should get a perfect null. No sound.
However, if the EQs are not exactly the same, despite your best efforts to match them, then when you'd null test them. You're left with the difference between the two pieces of gear.
What the Null Test Shows
The null test is a way for us to accurately measure the differences, if any, between two pieces of similar gear. You do not listen to the difference with qualitative ears. Often times the difference in amplitude is so small and low that we can safely call it inconsequential. By removing any qualitative listening, we are able to remove human bias. Just because one plugin is nice and shiny, it doesn't mean it is any better!
Setting Up a Null Test
Whilst you can use old-school analog tone generators, and suchlike, the easiest way to set up a null test is inside a DAW.
Here is the basic null test:
- Start by putting the play head at the very beginning of the DAW's timeline (00:00:00)
- Either using a built-in function or a tone-generating plugin, record one minute of a 1kHz sine wave. This sine wave should be anywhere from -10 to -20dB down, so that you have room to play with dynamics-based effects
- With the rendered sine wave, duplicate this sine wave onto another track and make sure it is exactly at 00:00:00
- Now flip the polarity on the second track and play back both sine-wave tracks. You should hear nothing. If you hear something, zoom in on the sine waves and ensure you did not accidentally move them by a few samples.
- With the sine-wave tracks now perfectly lined up, you can add any dynamic effects you like
Of course, you can—and should—test with other test tones as well. It just so happens that 1kHz is the standard test tone for most audio gear.
The only problem with the above test is that it does not allow for EQ plugins, or plugins that react differently to various frequencies.This is because there is only one tone.
By replacing the 1kHz sine wave with pink or white noise, you will be left with a nice full spectrum to work with. Again, just ensure the tracks phase cancel out first before you try them with effects.
Understanding Null Test Results
You may notice that it can be hard to hear the results of a null test, if at all. This is because unless one of the plugins you are comparing happens to be extremely nonlinear, a lot of them end up being about the same.
If this is the case, you can safely say that it should not matter which plugin you use.
You will want to test the plugins at various settings. Some compressors react very differently at low thresholds, some EQs have noise additions. If you still cannot readily hear the difference—perhaps because you are in a noisy environment—you can also add a FFT, gain or spectral analyzer to the master output track.
Using a visualization plugin will also help showcase any differences that the speakers may not be able to produce.
If you see a lot of activity at 20Hz, that can make a huge difference if you can hear it or not.
Here are few points to keep in mind:
- Any audio information from about -60dB and up left after a null test should be considered audible. Be sure, however, to double-check you are perfectly phase cancelling the original test tones
- Anything floating around -90dB can be regarded as functionally inaudible. While you technically could hear it if you absolutely blasted your speakers, chances are that will not happen. Furthermore, in the mix that difference will probably end up getting buried
- Anything around -110dB or lower is completely inaudible. Any two sources with a difference this low can be considered exactly the same
As stated earlier, the null test is a way to clearly and accurately show the differences between two similar pieces of equipment.
This can be a great cost saving measure if you are able to demo any gear before actually buying it. While this tutorial dealt with null testing plugins, next time I'll examine null testing physical equipment as well.