• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Beta-test: DeltaWave Null Comparison software

OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
A quick update to DeltaWave version 2.0.1:

Changes in 2.0.1
  • Fix: regression issue with color scaling of delta spectrograms
  • Changed: IIR filter replaced with the maximally flat in passband Butterworth version
  • Fix: FIR filter could previously cause a small error near DC frequency
  • Added: additional filter sizes (8M and 16M taps)
 

BeerBear

Active Member
Joined
Mar 9, 2020
Messages
264
Likes
252
I'm new to this and I encountered some errors in the few minutes that I've been using it.
This pops up after the first start of a clean install (after that, it doesn't anymore):
deltawave error.png


And I've also seen a couple of others, like this one:
deltawave error 2.png

I don't remember exactly what I was doing there, but I pretty much just loaded a couple of random files, pressed yes/no on some dialog and then switched between the tabs.
EDIT: OK, I can reproduce this. Indeed, as Blumlein 88 hinted at, the files are very different. They're just over a minute long (a few seconds apart) and one is much more heavily compressed (dynamically) than the other. And the error happens if I select 'No' on the "alternative drift correlation method" dialog.


But ignoring these errors, it seems to work fine. Thanks for making it!
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
I'm new to this and I encountered some errors in the few minutes that I've been using it.
This pops up after the first start of a clean install (after that, it doesn't anymore):
View attachment 170318

And I've also seen a couple of others, like this one:
View attachment 170319
(I don't remember exactly what I was doing there, but I pretty much just loaded a couple of random files, pressed yes/no on some dialog and then switched between the tabs.)

But ignoring these errors, it seems to work fine. Thanks for making it!
Paul can give you a better answer, but he said he was going to be away from computer access for a few days.

I've seen that sort of thing in two instances. One if comparing two files that are rather very different. Like maybe a recording of a song from an LP and same song from a CD. The variable timing of the LP can cause headaches. The other instance is using very large (long) files to compare. I think Paul says Dwave works best with files of 2-3 minutes.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
I'm new to this and I encountered some errors in the few minutes that I've been using it.
This pops up after the first start of a clean install (after that, it doesn't anymore):
View attachment 170318

And I've also seen a couple of others, like this one:
View attachment 170319
I don't remember exactly what I was doing there, but I pretty much just loaded a couple of random files, pressed yes/no on some dialog and then switched between the tabs.
EDIT: OK, I can reproduce this. Indeed, as Blumlein 88 hinted at, the files are very different. They're just over a minute long (a few seconds apart) and one is much more heavily compressed (dynamically) than the other. And the error happens if I select 'No' on the "alternative drift correlation method" dialog.


But ignoring these errors, it seems to work fine. Thanks for making it!

Hi BeerBear,

Thanks for reporting this. The first error seems to have to do with file permissions. This is strange, as this error should be caught before a message was displayed. It will only happen the first time you run the app. It's completely harmless and happens because it can't file any preferences saved from previous runs of the app and will instead, use default settings. I'll fix it, as it doesn't make for a very good experience for first time users.

The second one does happen more when the files are very mismatched. There are tons analysis routines that the files go through during a normal comparison, and while I try to prevent errors like this, sometimes computations go awry and exceed the normal/expected limits. This one is actually just the display routine showing the histogram of error values, and apparently due to the large range of errors the computation exceeded the range I allocated. Again, not fatal since only that one chart is affected, but I'll fix this one in the next update.

Glad you're able to use DW and finding it useful!
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
DeltaWave v2.0.2 is available for download. The primary change is the added support for reading/processing WAVEX files (extended WAV file format from Microsoft). While these are multi-channel files, DeltaWave will allow processing of only the first two channels, usually left and right.

Changes in 2.0.2

  • Added: support for reading WAVEX (extended WAVE format) files
 

manisandher

Addicted to Fun and Learning
Joined
Nov 6, 2016
Messages
652
Likes
607
Location
Royal Leamington Spa, UK
@pkane

Hi Paul, I've recently been using Deltawave to make comparisons of a few DA_AD loopbacks that I've taken, and have come across something that's baffling me. I'm hoping you'll be able to shed some light onto what might be happening.

