• 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!

Audibility thresholds of amp and DAC measurements

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
Something like this may be very audible (the 8th order all-pass filter from @xr100):

View attachment 46763

FYI the (wrapped) phase response of the 8th order all-pass filter as measured in VST Analyser:

ASR37.png


And, whilst I'm at it, the group delay:

ASR38.png
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
Did it occur to you that guy accused me for something I haven't remotely done? Not to mention he's not a Mod so it's definitely not his job to do policing on the forum.

And now my tone is more problematic than those 2 simple facts? You have interesting sense of justice.
No one used the language you used. And yes, the tone is super important. This is a professional forum and not some place where you speak that way.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
Curious...

@j_j : I would expect phase-shift detection, outside of extreme examples, to correlate to time shifts that can cause "smearing" and similar audible artifacts. That is, I would expect the amount of phase shift that is detectable to vary with frequency. Is that true? And if so, would a better metric be time shift (delay, whatever)? Or is the detectable phase shift unrelated to time (e.g. group delay -- does non-constant group delay matter)?

Chances are I'm oversimplifying, natch.

I know the answer for a radar or ultrasound system, not sure about audio and our ears (brain, etc.)

There are two effects, one is time smear, the other one is changes in the signal envelope when there is more than one frequency component in an ERB, which almost is always true in a real audio signal.

The combination is quite complicated. Above 2kHz, the sensitivity to waveform timing is mostly gone, but envelope sensitivity is there until you're past your ability to hear that frequency range. Below 500Hz, time arrival of the waveform itself is quite sensitive. Below 90 Hz or so, the actual arrival time is much more exaggerated due to the period of the signal, BUT your sensitivity goes to pot there, too, because the "leading edge" on the basilar membrane is too slow.

The question is not simply answered, but 15 degrees/ERB is a decent estimate.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
... but the result (according to df-metric) is very similar with this m-signal. Can you tell them apart by listening? In df-metric there is no need to look inside a DUT as it is a black box.
I would hope two approaches to least mean squares weren't TOOO different!
 
Last edited:

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
Oh, and some advice. "Phase shift" in the sense I used it, means FIRST you match the delay part of the signal (across linear frequency) and back it out with a linear fit. The difference is the actual phase (the slope of the line is pure delay). And the "difference" is to adjacent ERB's, not peak vs. minimum across all frequencies.

I'm sure there is also a peak vs. minimum factor, but I suspect that's much larger, and totally interpretable as differential time delay.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
There are many ways to match a system, and back out both linear and nonlinear parts. Hopefully most of them get about the same results, because they really should do so.

Of course, I should also say that dropping a distortion product at a frequency far removed from the signal spectrum will drive the ear batty, but not necessarily show up as substantial distortion.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
There are two effects, one is time smear, the other one is changes in the signal envelope when there is more than one frequency component in an ERB, which almost is always true in a real audio signal.

The combination is quite complicated. Above 2kHz, the sensitivity to waveform timing is mostly gone, but envelope sensitivity is there until you're past your ability to hear that frequency range. Below 500Hz, time arrival of the waveform itself is quite sensitive. Below 90 Hz or so, the actual arrival time is much more exaggerated due to the period of the signal, BUT your sensitivity goes to pot there, too, because the "leading edge" on the basilar membrane is too slow.

The question is not simply answered, but 15 degrees/ERB is a decent estimate.

Thanks JJ!
 

Serge Smirnoff

Active Member
Joined
Dec 7, 2019
Messages
240
Likes
136
Trial 1, user: A actual: A 1/1
Trial 2, user: B actual: B 2/2
Trial 3, user: B actual: B 3/3
Trial 4, user: A actual: A 4/4
Trial 5, user: B actual: A 4/5
Trial 6, user: B actual: B 5/6
Trial 7, user: B actual: B 6/7
Trial 8, user: A actual: A 7/8
Trial 9, user: B actual: B 8/9
Trial 10, user: A actual: A 9/10
Probability of guessing: 1.1%
A=Reference,B=Comparison
Test type: ABX
Does this correspond to my conclusion (from df measurements) that discerning them "needs very careful listening"?

