Null Tests

[KingsXJJ]

If you take a listen to Leo's vids for example, whatever you hear when the NULL test is done (when the signals have opposite phase) is what the profiling is missing. If the signals would be identical, a Null tests would rendered silence.

If the stuff you're hearing is a noticeable part of the frequency spectrum, then accuracy isn't as good and you're most likely going to be able to tell a difference.

If whatever you hear is just static etc then it means the profile is pretty spot on.
Just my 2c.

Thank you @2dor !

 

[KingsXJJ]

Because the residual of a null test only tells you if there is a difference, not if the difference is perceptually audible.

My point and question exactly put.

 

[KingsXJJ]

Does anyone suspect that software/firmware revisions will offer the possibility for improvement on null tests on units converting NAM files? I imagine each product will face some hardware limitations that will vary… but perhaps the conversion process can be tweaked toward a more lossless state? I am also intrigued about what difference a multi-layered approach would have upon null tests… if anything?

 

[paisleywookiee]

Does anyone suspect that software/firmware revisions will offer the possibility for improvement on null tests on units converting NAM files? I imagine each product will face some hardware limitations that will vary… but perhaps the conversion process can be tweaked toward a more lossless state? I am also intrigued about what difference a multi-layered approach would have upon null tests… if anything?

I can’t speak as an expert to NAM specifically, but it’s software, which can be improved; I can’t imagine they couldn’t optimize the code to make it a more 1:1 comparison. Sure, there’s limitations there with CPU and all that, but still.

 

[Orvillain]

@Orvillain @2dor @MirrorProfiles @AlbertA et al. Please bring me up to speed.

 

[Alex Kenivel]

I wonder what would happen if you try a null test between Jay and Andy?

 

[KingsXJJ]

Fuck off then.

 

[metropolis_4]

Null tests are the guitarist equivalent of photographers shooting test panels and pixel-peeping the results

 

[Boudoir Guitar]

Null tests are the guitarist equivalent of photographers shooting test panels and pixel-peeping the results

At least the pixel peeping of test panels is an actual analysis of an image.

 

[KingsXJJ]

Null tests are the guitarist equivalent of photographers shooting test panels and pixel-peeping the results

I, for one, think they are something measurable and interesting to explore. But I’ve always been interested in things like this.

 

[Boudoir Guitar]

I, for one, think they are something measurable and interesting to explore. But I’ve always been interested in things like this.

What would you measure?

 

[KingsXJJ]

What would you measure?

The difference between two signals.

 

[Boudoir Guitar]

The difference between two signals.

How? "difference" isn't a unit of measure.

 

[KingsXJJ]

How? "difference" isn't a unit of measure.

Ok.

 

[Chocol8]

The difference between two signals.

That's sounds like a great idea until you actually try to design and run a few tests...

What do you use as an input signal? How are differences measured and weighted? What meaning can be reliably gleaned from the resulting numbers?

If you actually try to do a few, and do it well with the intention of having a meaningful result, you will quickly realize the resulting numbers can vary tremendously with your choices, and can be what you want it to be if you are willing to be manipulative.

Results from someone else's null tests are meaningless, and doing your own is not much better.

 

[2dor]

That's sounds like a great idea until you actually try to design and run a few tests...

What do you use as an input signal? How are differences measured and weighted? What meaning can be reliably gleaned from the resulting numbers?

If you actually try to do a few, and do it well with the intention of having a meaningful result, you will quickly realize the resulting numbers can vary tremendously with your choices, and can be what you want it to be if you are willing to be manipulative.

Results from someone else's null tests are meaningless, and doing your own is not much better.

TBH the value with NULL tests is when you put different platforms against eachother in a controlled test setup. That's probably the scenario where you can get quantifiable results.
I for one wouldn't have gone down that rabbit hole unless some of the paid profiling tech I was trying a few years back wouldn't have been struggling with some of my favorite amp settings - it was about the same time when NAM came out & was liking what I heard when profiling my gear; got curious to understand what the difference was between NAM & what I had at that time in terms of commercial profiling products.

 

[Chocol8]

TBH the value with NULL tests is when you put different platforms against eachother in a controlled test setup.

There can be some value there, but when I fooled around with it, I could change the results pretty easily. For example, compare Kemper to Tonex using a clean capture and little input volume variation. Repeat the test with medium to high gain, a bunch of palm mutes etc. and significant roll back of the volume knob. The resulting LUFS values will be dramatically different. Which one is right?

 

[2dor]

There can be some value there, but when I fooled around with it, I could change the results pretty easily. For example, compare Kemper to Tonex using a clean capture and little input volume variation. Repeat the test with medium to high gain, a bunch of palm mutes etc. and significant roll back of the volume knob. The resulting LUFS values will be dramatically different. Which one is right?

Of course - and it proves the slippery slope with these things: done right, you can get some good insight; fiddle around with things during the process (deviate from the baseline input signal etc) and you're pretty much in "rubbish IN -> rubbish OUT" territory.

 

[ian_dissonance]

Captures are stupid anyway, so there’s that.

 

[AlbertA]

Captures are stupid anyway, so there’s that.