I'm using an RME ADI-2 Pro as the loopback device. The test file is a 6 second single guitar pluck that I recorded with a mic:

Guitar Pluck Waveform.jpg
Guitar Pluck Spectrogram.jpg

I've taken 10 loopbacks in succession, with absolutely no changes in SW or HW. Using Deltawave, I've compared the first to the subsequent nine. What I'm seeing is a pattern showing up over time in the rms difference and the PK Metric. I've made a little mpeg of how the PK Metric changes from 1 vs. 2 to 1 vs. 10, which you can download here: https://drive.google.com/file/d/1ZIxZCGEUYBRO7B0clC-JiSCGQ8YVOphx/view?usp=sharing.

My first thought was that it might be clock drift, but adding correction makes no difference to the results.

Any idea what could be causing this? And if so, any idea as to what I could do to remedy it?

Thanks,
Mani.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
@pkane

Hi Paul, I've recently been using Deltawave to make comparisons of a few DA_AD loopbacks that I've taken, and have come across something that's baffling me. I'm hoping you'll be able to shed some light onto what might be happening.

I'm using an RME ADI-2 Pro as the loopback device. The test file is a 6 second single guitar pluck that I recorded with a mic:

View attachment 177247
View attachment 177249

I've taken 10 loopbacks in succession, with absolutely no changes in SW or HW. Using Deltawave, I've compared the first to the subsequent nine. What I'm seeing is a pattern showing up over time in the rms difference and the PK Metric. I've made a little mpeg of how the PK Metric changes from 1 vs. 2 to 1 vs. 10, which you can download here: https://drive.google.com/file/d/1ZIxZCGEUYBRO7B0clC-JiSCGQ8YVOphx/view?usp=sharing.

My first thought was that it might be clock drift, but adding correction makes no difference to the results.

Any idea what could be causing this? And if so, any idea as to what I could do to remedy it?

Thanks,
Mani.

Hi Mani,

Not sure exactly what I'm looking at, but a few comments on what I do see. First, the recording appears to be very short (5-6 seconds?) which leaves no room for DeltaWave to compute clock drift, if any. Record 30 seconds, a minute is better if you suspect unsynched clocks.

Second, the video shows that PK Metric is finding an increasingly larger difference between recordings, but only at the start of the recording. At the start, it appears the level is fairly close to 0dBFS, so it is possible that some of the recordings were too loud, generating large distortion due to clipping at the start. Try adjusting RME ADC gain to lower this.
 

manisandher

Addicted to Fun and Learning
Joined
Nov 6, 2016
Messages
652
Likes
607
Location
Royal Leamington Spa, UK
Hi Mani,

Not sure exactly what I'm looking at, but a few comments on what I do see. First, the recording appears to be very short (5-6 seconds?) which leaves no room for DeltaWave to compute clock drift, if any. Record 30 seconds, a minute is better if you suspect unsynched clocks.

Second, the video shows that PK Metric is finding an increasingly larger difference between recordings, but only at the start of the recording. At the start, it appears the level is fairly close to 0dBFS, so it is possible that some of the recordings were too loud, generating large distortion due to clipping at the start. Try adjusting RME ADC gain to lower this.

Thanks Paul. I'll look into things further.
 

Tj99

Member
Joined
Dec 16, 2020
Messages
16
Likes
8
@pkane Thank you very much for this wonderful piece of software, directly donated.

I just have 2 questions:

1) Is there a way to see the applied EQ+Phase correction, when activating the option in the "Non-linear Calibration" section?
2) When trying to repeat the measurements on https://deltaw.org/gearslutz.html with the original files, I get much higher PK metric values.
I tested RME 9632 = -80,5 dBFS (instead of -106 dBFS) and Motu 828es = -88,4 dBFS (instead of -105,9 dBFS). Did there something change in the calculation, or is it a wrong setting on my side? I matched the settings 1:1.

Cheers!
 

Attachments

  • MOTU PK.png
    MOTU PK.png
    536.4 KB · Views: 72
  • RME PK metric.png
    RME PK metric.png
    473.5 KB · Views: 65
  • Settings DW.png
    Settings DW.png
    221.4 KB · Views: 68
Last edited:
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
@pkane Thank you very much for this wonderful piece of software, directly donated.

