DeltaWave null-testing audio comparator (beta)

[pkane2001]

4 hours ago, fas42 said:

In 54b, with DiffSpec range of say -6 to 0, should the exact same picture be shown as for 53b?

 

Yes, nothing's changed except how the min/max range is set. In 53 it was separate from Spectrogram 1 and 2, in 54 it is the same. But, if you set it to the same range in 54 as it was in 53, the display should match 53.

[fas42]

6 minutes ago, pkane2001 said:

 

Yes, nothing's changed except how the min/max range is set. In 53 it was separate from Spectrogram 1 and 2, in 54 it is the same. But, if you set it to the same range in 54 as it was in 53, the display should match 53.

 

But you can't set it the same - positive dB values aren't available, any longer. And I did try it with a gray scale, the a particular comparison, between 53 and 54 ... very different images.

[pkane2001]

6 minutes ago, fas42 said:

 

But you can't set it the same - positive dB values aren't available, any longer. And I did try it with a gray scale, the a particular comparison, between 53 and 54 ... very different images.

 

Oh! The scale is no longer the same. OK, that was another reason for keeping them unlocked and separate :) 

[pkane2001]

8 minutes ago, pkane2001 said:

 

Oh! The scale is no longer the same. OK, that was another reason for keeping them unlocked and separate :) 

 

On a positive note, spectrogram FFT size setting now allows much larger sizes to be selected. Unfortunately, a large FFT will result in a slower display, and also use a whole lot more memory. But, it will result in much finer frequency resolution. So, if you have enough memory on your PC, it's something to try :) Having three spectrogram tabs enabled will use up 3x the memory, so if you only chose one, you're less likely to run out.

 

 

[fas42]

Silly question 😉 ... why can't be the spectrogram windows be (un)locked, but the delta window have its own scale ranges?

[pkane2001]

Just now, fas42 said:

Silly question 😉 ... why can't be the spectrogram windows be (un)locked, but the delta window have its own scale ranges?

 

Because that's how it's designed :) All settings for the window, including range and scale are associated with the window type. All the windows of the same type share all the settings, not just some of them.

[fas42]

I think an executive decision, on design, is called for  ... 😜.

[pkane2001]

Just now, fas42 said:

I think an executive decision, on design, is called for  ... 😜.

 

I'll schedule a meeting to plan for a meeting to review your proposal 😂

[fas42]

Right, as a bandleader once said, One More Time ...

 

I created a test track of white noise, and on a dup of that mixed in, in the middle section, a fade in and fade out of a 3kHz sine wave, which peaked at a much higher level than the peak of the white noise - "sticks out like a sore thumb". And ran a compare of the two.

 

If I make the noise+plus sine clip the Reference the sore thumb looks like a toothpick of black in a sea of white, in the Delta Spectrogram. But if I swap the Reference and Compare, the toothpick disappears - just a tiny hint of a gray line to mark its presence.

 

IOW, the Delta can't show the detail that should be there, because of the limitations of the scale range.

[pkane2001]

2 hours ago, fas42 said:

Right, as a bandleader once said, One More Time ...

 

I created a test track of white noise, and on a dup of that mixed in, in the middle section, a fade in and fade out of a 3kHz sine wave, which peaked at a much higher level than the peak of the white noise - "sticks out like a sore thumb". And ran a compare of the two.

 

If I make the noise+plus sine clip the Reference the sore thumb looks like a toothpick of black in a sea of white, in the Delta Spectrogram. But if I swap the Reference and Compare, the toothpick disappears - just a tiny hint of a gray line to mark its presence.

 

IOW, the Delta can't show the detail that should be there, because of the limitations of the scale range.

 

Frank, Frank... We had the emergency board meeting to discuss your proposed change, and I'm afraid that it was unanimously voted down.

 

Despite this, I decided to go against the board decision and made the change anyway, considering it's you, Frank, that asked. Please try out the new version, but please don't tell anybody as I'll get in trouble with the board! 😎

[fas42]

Arghh! What do boards know, anyway ...

 

Though, sneaky fellow, it appears you have resolved this by reversing the locking group change - so, we're back to 53b, eh? 😉

[pkane2001]

2 hours ago, fas42 said:

Arghh! What do boards know, anyway ...

 

Though, sneaky fellow, it appears you have resolved this by reversing the locking group change - so, we're back to 53b, eh? 😉

It’ll need to be a feature for another release. Too many changes — the board was right ;)

[pkane2001]

18 hours ago, fas42 said:

Arghh! What do boards know, anyway ...

 

Though, sneaky fellow, it appears you have resolved this by reversing the locking group change - so, we're back to 53b, eh? 😉

 

Frank, please try to re-download and reinstall v1.0.55. I made some more adjustments to the spectrogram locking behavior. Not fully 100% perfect, but I think it should work for what you're trying to do.

