- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
Changing from IIR to FIR didn't make a difference. Changing from Hann to FIR Kaiser did not make a difference
-
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
- Thread Starter
- #422
No, you can use IIR or FIR. I was more concerned that somehow something corrupted filter size or Window type.
- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
If both filter one and two are on it works as it should. Filter one doesn't work with filter two off, and any value showing other than (no limit).
OP
- Thread Starter
- #424
Hmm. Strange. I can't get it to do the same. Notch filter is also turned on in your screen shot, so I did the same:
OP
- Thread Starter
- #425
Aha! The second filter has to be set @start only, and with a valid frequency, and turned off. Then I can reproduce it. OK, it has to be that specific combination of settings
I'll fix this.
- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
Yes, it does work if filter two is @end, I hadn't tried that combination.
OP
- Thread Starter
- #427
Updated DeltaWave version 1.0.54 is now available.
Changes in 1.0.54b
To activate FFT Window explorer, click on the question mark in settings, under Spectrum:
You can explore and compare other windows once there:
Changes in 1.0.54b
- Added: new FFT window types, including multiple Kaiser
- Added: FFT Window explorer with time- and frequency-domain plots and measurements
- Changed: Delta Spectrogram window scale is now locked to Spectrogram 1 and 2
- Fixed: under one specific combination of filter settings, filter 1 @start was being ignored
To activate FFT Window explorer, click on the question mark in settings, under Spectrum:
You can explore and compare other windows once there:
- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
All the FFT choices are nice edition and doubly so that you show the parameters of them when choosing.
Thanks Paul.
- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
This isn't a bug. The software functions properly, but could be confusing to the user.
If you set a notch filter at a frequency above the max possible for the file, the program stops and displays empty boxes. Unless you look at the results page where it says index was outside the bounds of the array you will not know why. For instance I compared two 44.1 khz files with a notch filter accidentally left on at 26.7 khz. Maybe if the notch filter box turned yellow or something when you do this.
You can set a HP or LP filter at too high a frequency and the program completes. If it is an LP filter then nothing is effected. If it is a HP filter the software completes and gives values, but the graphs are all wrong because you've filtered everything out. Either of those might warrant a yellow box on the filter or some sort of warning.
If you set a notch filter at a frequency above the max possible for the file, the program stops and displays empty boxes. Unless you look at the results page where it says index was outside the bounds of the array you will not know why. For instance I compared two 44.1 khz files with a notch filter accidentally left on at 26.7 khz. Maybe if the notch filter box turned yellow or something when you do this.
You can set a HP or LP filter at too high a frequency and the program completes. If it is an LP filter then nothing is effected. If it is a HP filter the software completes and gives values, but the graphs are all wrong because you've filtered everything out. Either of those might warrant a yellow box on the filter or some sort of warning.
I-Love-Null-Tests
Member
- Joined
- Oct 26, 2020
- Messages
- 7
- Likes
- 2
The results I have been getting with DeltaWave seems to be somewhat random.
I have been testing a fully digital loop. I made three different recordings in a row, of the same test tone, where the only variable between the recordings was the randomness of the triangular dither.
One of the recordings measured 144dB of difference, 220 dB of correlated null and 4ps of jitter.
The second measured 72dB of difference, 95 dB of correlated null and 1.2 µs (!!) of jitter.
The third was somewhere inbetween.
All three recordings sounds exactly the same to my ear and the diffmaker calculated corr depth around 190dB on all three.
I think the algorithm of DeltaWave gets confused with very complex test tones - the more complex and chaotic the test tone, the more obviously wrong measurements I'm getting.
I have been testing a fully digital loop. I made three different recordings in a row, of the same test tone, where the only variable between the recordings was the randomness of the triangular dither.
One of the recordings measured 144dB of difference, 220 dB of correlated null and 4ps of jitter.
The second measured 72dB of difference, 95 dB of correlated null and 1.2 µs (!!) of jitter.
The third was somewhere inbetween.
All three recordings sounds exactly the same to my ear and the diffmaker calculated corr depth around 190dB on all three.
I think the algorithm of DeltaWave gets confused with very complex test tones - the more complex and chaotic the test tone, the more obviously wrong measurements I'm getting.
Last edited:
OP
- Thread Starter
- #431
It’s not the chaotic nature of the test tones but usually periodic test signal or too short a sample that’s the culprit. White noise samples work just fine, for example.
I-Love-Null-Tests
Member
- Joined
- Oct 26, 2020
- Messages
- 7
- Likes
- 2
How short sample is too short?
OP
- Thread Starter
- #433
30 seconds is what I usually use at a minimum. DeltaWave tries to compute clock drift and that requires a number of samples of FFT size that are spaced out to measure clock differences over the length of the track.
I-Love-Null-Tests
Member
- Joined
- Oct 26, 2020
- Messages
- 7
- Likes
- 2
Yes, that's about the length I was using - 30-50s. What would you say is the optimal length for the most accurate results?