I just have 2 questions:

1) Is there a way to see the applied EQ+Phase correction, when activating the option in the "Non-linear Calibration" section?
2) When trying to repeat the measurements on https://deltaw.org/gearslutz.html with the original files, I get much higher PK metric values.
I tested RME 9632 = -80,5 dBFS (instead of -106 dBFS) and Motu 828es = -88,4 dBFS (instead of -105,9 dBFS). Did there something change in the calculation, or is it a wrong setting on my side? I matched the settings 1:1.

Cheers!
Thank you, TJ99!

Will take a look when I get back to my computer :) There were some changes made in how the non-linear EQ is applied to make it less sensitive to non-repeating errors, so that may account for some of the difference. PK Metric calculation hasn't changed.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
@pkane Thank you very much for this wonderful piece of software, directly donated.

I just have 2 questions:

1) Is there a way to see the applied EQ+Phase correction, when activating the option in the "Non-linear Calibration" section?
2) When trying to repeat the measurements on https://deltaw.org/gearslutz.html with the original files, I get much higher PK metric values.
I tested RME 9632 = -80,5 dBFS (instead of -106 dBFS) and Motu 828es = -88,4 dBFS (instead of -105,9 dBFS). Did there something change in the calculation, or is it a wrong setting on my side? I matched the settings 1:1.

Cheers!

Hi TJ,

Looked at RME HDSP 9632, since I had that file. My comments:

1. You should use just Left or Right channels for comparison, as adding them (L+R) will compound some of the errors/distortions

2. It looks like I removed a limit on how small a drift can be detected, and in the process, made it possible for some noise in the recording to affect the drift measurement. While the detected drift is tiny (0.001ppm) it still affects negatively the overall error and increases PK Metric. I'll see if I can handle this better in a future update.

EDIT:
3. I also made a change recently to stop DW from subtracting DC offset when that option was unchecked. Previously, DW could ignore that option if the result was a lower error value, which meant that DC offset may still be removed even if that option is not selected. Now, the option is honored in all cases. This causes a larger error with this file, since there is a DC offset. The two changes are: don't correct clock drift, and remove DC offset.

For now since these recordings were all done with a synchronized clock, there can be no drift. So it's safe to turn off drift correction in DW:

1641942879878.png


After making these changes, the result is much closer to the one I originally posted:
1641941743886.png
 
Last edited:

Tj99

Member
Joined
Dec 16, 2020
Messages
16
Likes
8
Hi TJ,

Looked at RME HDSP 9632, since I had that file. My comments:

1. You should use just Left or Right channels for comparison, as adding them (L+R) will compound some of the errors/distortions

2. It looks like I removed a limit on how small a drift can be detected, and in the process, made it possible for some noise in the recording to affect the drift measurement. While the detected drift is tiny (0.001ppm) it still affects negatively the overall error and increases PK Metric. I'll see if I can handle this better in a future update.

EDIT:
3. I also made a change recently to stop DW from subtracting DC offset when that option was unchecked. Previously, DW could ignore that option if the result was a lower error value, which meant that DC offset may still be removed even if that option is not selected. Now, the option is honored in all cases. This causes a larger error with this file, since there is a DC offset. The two changes are: don't correct clock drift, and remove DC offset.

For now since these recordings were all done with a synchronized clock, there can be no drift. So it's safe to turn off drift correction in DW:

View attachment 178333

After making these changes, the result is much closer to the one I originally posted:
View attachment 178324
Thank you very much for the detailed answer!

And about the applied EQ+Phase correction? Is there (or will there be) a way to show this?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
Thank you very much for the detailed answer!

And about the applied EQ+Phase correction? Is there (or will there be) a way to show this?

I'm not sure how you'd want to see it. The results can certainly be seen. For the same RME loopback, with phase and level EQ:

1641990219327.png

1641990251435.png


If you turn on Impulse Response recording before doing a Match, then you can also see the IR that was used for phase/level EQ correction:

1641990537830.png
 

BeerBear

Active Member
Joined
Mar 9, 2020
Messages
264
Likes
252
A request: can you add WASAPI shared mode for the audio output? Exclusive mode is annoying most of the time and--except in some rare cases--I'm not really concerned about Windows' resampling quality.

