MiniDSP SHD vs Pioner VSX-LX505 : noticeable difference in sound quality

[Kachda]

I recently purchased a Pioneer VSX-LX505 to see if I could move from a stereo to 7.1.2 setup. I made a head to head comparison, using the same amplifier and speaker, and with all processing off (Pure Direct on AVR, Dirac off on Minidsp). Both speakers were placed right next to each other and same channel was used. I used pink noise and a UMIK-1 to make sure they are both producing SPL within 1db of each other.

Setup was
Apple TV -> Pioneer -> TV -> TV Optical Out -> Minidsp SHD -> Hypex NC502MP Amp Channel 1 -> Speaker (Wharfedale Linton)
Apple TV -> Pioneer -> Pre Out -> Hypex NC502MP Channel 2 -> Speaker (Wharfedale Linton)

I can distinctly make out that the Pioneer Pre-outs lack as much attack and detail as the SHD. I also took a recording of both and performed an ABX test in Foobar, and got 16/16 (test.txt).

Amir found the LX505 to have a SINAD of 97, and MiniDSP is 111, both of which clear CD quality (96db if I am not mistaken). Given this, why is the sound from the Pioneer worse given everything else being equal (and no sound processing taking place). ?

 

[Miker 1102]

I recently purchased a Pioneer VSX-LX505 to see if I could move from a stereo to 7.1.2 setup. I made a head to head comparison, using the same amplifier and speaker, and with all processing off (Pure Direct on AVR, Dirac off on Minidsp). Both speakers were placed right next to each other and same channel was used. I used pink noise and a UMIK-1 to make sure they are both producing SPL within 1db of each other.

Setup was
Apple TV -> Pioneer -> TV -> TV Optical Out -> Minidsp SHD -> Hypex NC502MP Amp Channel 1 -> Speaker (Wharfedale Linton)
Apple TV -> Pioneer -> Pre Out -> Hypex NC502MP Channel 2 -> Speaker (Wharfedale Linton)

I can distinctly make out that the Pioneer Pre-outs lack as much attack and detail as the SHD. I also took a recording of both and performed an ABX test in Foobar, and got 16/16 (test.txt).

Amir found the LX505 to have a SINAD of 97, and MiniDSP is 111, both of which clear CD quality (96db if I am not mistaken). Given this, why is the sound from the Pioneer worse given everything else being equal (and no sound processing taking place). ?

I bought the 503 model. My subjective listening was a poor experience.

 

[YSDR]

It might be worth mentioning that if we don't use 2V (LX505) or 4V (SHD) output levels, the SINAD of the devices will be different than what Amir measured, but how much and what other factors might come in, we don't know. And you probably don't use such high signal levels, because with the NC502MP it would be very loud (if the speakers can handle that power at all) and with 4V the amp would be overloaded anyway.

 

[MaxwellsEq]

I used pink noise and a UMIK-1 to make sure they are both producing SPL within 1db of each other.

I think that using pink noise with a calibrated microphone, whilst better than nothing, is probably not accurate enough. It needs to be done with a voltmeter and and tone (e.g. at 400Hz) with 0.1V difference.

There was an excellent article by Bob Carver that is no longer publicly available, that included:

"When we were finally able to get the output levels of the two power amplifiers exactly matched, there was absolutely no audible difference when switching between them while listening to either white noise or music. During the adjustments of the amplifiers, it was demonstrated dramatically that minute differences in volume level (sound quantity) that are too subtle to be heard as such are interpreted by the ear as "obvious" differences in sound quality. Everyone was startled by the effect — everyone, that is, except Larry Klein, who had touched upon the phenomenon some time ago in his Audio Questions and Answers column

 

[welwynnick]

Everybody bends over backwards, leaving absolutely no stone unturned, to find evidence that all audio electronics sound the same.

The simple answer could be that the MiniDSP sounds better.

If curiosity demands, you could test again with the Pioneer playing 1 dB louder. Whenever I've done this, it never made any difference.

Nick

 

[GXAlan]

I haven’t had a chance to run deltawave on your recordings, but that’s your next step

 

[Sokel]

Visuals:



No wonder differences were so prominent.

