• 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.

Listening test of 2 power amplifiers - files recorded for download - disclosed

Can you hear a difference and which file do you prefer

  • I can hear a difference but have no ABX result

    Votes: 4 14.3%
  • I cannot hear a difference but have no ABX result

    Votes: 14 50.0%
  • I can hear a difference and have an ABX result

    Votes: 2 7.1%
  • I cannot hear a difference and have an ABX result

    Votes: 4 14.3%
  • I prefer witch1 file

    Votes: 1 3.6%
  • I prefer witch2 file

    Votes: 8 28.6%

  • Total voters
    28
  • Poll closed .

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
Yes, these don't match nearly as well as 1 and 4.
I see now.

So, amp 1 is the one with larger imbalance. It is down by 0.3db for right channel as compared to orig.

Amp2 is closer to orig in term of balance.

I thought one amp was louder on the right as compared to orig, thus when I see #3 is louder than #4 for right channel, I automatically thought #3 is not the orig.
 
Last edited:

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,258
Likes
1,046
@pma would you be willing to do a loopback recording to compare?
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
Love your efforts on this - you're teaching a lot of folks important stuff here. : )

You're absolutely right, the interchannel dynamic range (difference between the loudest and quietest portions of the song) is nearly identical. I was wrong to use that term, but the right channel of the better sounding track in both cases has higher peaks and averages a bit louder. This is also the case with witch3 compared to witch4. I clearly heard this on my Klipschorns when comparing witch1 to witch2, but with the witch3 and witch4 files I simply looked at the data you presented above, but for tracks 3 and 4 (below).

I still say that witch3 is the original and if I'm right I want an invitation to Prague! : )

God bless you and your precious family - Langston

PS: I wasn't able to hear such small differences until I started using horn loaded loudspeakers.

View attachment 147894

View attachment 147895
Curious, do you hear a difference between #2 and #3 in your setup?
 

cany89

Active Member
Joined
Dec 22, 2020
Messages
238
Likes
115
Have you made any progress here?

Unfortunately one of my monoblocks gone rogue. Took it to the shop and still waiting. I’ll give the tracks some time when it’s back. :/
 
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,289
Likes
4,057
Location
Prague
#2 should not be directly compared to #3 in an ABX test because there is a big difference in the time when the files start. Time shift is 77936 samples and this will be of course easily detectable. For the ABX, only #1 vs. #2 or #3 vs. #4 are possible. It was the goal not to identify easily which of #3 and #4 is the original, to prevent further speculations.
 
Last edited:
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,289
Likes
4,057
Location
Prague
For once more I am posting amplitude and rms statistics (Adobe Audition) of the files under test

witch1 x witch2
witch1_stat.png
witch2_stat.png



witch3 x witch4
witch3_stat.png
witch4_stat.png


Time difference between witch1&2 against witch3&4 is 77936 samples. Witch3&4 have later beginning, 77936 first samples with noise (from the original track) only were cut. Thus the only valid ABX pairs are 1x2 and 3x4.
witch_stats.png


DW reports (stereo) for 3x4 and 3x2
 

Attachments

  • reports.pdf
    25.4 KB · Views: 14
Last edited:

pogo

Active Member
Joined
Sep 4, 2020
Messages
261
Likes
63
My favourite is definitely witch3 in a simple listening comparison.
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
For once more I am posting amplitude and rms statistics (Adobe Audition) of the files under test

witch1 x witch2
View attachment 147984 View attachment 147985


witch3 x witch4
View attachment 147986 View attachment 147987

Time difference between witch1&2 against witch3&4 is 77936 samples. Witch3&4 have later beginning, 77936 first samples with noise (from the original track) only were cut. Thus the only valid ABX pairs are 1x2 and 3x4.
View attachment 147990

DW reports (stereo) for 3x4 and 3x2
Don't you think your comparisons are kind of flawed?

Due to channel differences, most people are going to prefer the louder one.

You don't know for sure if the preference is due to other factors....
 
OP
pma

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
2,289
Likes
4,057
Location
Prague
Don't you think your comparisons are kind of flawed?

Due to channel differences, most people are going to prefer the louder one.

You don't know for sure if the preference is due to other factors....