OP
- Thread Starter
- #435
30-50s is plenty. If you are using nonlinear-EQ settings, then I'd say at least a minute to two minutes. Non-linear EQ benefits from more averaging over a larger number of samples.
OP
- Thread Starter
- #436
Here's an example of white noise with and without triangular dither. Generated in Audacity, with dither applied due to the conversion from 32-bit to 16-bit WAV files. -96dBFS RMS is about what you might expect with 16-bit file:
And here's 24-bit triangular dithered file. Same white noise as before:
Again, -145dB or so is what you'd expect with 24 bit file.
Oh, and the original spectrum of the file:
I was curious, so here's Audacity Shaped dither with 16bit file. Result is over the entire 48kHz spectrum of the file, but looks really good in the audible range:
Last edited:
111MilesToGo
Member
Jumped the wagon yesterday. Wow - I can say I am deeply impressed by DeltaWave, great ideas and great implementation, dear @pkane! I had tried to use Audio DiffMaker, but ultimately I did fail, as presumably did many others. DeltaWave does so much more, and it proved to be stable up to now. I am pretty sure DeltaWave is going to be THE go-to reference software for such purposes.
Let me just start out with a minor "housekeeping" bug report: On my Windows 8.1 (don't know about W10), the installer places the software into "C:\Program Files (x86)\PKAudio", no matter what. Of course, the 64-bit app should end up in "C:\Program Files\PKAudio". The installer doesn't pick up my changes for the installation directory. Task Manager of course shows that the app is 64-bits.
So I embarked on a long-standing project of mine, the obnoxious audio watermarking of files purchased / downloaded from various web-based (HiRes and CD quality) shops. Wow again, DeltaWave worked like a breeze! My current results are really saddening. Personally, I consider watermarking an insult (to say the least) on the artists, the original sound engineers, and last but not least us customers, many of whom buy the same music over and over.
This forum nicely can display screen shots. If some of you folks want it, and if the moderators don't object, I will post screen shots with my watermarking analysis on this thread here. I think DeltaWave shows the watermarking in an astoundingly clear manner. Of course, if watermarking has been detected and analyzed sufficiently elsewhere, I would like to receive some links to such publications and be quiet on it. I am aware of https://www.mattmontag.com/music/universals-audible-watermark and the links therein. The HiRes purchases I looked at were done only about a year ago, they are ECM Records releases, who are an independent company, but rely on Universal Music Group for distribution - too bad. UMG seems to continue watermarking to this day, at least until a year ago.
Let me just start out with a minor "housekeeping" bug report: On my Windows 8.1 (don't know about W10), the installer places the software into "C:\Program Files (x86)\PKAudio", no matter what. Of course, the 64-bit app should end up in "C:\Program Files\PKAudio". The installer doesn't pick up my changes for the installation directory. Task Manager of course shows that the app is 64-bits.
So I embarked on a long-standing project of mine, the obnoxious audio watermarking of files purchased / downloaded from various web-based (HiRes and CD quality) shops. Wow again, DeltaWave worked like a breeze! My current results are really saddening. Personally, I consider watermarking an insult (to say the least) on the artists, the original sound engineers, and last but not least us customers, many of whom buy the same music over and over.
This forum nicely can display screen shots. If some of you folks want it, and if the moderators don't object, I will post screen shots with my watermarking analysis on this thread here. I think DeltaWave shows the watermarking in an astoundingly clear manner. Of course, if watermarking has been detected and analyzed sufficiently elsewhere, I would like to receive some links to such publications and be quiet on it. I am aware of https://www.mattmontag.com/music/universals-audible-watermark and the links therein. The HiRes purchases I looked at were done only about a year ago, they are ECM Records releases, who are an independent company, but rely on Universal Music Group for distribution - too bad. UMG seems to continue watermarking to this day, at least until a year ago.
Last edited:
- Joined
- Feb 23, 2016
- Messages
- 22,962
- Likes
- 42,618
Welcome to ASR.
Lots of good members here. Pkane is one of the best. He put together Deltawave after seeing/hearing about our complaints of Diffmaker being a good idea, but unusable as it was so unstable. Then he kept adding features as people requested them. Deltawave is quite something now with so many uses.
Last edited:
OP
- Thread Starter
- #439
Hi @111MilesToGo,
Glad to hear that you're enjoying DeltaWave!
Location of the app under Windows is not that important. Most apps just pick one or the other location. Since the app originally was both, 64- and 32-bit, I kept the original location where it was installed.
There were a few threads here on ASR on watermarking, some using DeltaWave. If you use the search function, I bet you'll find these of interest!
111MilesToGo
Member
Update wrt watermarking: Please refer to https://www.mattmontag.com/music/an-update-on-umg-watermarks. Good news, but it still is not known definitely how many of the spoiled UMG releases on hi-res shops and streaming services have been cleaned up by now.
Similar threads