At least now you have an idea how audible is the difference in 1.3dB of df levels in case of similar artifact signatures (distance = 0.18dB). This will help us to draw conclusions for your other test items:

4-Bit Dither.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-16.5081-15.0720-0.1327.png

4-Bit Dither.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-16.5081-15.0720-0.1327

4-Bit Truncated.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-21.2348-19.6598-0.2708.png

4-Bit Truncated.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-21.2348-19.6598-0.2708

8-Bit Dither and Noise Shaping.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-40.1551-23...png

8-Bit Dither and Noise Shaping.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-40.1551-23.5380-0.1230

8-Bit Dither.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-40.4684-38.9241-0.1165.png

8-Bit Dither.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-40.4684-38.9241-0.1165

8-Bit Truncated.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-45.2279-43.7093-0.1718.png

8-Bit Truncated.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-45.2279-43.7093-0.1718

8-Bit Waves L1 Type 1 Dither and Ultra Noise Shaping.flac_cut.wav(44)__Original (-6dB).flac(44...png

8-Bit Waves L1 Type 1 Dither and Ultra Noise Shaping.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-22.9337-21.1947-0.2337

Sonnox Oxford Dynamics (Compressor ONLY).flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-...png

Sonnox Oxford Dynamics (Compressor ONLY).flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-69.0738-38.9037-0.0178

WaveArts Tube Saturator Classic.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-32.6024-2...png

WaveArts Tube Saturator Classic.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-32.6024-23.9921-0.0075

Waves L316.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-83.3550-40.8946-0.1463.png

Waves L316.flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-83.3550-40.8946-0.1463

Waves SSL Channel E (Dynamics and EQ OFF).flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400...png

Waves SSL Channel E (Dynamics and EQ OFF).flac_cut.wav(44)__Original (-6dB).flac(44)__mono_400-86.2978-68.0249-0.0113

Similarity of their artifact signatures:

dendro_xr100_13.png

Perceived audio quality of eight DUTs (from bottom of the dendrogram) can be assessed by their df levels. Probably WaveArts Tube Saturator Classic can be also included into the cluster of close DUTs. Df levels of other DUTs are not indicative of their SQ.

Now you can take df levels of DUTs from "good cluster" and compare them to perceived closeness to original. For example one can take samples from good cluster, listen them and sort according to closeness to original, then compare the found order with df levels.
 

Attachments

  • dendro_xr100_12.png
    dendro_xr100_12.png
    7.3 KB · Views: 103
Last edited:

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
Firstly, thank you, Serge, for responding with those test results and your writeup.

Does this correspond to my conclusion (from df measurements) that discerning them "needs very careful listening"?

About as careful as I'd expected--especially via desktop PC speakers in a room with a noisy bunch of fans in a PC case and a fan heater...

ABX tests are "torture tests." It was definitely easier than ABX'ing, for instance, 320kbps AAC with an uncompressed version. (Which I have done, and yes, I could. I think one of the tests I did was over 12 trials; I have the results somewhere. I used IEM's for those tests.)

I'd want to also do the other pairs--particularly, the "Original" and "Softube Trident A-Range." Off the top of my head, I wouldn't necessarily bet on being able to tell the difference between those two in an ABX Test.

Here's the "Delta Waveform" plots from DeltaWave for the "Soft Clip" and "Softube Trident A-Range":

ASR40.png



ASR41.png


Once again, it can be seen that the "Soft Clip" process (top) has a much larger effect on transient peaks. That's what to listen out for--it's what I listened for when presented with "X."

At least now you have an idea how audible is the difference in 1.3dB of df levels in case of similar artifact signatures (distance = 0.18dB). This will help us to draw conclusions for your other test items:

[snip...]

Similarity of their artifact signatures:

I'm afraid the "8-Bit Dither and Noise Shaping" result is invalid--as I mentioned, this was performed by a Reaper "JS Script" plug-in, and the noise shaping loop produced an output signal exceeding 0dBFS, which causes the plug-in to turn off its noise shaper. Presumably at 8-bits the noise shaper was unstable? Anyway, I used a version of the script that I'd modified to turn noise shaping back on in case of overs. It does this by waiting for a certain number of samples before doing so; I wrote it so I could use it on the Master bus without having to keep resetting the noise shaper in case of overs. (Usually 24-bit output, so that would be "my" fault for going over the Master bus rather than the noise shaper going off the rails...)

In a nutshell, the result was that the noise shaper was turned on and off throughout the file.

Perceived audio quality of eight DUTs (from bottom of the dendrogram) can be assessed by their df levels. Probably WaveArts Tube Saturator Classic can be also included into the cluster of close DUTs.

OK, before proceeding further, I think it would be helpful if you define/explain some of your terminology. (Please forgive me if you've already done so previously in this thread.)

Particularly--artifact signature, distance, what the dendrogram represents, what the plots that you've posted for each file are.

Df levels of other DUTs are not indicative of their SQ.

Because...?