Comparisons are not flawed, the files are as recorded without further manipulation with level. It may be interesting to see how the files are distinguishable. I have not seen a statistically strong ABX result yet.

For those who would declare they hear a difference, especially in the new set witch #3 and #4, I would ask to post a valid ABX protocol to consider them serious. It looks like the one I made (unsuccessful).

Code:
foo_abx 2.0.2 report
foobar2000 v1.4.8
2021-08-16 16:13:13

File A: witch3.wav
SHA1: c5baa55f2e069aaf20d1dd68fab58e0b45f6e22d
File B: witch4.wav
SHA1: ed45cd1f28fabe63260873b04cafc0381c488a86

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

16:13:14 : Test started.
16:14:00 : 01/01
16:14:13 : 01/02
16:14:23 : 01/03
16:14:35 : 01/04
16:14:43 : 02/05
16:14:52 : 02/06
16:14:59 : 02/07
16:15:10 : 03/08
16:15:17 : 04/09
16:15:24 : 05/10
16:15:38 : 06/11
16:15:46 : 07/12
16:15:56 : 07/13
16:16:03 : 07/14
16:16:10 : 08/15
16:16:18 : 09/16
16:16:18 : Test finished.

 ----------
Total: 9/16
Probability that you were guessing: 40.2%

 -- signature --
980c331994da1bafee09124b9a375e86a90f09fb
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
Comparisons are not flawed, the files are as recorded without further manipulation with level. It may be interesting to see how the files are distinguishable. I have not seen a statistically strong ABX result yet.

For those who would declare they hear a difference, especially in the new set witch #3 and #4, I would ask to post a valid ABX protocol to consider them serious. It looks like the one I made (unsuccessful).

Code:
foo_abx 2.0.2 report
foobar2000 v1.4.8
2021-08-16 16:13:13

File A: witch3.wav
SHA1: c5baa55f2e069aaf20d1dd68fab58e0b45f6e22d
File B: witch4.wav
SHA1: ed45cd1f28fabe63260873b04cafc0381c488a86

Output:
ASIO : Focusrite USB ASIO
Crossfading: NO

16:13:14 : Test started.
16:14:00 : 01/01
16:14:13 : 01/02
16:14:23 : 01/03
16:14:35 : 01/04
16:14:43 : 02/05
16:14:52 : 02/06
16:14:59 : 02/07
16:15:10 : 03/08
16:15:17 : 04/09
16:15:24 : 05/10
16:15:38 : 06/11
16:15:46 : 07/12
16:15:56 : 07/13
16:16:03 : 07/14
16:16:10 : 08/15
16:16:18 : 09/16
16:16:18 : Test finished.

----------
Total: 9/16
Probability that you were guessing: 40.2%

-- signature --
980c331994da1bafee09124b9a375e86a90f09fb
You have seen my ABX for #1 and #2, correct? I surely hope it is good enough to show that indeed channel imbalance of 0.3db is audible for certain people with sensitive hearing. You can't pass #1 vs #2 either.....so you can't pass #3 vs #4, well, means not much.

This imbalance is muddy up all other conclusions you may draw from.

Why do you think #3 and #4 will be any different?
 
Last edited:

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,258
Likes
1,046
Non-linear EQ settings are designed to compensate for the effects of filters on frequency and phase. It is best not to use these unless you know that you need to correct for these (for example, when comparing two different DACs or same DAC with different filters).
Back to this question: it seems that all of the recordings of amplifier output compared to a loopback or original file have worse matching at low frequencies. Is this just mismatch from the phase shift of decoupling the amplifier from DC? If that is the case, doesn't it make sense to use the Phase EQ option in DeltaWave when comparing amplifier voltages to loopbacks (or comparing amplifier output to original tracks)?
 

pkane

Major Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
2,563
Likes
4,235
Location
North-East
Back to this question: it seems that all of the recordings of amplifier output compared to a loopback or original file have worse matching at low frequencies. Is this just mismatch from the phase shift of decoupling the amplifier from DC? If that is the case, doesn't it make sense to use the Phase EQ option in DeltaWave when comparing amplifier voltages to loopbacks (or comparing amplifier output to original tracks)?