 

[fas42]

Quick look ... yep, seems fine - thanks! 👍

[fas42]

Okay, clock drift ... if I disable Correct phase drift, I still get Raw drift and Corrected drift plots in the the Clock Drift window - and they differ. Does this mean that adjustments were made, or not?

 

If I'm certain that the clocks are 100% correct, for both files, can I have it be guaranteed that no attempt at drift compensation happens?

 

Thanks, Paul 🙂

[pkane2001]

9 minutes ago, fas42 said:

Okay, clock drift ... if I disable Correct phase drift, I still get Raw drift and Corrected drift plots in the the Clock Drift window - and they differ. Does this mean that adjustments were made, or not?

 

If I'm certain that the clocks are 100% correct, for both files, can I have it be guaranteed that no attempt at drift compensation happens?

 

Thanks, Paul 🙂

 

No drift correction is applied if Correct Phase Drift is turned off (but also make sure the non-linear phase EQ is turned off as well).

 

What's the scale of the differences? Very small differences could be due to computational errors of values that are below the precision of what DW can compute.

[fas42]

Could be the logic controlling updating of the plots - if I do a start of DW, and immediately disable Correct phase drift, and then match, no data appears in the Clock Drift window. Doing subsequent matches with the Correct enabled, and then disabled doesn't clear the Clock Drift window data, from the prior match - which is confusing.

[pkane2001]

Just now, fas42 said:

Could be the logic controlling updating of the plots - if I do a start of DW, and immediately disable Correct phase drift, and then match, no data appears in the Clock Drift window. Doing subsequent matches with the Correct enabled, and then disabled doesn't clear the Clock Drift window data, from the prior match - which is confusing.

 

I'll have to take a look to see why this might happen, that's not the way it should work.

[pkane2001]

Version 1.0.56 of DeltaWave is now available with a couple of bug fixes:

• Fix: Drift correction could result in a decrease in higher frequencies in the spectrum plot
• Fix: Turning off drift correction would not clear the previously computed data in the Clock Drift plot

 

[fas42]

Paul, may I suggest a little refinement with how the locking works, please. If I have a single window unlocked, and the rest locked - and I zoom into some part of a locked window; and then go back that unlocked window it hasn't moved of course. So, I now lock it - and it snaps to the same time window as the rest - exactly as I want. But I now want to explore that now locked window without disturbing the others, so, I unlock it to do that ... which loses the alignment - it jumps back to the previous view, at the time I changed its status from unlocked to locked! ... So, I've got nowhere - with trying to synch the windows in one dimension; I have to do a full manual zoom to try and match the time axis.

 

Thanks!

[pkane2001]

On 1/25/2021 at 8:08 PM, fas42 said:

Paul, may I suggest a little refinement with how the locking works, please. If I have a single window unlocked, and the rest locked - and I zoom into some part of a locked window; and then go back that unlocked window it hasn't moved of course. So, I now lock it - and it snaps to the same time window as the rest - exactly as I want. But I now want to explore that now locked window without disturbing the others, so, I unlock it to do that ... which loses the alignment - it jumps back to the previous view, at the time I changed its status from unlocked to locked! ... So, I've got nowhere - with trying to synch the windows in one dimension; I have to do a full manual zoom to try and match the time axis.

 

Thanks!

 

That's a feature, Frank :) This way you have two zoom levels to switch between easily.

 

Here's a way for you to do what you want: when in the "locked" mode, click on the manual Zoom Settings icon next to the lock:

 

 

This will display the manual zoom level window. Keep this window floating, then unlock by clicking the lock icon. The zoom level will change. But now you can click on the Apply! button in Zoom Settings window, and the plot will reset to the same settings as in the "locked" zoom view.

[pkane2001]

New version of DeltaWave is available: v1.0.57

Changes in this version:

• Added: ERB trend-line option for FFT plots in settings
• Added: PK Error Metric and plot
• Changed: phase unwrap behavior and algorithm

 

PK Error Metric is a beta version of a new audibility (perception-weighted) metric meant to supplement the more engineering-focused RMS error metric. Here's an example:

 

Discussion of PK Metric on ASR.

 

ERB-smoothed trend-lines for frequency plots can be enabled in Settings:


The trend-line looks like this:

 

 

[rando]

Very interesting.  Time to dig out the 1st-8th gen test files and see what else I have hanging around in the near vicinity.

[fas42]

2 hours ago, pkane2001 said:

 

That's a feature, Frank :) This way you have two zoom levels to switch between easily.

 

Here's a way for you to do what you want: when in the "locked" mode, click on the manual Zoom Settings icon next to the lock:

 

 

 

Got it!! ... Thanks!

[fas42]

Naughty Paul ... not keeping us up to date on latest offerings ... 😈

 

Has got himself a fan base on that forum that shall not be named - what's the world coming to !? 🤪

 

 

But seriously, lots of good input from members there - and DW keeps getting better, 😉.