Now you can take df levels of DUTs from "good cluster" and compare them to perceived closeness to original. For example one can take samples from good cluster, listen them and sort according to closeness to original, then compare the found order with df levels.

What is "good" about the "cluster?" And what do you suggest to use for "listening and sorting"?
 
Last edited:

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
So, Serge how can I send you 8 files (2 originals, 3 distortions of each) They are synthetic, and only to be listened to at low levels.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
So, Serge how can I send you 8 files (2 originals, 3 distortions of each) They are synthetic, and only to be listened to at low levels.

I think it would be good if you were able to share the files with all participants?

If they are too large to be attachments on this site, then there are plenty of other options. I used https://wetransfer.com/ -- free up to 2GiB. (The free service provided by WeTransfer will automatically delete the files after a week.)
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
I think it would be good if you were able to share the files with all participants?

If they are too large to be attachments on this site, then there are plenty of other options. I used https://wetransfer.com/ -- free up to 2GiB. (The free service provided by WeTransfer will automatically delete the files after a week.)

Never shared a file on this site, other than a photo. Where do I find the mechanism?

I will share them with one requirement KEEP THE LEVEL DOWN!!!!!!

Note wavefiles are not included in the file attachments to a post. They are not large.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
https://we.tl/t-4RUHhDpPL5

bzzzz44 and altbuzz are the originals, d2 and d3 are distorted versions. I'm curious to see what you get from these.

KEEP LEVEL VERY FAR DOWN IF YOU LISTEN. These are NOT FOR LISTENING TO EXCEPT AT VERY LOW LEVELS.

Remember this is very much transitory storage.
 
Last edited:

scott wurcer

Major Contributor
Audio Luminary
Technical Expert
Joined
Apr 24, 2019
Messages
1,501
Likes
2,821
There are two effects, one is time smear, the other one is changes in the signal envelope when there is more than one frequency component in an ERB, which almost is always true in a real audio signal.

The combination is quite complicated. Above 2kHz, the sensitivity to waveform timing is mostly gone, but envelope sensitivity is there until you're past your ability to hear that frequency range.

What would you use to compute the signal envelope in this case? This came up elsewhere and I was playing with the magnitude of the analytic signal, but I have never tried this with complex audio signals just sweeps, chirps, or other standard applications.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
2,267
Likes
4,758
Location
My kitchen or my listening room.
What would you use to compute the signal envelope in this case? This came up elsewhere and I was playing with the magnitude of the analytic signal, but I have never tried this with complex audio signals just sweeps, chirps, or other standard applications.

I use analytic signal, i.e. abs(ifft(positive spectrum)) You can weight the spectrum appropriately to a given ERB to get a real good idea of what's up in that ERB.
 

Serge Smirnoff

Active Member
Joined
Dec 7, 2019
Messages
240
Likes
136
In a nutshell, the result was that the noise shaper was turned on and off throughout the file.
It is visible on the diffrogram of this sample, which basically shows how a signal is distorted in time.

Particularly--artifact signature, distance, what the dendrogram represents, what the plots that you've posted for each file are.
As I already explained this in another thread I give the link - https://www.audiosciencereview.com/...od-for-measuring-distortion.10282/post-284457

What is "good" about the "cluster?" And what do you suggest to use for "listening and sorting"?
The eight/nine DUTs from the bottom of the dendrogram. They have similar artifact signatures (the processings distort your test vector in a similar way) and so, df levels for them are indicative of their closeness to the original.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK

Thank you very much--I shall read and digest...

The eight/nine DUTs from the bottom of the dendrogram. They have similar artifact signatures (the processings distort your test vector in a similar way) and so, df levels for them are indicative of their closeness to the original.

I find it troubling that the "4-bit dithered" and "4-bit truncated" are so close together and that the "4-bit dithered" gets a worse score. I wonder if you have listened to them?

If an objective is to characterise the performance of DAC's, then how is this helpful, considering that, for example, noise-shaping is fundamental to the operation of delta-sigma DACs?
 
Last edited:

EB1000

Senior Member
Joined
Jan 8, 2020
Messages
484
Likes
579
Location
Israel
Here are a few biological facts:

No human can hear outside 20Hz to 20kHz frequency response.

No human can hear harmonic distortion below 1% for music or below 0.5% for a pure sinewave.

No human can tell the difference between music played using 120db dynamic range, and the same music played using 70db+ dynamic range.

No human can hear background noise of music played with SNR of 70db and above


IOW, any audio component that meets the above minimum scores, will sound just as good to any other component that scores higher measurements results...
 
Top Bottom