Remember, the goal is to determine what differences there are between the two devices/amps. If you correct for phase/frequency differences using non-linear EQ, you'll get a better match in DeltaWave, but you will be ignoring some real differences in this case.
 

levimax

Major Contributor
Joined
Dec 28, 2018
Messages
1,073
Likes
1,504
Location
San Diego
You have seen my ABX for #1 and #2, correct? I surely hope it is good enough to show that indeed channel imbalance of 0.3db is audible for certain people with sensitive hearing. You can't pass #1 vs #2 either.....so you can't pass #3 vs #4, well, means not much.

This imbalance is muddy up all other conclusions you may draw from.

Why do you think #3 and #4 will be any different?

When I do Foobar2000 ABX it first "normalizes" the volume of the 2 tracks... wouldn't that make the 0.3 dB channel balance difference disappear? Did you do the Foobar 2000 ABX without normalizing the tracks?
 

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,258
Likes
1,046
Remember, the goal is to determine what differences there are between the two devices/amps. If you correct for phase/frequency differences using non-linear EQ, you'll get a better match in DeltaWave, but you will be ignoring some real differences in this case.
I see, but if the differences are dominated by this effect, which may not be important, does it make sense to take it out? We have to compare the amps to the original file or loopback to know if one amp gives a more faithful output than another (maybe showing that one is better than another is not the goal of the OP but that is what I would want to get at). Maybe there is something else that should stand out even with the HP filter effect left in the comparison?
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
When I do Foobar2000 ABX it first "normalizes" the volume of the 2 tracks... wouldn't that make the 0.3 dB channel balance difference disappear? Did you do the Foobar 2000 ABX without normalizing the tracks?
Yes, no normalized. This is to ABX the files as is. Normalize them defeats the purpose of sensing a real different between two different amps.
 

pogo

Active Member
Joined
Sep 4, 2020
Messages
261
Likes
63
Back to this question: it seems that all of the recordings of amplifier output compared to a loopback or original file have worse matching at low frequencies.
In the real world, you still see the mass-dependent mechanical speaker swing-in/out behavior, which is not contained in the original signal. The back driven energy can be better handled by a higher system damping factor and the original signal is followed more precisely.
 
Last edited:

pkane

Major Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
2,563
Likes
4,235
Location
North-East
I see, but if the differences are dominated by this effect, which may not be important, does it make sense to take it out? We have to compare the amps to the original file or loopback to know if one amp gives a more faithful output than another (maybe showing that one is better than another is not the goal of the OP but that is what I would want to get at). Maybe there is something else that should stand out even with the HP filter effect left in the comparison?

How do you define "important"? Is frequency response important? Is phase? If you want to know the approximate audibility of the differences, look at PK Metric plot in DeltaWave. Don't use non-linear EQ corrections.
 

levimax

Major Contributor
Joined
Dec 28, 2018
Messages
1,073
Likes
1,504
Location
San Diego
Yes, no normalized. This is to ABX the files as is. Normalize them defeats the purpose of sensing a real different between two different amps.
OK I get the point when it comes to phase or FR but when it comes to small volume level differences those are just a "recording variation" ..... certainly either amp can play a little louder or softer so not sure you can say you are testing the amps for a difference... more like you are testing the "capture" for differences.
 

Pdxwayne

Major Contributor
Joined
Sep 15, 2020
Messages
2,673
Likes
904
OK I get the point when it comes to phase or FR but when it comes to small volume level differences those are just a "recording variation" ..... certainly either amp can play a little louder or softer so not sure you can say you are testing the amps for a difference... more like you are testing the "capture" for differences.
Well, that depends....

Only PMA can confirm that, but if after multiple recording tries and the the captures are consistent for both amps, I would say it is no longer recording variations.
 

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,258
Likes
1,046
How do you define "important"? Is frequency response important? Is phase? If you want to know the approximate audibility of the differences, look at PK Metric plot in DeltaWave. Don't use non-linear EQ corrections.
If I look at the PK Metric of original vs amplifier recordings for a given track, they look similar no matter which amplifier. So it seems difficult to identify a better amplifier this way. Maybe that is just what we are up against when the differences in direct amp to amp comparison are small.
 
Top Bottom