Mastering software

[cuttime]

If we ever get quantum computers we will be soooo screwed.

[bayswater]

If we ever get quantum computers we will be soooo screwed.

We'll be screwed and not screwed, and everything in between at the same time

[Shooshie]

Let me explain my supposed "passion for null." It doesn't exist. I go for years without thinking about a null test. My passion is for precision. When we're going to talk about null tests or the concept of nullification in any form whatsoever (for example: it's used every day in noise-reduction software), we absolutely, positively, resolutely, timbuktootly MUST agree on one thing: the definition of null.

Now that we can agree on what it means, we can go on to have meaningful discussions about how it can be used, when and how it can be trusted, whether it's a practical concept when used as a tool, and so forth.

But before any of those conversations can have any meaning, we HAVE to agree that nullification is the result obtained when inversely combining two identical strings of digital audio samples. And for that to have any meaning, we must understand the word "identical." It leaves no room for error. If two tracks of digital audio samples are identical, there can be no difference between them. Any perceived differences are merely the inconsistencies of the mind as it listens twice to the same thing.

So, subtracting one digital string from its identical twin leaves you with...
000000000000000000000000
000000000000000000000000
000000000000000000000000
000000000000000000000000
000000000000000000000000
000000000000000000000000
000000000000000000000000
000000000000000000000000...

Each of those is 24 digital bits. Put together 44,100 of them, and you've described a wave (or lack thereof) lasting one complete second at 44.1K/24 bit resolution.
That is null.

From that point, we can talk about degrees of nullification. For example, 99.99% null might yield a different value after every thousand samples. 90% null gives you a different value after every 9 samples.

Then we can discuss how far off that value can be. Is it -132dB? Is it -10dB? Are the non-null values concentrated in short strings, forming a wave?

Once we start talking about this kind of thing, the concept of null begins to have some usefulness. Absolute nullification (100% null) probably occurs very rarely.

So why is it so important to come to grips with the definition of null? Because we must have some way of defining real differences vs. perceived differences. In the case of two identical files that produce a null result in a test, any perceived differences between the two are put there either by fluctuations in one's ears and mind, or else some other variable is jacking with the audio during playback. If one person claims to hear a difference, but nobody else can, or if people claim to hear a difference, but in subsequent playbacks they cannot consistently pick out the same file, then we know that they are hearing things that aren't there.

No shame in that. It happens to me and to every last one of you. I'm forever astounded by the difference I hear at the end of the day vs. the start of a day. My ears literally grow numb. Come back to a long mix after a good night's sleep, and I want to do it all over again, because I wasn't hearing the glaring things I'm hearing now.

During playback you might even glance over at your girlfriend and remember a moment of passion with her for just one second, and in that second the audio will sound magically different. Our perceptions are influenced by our thoughts. Our minds are fluid and imaginative. And thank goodness for that!

Or maybe the engineer will push the magic red button on her mixing board, and you'll swear you heard a difference, but the button isn't connected to anything.

We are merely trying to establish that there can be identical files with NO difference, but some people may still hear differences, because their minds are biological computers that are not consistent from one second to the next.

So, the null test gives us one method of identifying when this is happening. The next debate would probably concern at what percentage and degree of nullification does a difference cease to be audible even to the trained ear. At what degree and percentage does the difference rise above the self-noise of hardware and amplification and become a real "sound?"

When I was a freshman in college in the early 1970s, taking physics, the professor was using a wave generator to produce ever higher wavelengths. I had good ears, and could hear high pitches then. He asked people to hold their hands up if they could hear the pitch. As he went higher and higher, the hands dropped until I was the only one with my hand up. The professor said "the machine is off." The class laughed uproariously, and I was embarrassed, but I heard that pitch plain as day. Someone called out to me "you've got tinnitus." Until that day, I'd never heard that word. Suddenly I learned about the ringing in my ears that never goes away. Right now it's producing an Eb at some extreme octave. So, that day I learned several valuable lesson: I really could hear high pitches, but at some point they were indistinguishable from my tinnitus. If they moved, I could hear them. But if they were static waves, I could NEVER be sure whether I was hearing an actual sound or the ringing in my ears. So, I have instruments that will tell me that, and I have to trust those instruments. We're talking frequencies at the 17K-20K range, so it's not something that impedes my abilities as an audio engineer, and as I said, if the sound is moving, I can tell the difference. But more importantly, I learned that the human mind can be fallible even when we're 100% sure of what we perceived.

SO... before we can talk about anything like percentage of nullification, we MUST agree on the definition of nullification, and I think we have at least gotten that far today.

