• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Can you tell diode limiter distortion in the ABX test (sine, music)?

Would you like to participate in the diode distortion test?

  • Yes

    Votes: 4 100.0%
  • No

    Votes: 0 0.0%

  • Total voters
    4
  • Poll closed .

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
Please let me know if you would be interested to participate in the test with diode limiter distortion. BAT Schottky diode is connected in parallel to the output, behind 1k resistor. The resulting distortion at -2dB, -11dB and -21dB is shown in the following plots. The upper (left) channel is always the distorted one, the lower (right) channel is a plain loopback.

Here is the link to the test files - music samples
https://www.dropbox.com/s/x9jiu26haku8wpw/diodelimiter.zip?dl=0
Please download the file and then unzip, you will get 2 flac files. It is impossible to play directly from dropbox.

Sine distortions:

diode_-2dB.png


diode_-11dB.png


diode_-21dB.png
 
Last edited:

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
1,587
Likes
3,426
Location
Berlin, Germany
Excellent, Pavel!

You outraced me, as I was just about to do the same thing. Actually I planned to use a series R + D accross the 100R output impedance of my RME, with the R dialed in so that the distortion is just a tad below (buried in) the normal loopback distortion, making it invisible. Then using the power of heavy block-averaging prior to diff'ing the signals with the circuit in vs out, to make it visible/audible.

Not the same kind of test of course, nothing will ever be heard in my test. In your test, probably nothing will be heard either, as I found static distortion is always audible quite benign.

Therefore I'm thinking of a dynamic distortion circuit, adding a C//R in series with the D+R shunt. That would change the bias voltage the diode is shorting to, depending on the signal and its history.... in a sine distortion test, the measured distortion would be quite low as the cap remains charged. In a large-period IMD, notably one optimized for high crest factor (all components starting with zero cosine phase) the distortion would be higher, and with real music tracks it will be rather high.
 
  • Like
Reactions: pma
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
Hi Klaus,

Thanks! Let's hope that something might be heard, this is the shot of the original and limited file. I will post the test files.

1612606789977.png
 
Last edited:
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
This time it is really easy, if you concentrate on sibilants. So again, transients!! Distortion spectra in isolation do not tell the story.

Code:
foo_abx 2.0.2 report
foobar2000 v1.4.8
2021-02-06 11:54:35

File A: toomuchin_origrec_1.flac
SHA1: c3a9d915d6cd4980f931416b005ff2cd8899dd56
File B: toomuchin_origrec_2.flac
SHA1: 3d2c820a1c691608fb8d319e5faf87fc54fa0213

Used DSPs:
Resampler (PPHS), Convolver

Output:
ASIO : DUO-CAPTURE EX
Crossfading: NO

11:54:36 : Test started.
11:56:00 : 01/01
11:56:10 : 02/02
11:56:19 : 03/03
11:56:23 : 04/04
11:56:27 : 05/05
11:56:36 : 06/06
11:56:44 : 07/07
11:56:54 : 08/08
11:57:04 : 09/09
11:57:13 : 10/10
11:57:13 : Test finished.

----------
Total: 10/10
Probability that you were guessing: 0.1%

-- signature --
638dc6d645b433cb19ceda89eb5480638fa2c8b3
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
1,587
Likes
3,426
Location
Berlin, Germany
The distorted vowels of the voice (at 0:37 for example) were a dead giveaway for me, so its not just transient alone.. no ABX needed on that ;-)
 
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
The distorted vowels of the voice (at 0:37 for example) were a dead giveaway for me, so its not just transient alone.. no ABX needed on that ;-)

I mean that from distortion spectra of the sine wave I would not assume such severe clipping in one polarity. This says that one kind of plot analysis is not enough to predict the effect. Similar distortion got regularly from tubes does not result in same audible effect :). That's why audio is so exciting.
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
1,587
Likes
3,426
Location
Berlin, Germany
Yep, exactly. In this case, the (seldomly shown) phases of the distortion components matter as much as the magnitudes do, in the simple sine THD measurement, as that decides what the actual waveform looks like. At some point @pkane might implement adjustable phases in his distortion generator so this can more readily investigated.
Simple scalar steady-state metrics like THD (and SINAD) are useless, which is what we both know, but others may not.
 
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
Just for fun, these are the sine waves according to this profile https://www.audiosciencereview.com/forum/index.php?attachments/diode_-11db-png.110851/ , clean loop vs. diode limiter, 1kHz sine

P.S.: the transient part of the diode limiter response is during about 0.4s, if you look at the time series. It is because of the time constant associated with the low frequency HP cut-off. The diode connected behind the series 1k resistor would normally limit only the negative part of the wave. However, as the output coupling is through an electrolytic cap, the DC disappears with the HP cut-off. This also affects the distortion, which is highest at the beginning. This also explains why music sample is highly clipped and distorted. Nothing is as simple as it looks at first sight ;).
diode_clip.png
 

Attachments

  • sine_diode_or.zip
    550.9 KB · Views: 33
Last edited:
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
Distortion vs. amplitude [V] of the loop and diode limiter path. Please note how early the diode affects distortion.

diode_vs_loop_thdampl.png
 
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,398
Likes
4,415
Location
Prague
Top Bottom