 

[GXAlan]

@Kachda
Now you just need to record the same device twice to verify that your UMIK is consistent (negative control).

 

[Kachda]

Visuals:

View attachment 349124 View attachment 349125 View attachment 349126 View attachment 349127

No wonder differences were so prominent.

Thanks for running this. Can you tell me which recording is the blue vs white line?

 

[MaxwellsEq]

The simple answer could be that the MiniDSP sounds better.

Or there are some other differences, as shown by deltawave.

 

[Sokel]

Thanks for running this. Can you tell me which recording is the blue vs white line?

The white one is "Recording" and the blue one is "Recording (2) "

 

[welwynnick]

Or there are some other differences, as shown by deltawave.

Of course there are differences if they sound different.

 

[MaxwellsEq]

Of course there are differences if they sound different.

Causality.

They sound different because they measure quite differently, which they shouldn't, given this was supposed to be an A vs B test.

Everyone knows if product R has a measurable bass and treble boost and product N has a measurable mid boost, the two products must sound different.

 

[welwynnick]

I'm not sure what the point is there. They ARE different AND measure different AND sound different. What's the problem?

Whether they should or should not measure differently is another matter. Going by Amir's measurements I'd expect the difference between them to be small, but it appears that's not the case. My money is on an unknown difference in the settings and processing, but it could simply be load sensitivity.

 

[Kachda]

I'm not sure what the point is there. They ARE different AND measure different AND sound different. What's the problem?

Whether they should or should not measure differently is another matter. Going by Amir's measurements I'd expect the difference between them to be small, but it appears that's not the case. My money is on an unknown difference in the settings and processing, but it could simply be load sensitivity.

I also suspect some internal processing happening in one of them, even though I’ve tried my best to turn off everything.

 

[GXAlan]

I also suspect some internal processing happening in one of them, even though I’ve tried my best to turn off everything.

The only wildcard is if the microphone is inaccurate. The way you prove this is not the case is to show that playing the same system twice generates much deeper nulls. So if A1 is like A2 and B1 is like B2 but A is very different than B, you are done.

Then you have proven, that two similarly measuring DACs that theoretically are transparent in a 1 kHz SINAD result in an audible difference at the speaker level.

 

[voodooless]

Both speakers were placed right next to each other and same channel was used. I used pink noise and a UMIK-1 to make sure they are both producing SPL within 1db of each other.

So, once and for all, it should be clear now why we keep repeating that using microphones or SPL meters to level match is not good enough, because clearly the levels were nowhere near close enough.

Next we have more differences. You used two speakers, and they were next to each other? Was the microphone moved between measurements? Do we know for sure that the speakers measure the same? Since they are not at the same position, this could result in differences in measured FR. That alone could be enough to easily pass an ABX.

 

[Kachda]

So, once and for all, it should be clear now why we keep repeating that using microphones or SPL meters to level match is not good enough, because clearly the levels were nowhere near close enough.

Next we have more differences. You used two speakers, and they were next to each other? Was the microphone moved between measurements? Do we know for sure that the speakers measure the same? Since they are not at the same position, this could result in differences in measured FR. That alone could be enough to easily pass an ABX.

Sorry, what makes you conclude the levels aren’t close enough? @Sokel ’s graph shows differences in FR but not significant differences in average levels?

 

[voodooless]

Sorry, what makes you conclude the levels aren’t close enough? @Sokel ’s graph shows differences in FR but not significant differences in average levels?

Look again, first image.

 

[GXAlan]

So, once and for all, it should be clear now why we keep repeating that using microphones or SPL meters to level match is not good enough, because clearly the levels were nowhere near close enough.

Next we have more differences. You used two speakers, and they were next to each other? Was the microphone moved between measurements? Do we know for sure that the speakers measure the same? Since they are not at the same position, this could result in differences in measured FR. That alone could be enough to easily pass an ABX.

Look again, first image.

@pkane
DeltaWave takes care of level errors and this seems reasonably close. So if he heard a difference, but it didn’t have a difference on DeltaWave, you could say it was just the level. Since DeltaWave is showing a difference what needs to be done is to show that comparing the same setup twice also nulls deeply.

Using a microphone has more noise so it should be harder to show a difference as long as your negative controls are good.

But moving the microphone is bad…