(Alternatively: resample all the audio from DeltaWave to some user defined value. This would at least let us use ASIO without conflicts with other apps. But adding WASAPI shared is probably easier and would solve the problem for more people.)
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
A request: can you add WASAPI shared mode for the audio output? Exclusive mode is annoying most of the time and--except in some rare cases--I'm not really concerned about Windows' resampling quality.

(Alternatively: resample all the audio from DeltaWave to some user defined value. This would at least let us use ASIO without conflicts with other apps. But adding WASAPI shared is probably easier and would solve the problem for more people.)

Everything's possible :) Auto-resample for playback option is available which will use Windows built-in resampler, or you can use DeltaWave higher-quality resampler to change the sampling rate of the waveform to a specific rate from the drop-down:

1646917084822.png
 
  • Like
Reactions: trl

BeerBear

Active Member
Joined
Mar 9, 2020
Messages
264
Likes
252
D'oh! I missed those settings. However...

The "Auto-resample for audio playback" setting doesn't work for me, because
1. it only works with WASAPI exclusive mode and
2. it only resamples if the driver doesn't support the sample rate of the files. And DeltaWave can dictate the sample rate to the driver.
DAWs have a user-selectable sample rate setting and the audio is resampled on the fly just for playback. That's what I had in mind with the alternative idea.

The General->Resample option does solve my problem. But it's a rather clumsy solution, which changes the files. If I'm comparing two 44.1k files, I would prefer to keep them at that rate, also for visual reasons, instead of converting them to, say, 96k of my sound card (which also increases processing time, I think).
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,629
Likes
10,202
Location
North-East
By request, a few new enhancements are now available in DeltaWave v2.0.3:
  • Added: WASAPI Exclusive/Shared setting (@BeerBear )
  • Added: % match to the output of the batch-file mode (command line execution)
The WASAPI Exclusive/Shared mode setting is added but I've not been able to test it here, so hopefully it works :) Just uncheck the Exclusive Mode option to use Shared:

1647115443969.png


The command line enhancement is to add % match to the output when running DeltaWave in batch mode, using command-line arguments. You can start DeltaWave from command line and have it output results appended to a text file. The way to run it is as follows:

Bash:
 C:\Bin> DeltaWave "z:\test\reference.wav" "z:\test\comparison.wav" "z:\test\output.txt"


The first parameter is the reference file, the second parameter is the comparison. Third is output file that the results will be written to (appended). There is an optional fourth parameter that can be used to specify a DeltaWave settings file to be used. If not specified, last saved settings in DeltaWave will be used for analysis.

While performing the comparison, you will see the DeltaWave window pop-up, but it will go away when the analysis is finished. In this way, you can run multiple comparisons by executing DeltaWave multiple times, with different parameters. This can also be done from a batch file or a script.

Here's an example output from a run (all output values are tab-separated):

Code:
c:\wavfiles\DA_AD_test.wav    Gain(dB)=0.00101504194079131    ClockDrift=-4.75211887012661E-10    RMS(dBFS)=-53.5933309443475    RMS(dBA)=-68.9271096211634    RMSEQ(dBFS)=-77.3854156368189    RMSEQ(dBA)=-78.2824387470902    Linearity=0    Jitter=9.92105896272957E-08    Phase10k=0.128405279833697    PKMetric=-81.3    DFMetric=0    PercentMatch=0.018415384100269
 
Last edited:

dadregga

Active Member
Joined
May 27, 2021
Messages
154
Likes
340
Just now finding this tool, which is considerably more useful at this specific task than Audacity - thanks much @pkane !
 

BeerBear

Active Member
Joined
Mar 9, 2020
Messages
264
Likes
252
This is maybe a more general question, but I'm trying to make sense of the various null depth figures.

I've been used to loading files in a DAW, inverting the phase of one of the two and playing them together.

If I do this for a couple of 24bit files I have here, I get a signal at -138dB (the only difference between the two is dither).
But DeltaWave says the null depth is 120dB. And the correlated null depth is 193dB.
What's the explanation for that difference? And I guess correlated null depth measures something else... but what?
 
Top Bottom