Phase shift and echo trails length illusion

[Pdxwayne]

May I know phase shift difference between DAC could cause echo trails' length differences between DAC? When I first got the X16, I thought it is lacking the sparkles of E30. Now that I did more measurements, I am wondering the extra sparkles could be due to E30 have longer echo trails because it phase shift differently than X16 in the high frequencies.

Please let me know if there is any such relationship and why. Thanks!

Here are the DeltaWave comparisons of E30 vs X16, playing a song with echo trails effects.

Matched spectrum:

Diff Spectrum:

Phase diff:

 

[gvl]

Were the digital filter types matched for this test? The phase difference may be just because one runs minimal and the other linear phase type of filters.

 

[Pdxwayne]

Were the digital filter types matched for this test? The phase difference may be just because one runs minimal and the other linear phase type of filters.

E30 filter 1. X16 filter 2. According to @solderdude , this is as close as a match as possible....

 

[solderdude]

In frequency extension these filters will be close enough, in any case better than filter 1 and slow filters.
Phase wise they can differ. It's very little and gentle sloping. Only very fast and substantial phase differences may become audible.
Here too there is nothing to worry about.
Any speaker and even headphones have much larger phase shifts.
Headphones would show phase shifts better than speakers.
With music phase shifts aren't audible.
With only a single voice people seem to be able to discern phase shifts but not small ones and at midrange frequencies.

 

[Pdxwayne]

In frequency extension these filters will be close enough, in any case better than filter 1 and slow filters.
Phase wise they can differ. It's very little and gentle sloping. Only very fast and substantial phase differences may become audible.
Here too there is nothing to worry about.
Any speaker and even headphones have much larger phase shifts.

This is just a left channel captures. Would capture right channel and compare the channels of the same DAC meaningful?

Thanks!

 

[solderdude]

Only when you have a mono signal.

 

[solderdude]

Echo trails cannot suddenly become longer or shorter. It is part of the waveform and doesn't have a separate status.
Besides you would need non linear behavior for that (waveform compression) for trails only which would go hand in hand with 3rd harm. distortion.
This would be measurable with nulling which was not the case.
Here too: only blind listening tests can bring clarity.

 

[Pdxwayne]

Echo trails cannot suddenly become longer or shorter. It is part of the waveform and doesn't have a separate status.
Besides you would need non linear behavior for that (waveform compression) for trails only which would go hand in hand with 3rd harm. distortion.
This would be measurable with nulling which was not the case.
Here too: only blind listening tests can bring clarity.

Thanks!

@pkane said this:
Anything above -50dBFS with PK Error Metric might be audible.

This PK Error part is interesting. I will check later for 3 different songs that I already captured.

 

[Pdxwayne]

Here is the Delta of Waveform and PK Metric chart. According to @pkane, the difference in PK Metric should above -50dbFS. The PK of this comparison is at most -67. So, I guess unlikely to be audible?

@solderdude, the interesting part of the charts is that the vocal part with echo trails start around 29 seconds or so. 39 seconds also vocal part with echos. How subtle of echo trails difference can one detect? I would guess in blind tests, I would struggle based on the PK Metric numbers....

Delta of Waveform:

PK Metric

 

[pma]

I also think there is no correlation of these plots and audible differences.

 

[pkane]

I also think there is no correlation of these plots and audible differences.

There is some, by design.

 

[Pdxwayne]

There is some, by design.

Anything audible based on the PK chart? 39 seconds is a part with vocal with echos. Thanks!

 

[pkane]

Anything audible based on the PK chart? 39 seconds is a part with vocal with echos. Thanks!

Unlikely. You can listen using an AB/X type of a test for any differences between 38 and 46 seconds of the track, where the differences are largest. But even there, it's very unlikely you'll hear them.

 

[Pdxwayne]

Unlikely. You can listen using an AB/X type of a test for any differences between 38 and 46 seconds of the track, where the differences are largest. But even there, it's very unlikely you'll hear them.

Yes, I agree it is very hard to hear a difference. But measurements wise, there were differences, correct? Thanks!

 

[pkane]

Yes, I agree it is very hard to hear a difference. But measurements wise, there were differences, correct? Thanks!

There will always be differences between two recordings passing through an analog stage, and sometimes through digital. The important question is whether they are audible to you. Even if not audible, the next question you may want to ask is which of the two devices distorts less (or is a better engineered device). This is where you'd want to compare each of the loopback recordings to the original, digital file. This will tell you what effect each of these devices has on the source.

 

[Pdxwayne]

There will always be differences between two recordings passing through an analog stage, and sometimes through digital. The important question is whether they are audible to you. Even if not audible, the next question you may want to ask is which of the two devices distorts less (or is a better engineered device). This is where you'd want to compare each of the loopback recordings to the original, digital file. This will tell you what effect each of these devices has on the source.

Unfortunately, for this particular song with echo trails, I do not have original digital file on hand. It is a Tidal streaming song.

Interesting that, if I compare the same two DAC using 15 seconds captured bagpipe music, which have no echo trails, the differences looks smaller.

Delta of waveform:

PK Metric:

 

[pkane]

Unfortunately, for this particular song with echo trails, I do not have original digital file on had. It is a Tidal streaming song.

Interesting that, if I compare the same two DAC using 15 seconds captured bagpipe music, which have no echo trails, the differences looks smaller.

Delta of waveform:
View attachment 114188

PK Metric:
View attachment 114189

DeltaWave can't discover differences that are not present in the data you feed it. If the source track is too short, doesn't contain enough frequencies or levels, you will get an incomplete picture. I recommend using at least 60 seconds test tracks with DeltaWave. 120 seconds will produce a more accurate result. Some full orchestra, large dynamic range music would work best.

 

[Pdxwayne]

DeltaWave can't discover differences that are not present in the data you feed it. If the source track is too short, doesn't contain enough frequencies or levels, you will get an incomplete picture. I recommend using at least 60 seconds test tracks with DeltaWave. 120 seconds will produce a more accurate result. Some full orchestra, large dynamic range music would work best.

Good to know!

I will do longer captures and compare again.

Thanks!

 

[AnalogSteph]

And even then, I suspect in a lot of cases this is just a convoluted way of detecting benign differences related to bass and highs rolloff that RMAA would have shown much more quickly and clearly... or maybe REW, which would have the advantage of displaying phase as well, so you can tell what is FIR and what is IIR.

(FYI, RMAA does have a "mono mode" option.)

 

[Pdxwayne]

And even then, I suspect in a lot of cases this is just a convoluted way of detecting benign differences related to bass and highs rolloff that RMAA would have shown much more quickly and clearly... or maybe REW, which would have the advantage of displaying phase as well, so you can tell what is FIR and what is IIR.

(FYI, RMAA does have a "mono mode" option.)

I am a newbie at this....Had to Google some of the terms and I am not sure I understand FIR vs IIR much.....And how it is related to hearing longer echo trails....

Thanks for suggesting RMAA. Another tool to look into. I do have REW, but not sure how to get started to check phase, FIR, and IIR....More learnings to do.... : )