Shooshie

[Shooshie]

If we ever get quantum computers we will be soooo screwed.

We'll be screwed and not screwed, and everything in between at the same time

Oh gawd! Schroedinger's Mixing Board!



Shoosh

[bayswater]

Would it be worth emphasizing that the null test it merely a simple arithmetic trick to see if two files are identical? If we agree that identical files are identical, and therefore sound the same, the only question is whether the null test actually shows two files are identical. Most first year Algebra students could provide a proof. It's been too long for me.

[Timeline]

No shame in that. It happens to me and to every last one of you. I'm forever astounded by the difference I hear at the end of the day vs. the start of a day. My ears literally grow numb. Come back to a long mix after a good night's sleep, and I want to do it all over again, because I wasn't hearing the glaring things I'm hearing now.

Someone called out to me "you've got tinnitus." Until that day, I'd never heard that word. Suddenly I learned about the ringing in my ears that never goes away. Right now it's producing an Eb at some extreme octave. So, that day I learned several valuable lesson: I really could hear high pitches, but at some point they were indistinguishable from my tinnitus. If they moved, I could hear them. But if they were static waves, I could NEVER be sure whether I was hearing an actual sound or the ringing in my ears. So, I have instruments that will tell me that, and I have to trust those instruments. We're talking frequencies at the 17K-20K range, so it's not something that impedes my abilities as an audio engineer, and as I said, if the sound is moving, I can tell the difference. But more importantly, I learned that the human mind can be fallible even when we're 100% sure of what we perceived.

SO... before we can talk about anything like percentage of nullification, we MUST agree on the definition of nullification, and I think we have at least gotten that far today.

Shooshie

We agree so lets move on Shooshie.

I mix till i'm fatigued then refreshed in the AM do small tweaks to balance the final mix. Works for me.

I have this ringing issue on high blood sugar moments, then i take my meds and it fades away. Pretty high frequency though. Used to be able at 19 to hear in one ear almost to 21k the other 19k. We mixers would test each other in the control room at Sunwest Studios Hollywood where i started.

We do loose most of this ear bandwidth over time. Used to hate the 19k stereo pilot tone from FM broadcasts and store movement alarms drove me crazy. FM just don't bug me any more. I knew an old engineer that worked at Sound City where I worked in the early '70s that could only hear to about 3k and his recordings were excellent. He just put the mic in the same place he learned over his life and hit the record button. Worked every time.

Yes we all get fooled with psycho acoustics from time to time, some more psycho than others, .

You have shown numbers to demonstrate Null. How do you see in a math configuration? What procedure are you using for this?

[Dan Worley]

When I was a freshman in college in the early 1970s, taking physics, the professor was using a wave generator to produce ever higher wavelengths. I had good ears, and could hear high pitches then. He asked people to hold their hands up if they could hear the pitch. As he went higher and higher, the hands dropped until I was the only one with my hand up. The professor said "the machine is off." The class laughed uproariously, and I was embarrassed, but I heard that pitch plain as day. Someone called out to me "you've got tinnitus." Until that day, I'd never heard that word. Suddenly I learned about the ringing in my ears that never goes away. Right now it's producing an Eb at some extreme octave. So, that day I learned several valuable lesson: I really could hear high pitches, but at some point they were indistinguishable from my tinnitus. If they moved, I could hear them. But if they were static waves, I could NEVER be sure whether I was hearing an actual sound or the ringing in my ears. So, I have instruments that will tell me that, and I have to trust those instruments. We're talking frequencies at the 17K-20K range, so it's not something that impedes my abilities as an audio engineer, and as I said, if the sound is moving, I can tell the difference. But more importantly, I learned that the human mind can be fallible even when we're 100% sure of what we perceived.

That's a great point. I was just going to say we can't always trust our ears when doing null tests. If the difference is up in the high frequency range and we can't hear up in that range, or don't have our monitors turned up loud enough, we have to rely on meters and analyzers to tell us what's coming through.

Also, time-based effects can be different with each pass, so even though we can have two mixes that will not null, the mixes can be as close to identical as you can get them.

I'm a null testing fanatic. I think it's a wonderful tool for learning what our tools are doing and for checking equipment and software releases, and our ears and brain to make sure they're not playing tricks on us.

A great way to learn how fine or coarse settings can be is do a 2-mix, then bring it back in and null it out against the multi-track, then change the settings in an EQ or compressor (or whatever) on any of the tracks. It's there when we start to realize how precise all this stuff is, or not.