MQA Testing

[The Computer Audiophile]

Hi Guys - I'm writing an MQA article and an idea came to mind.

 

There has been much discussion about MQA and some people believing un-decoded MQA is inferior to standard redbook CD rips.

 

So, here's a challenge to those who believe it's inferior. Keep in mind that those who believe it's inferior are almost always very objective-minded and almost always demand double blind testing.

 

 

Make sure to use a NON-MQA DAC. Find MQA and non-MQA versions of the same tracks. We can use the 2L test bench for samples if needed. Using the ABX feature of Foobar or similar feature in another app, play both MQA and non-MQA versions un-decoded. If MQA is inferior, it should be pretty easy to identify the un-decoded MQA version.

 

 

What will we find?

 

I don't know.

 

If un-decoded MQA measures worse, according to some people, but it doesn't sound worse, we could concluded that not everything that can be measured matters. Would this make all the discussion about un-decoded MQA being inferior moot? I don't know.

 

We could find that the un-decoded MQA files are easily identifiable as worse. This would put real world experience and the measurements in harmony.

 

We could find the un-decoded MQA sounds better.

 

We could find the two sound the same.

 

This could be a fun, practical exercise. This is just an idea that came to mind, and it's not thought through all the way. I encourage suggestions to improve it.

 

 

Here is the 2L test bench is needed - http://www.2l.no/hires/

[Samuel T Cogley]

Hi Guys - I'm writing an MQA article and an idea came to mind.

 

There has been much discussion about MQA and some people believing un-decoded MQA is inferior to standard redbook CD rips.

 

So, here's a challenge to those who believe it's inferior. Keep in mind that those who believe it's inferior are almost always very objective-minded and almost always demand double blind testing.

 

 

Find MQA and non-MQA versions of the same tracks. We can use the 2L test bench for samples if needed. Using the ABX feature of Foobar or similar feature in another app, play both MQA and non-MQA versions un-decoded. If MQA is inferior, it should be pretty easy to identify the un-decoded MQA version.

 

 

What will we find?

 

I don't know.

 

If un-decoded MQA measures worse, according to some people, but it doesn't sound worse, we could concluded that not everything that can be measured matters. Would this make all the discussion about un-decoded MQA being inferior moot? I don't know.

 

We could find that the un-decoded MQA files are easily identifiable as worse. This would put real world experience and the measurements in harmony.

 

We could find the un-decoded MQA sounds better.

 

We could find the two sound the same.

 

This could be a fun, practical exercise. This is just an idea that came to mind, and it's not thought through all the way. I encourage suggestions to improve it.

 

 

Here is the 2L test bench is needed - http://www.2l.no/hires/

 

Hi Chris. I've tried this, but a bug in the Meridian Explorer 2 (described in this post) prevents this kind of testing. That unit is a relatively inexpensive way to conduct the test. Now you're down to Mytek Brooklyn or Bluesound Node 2. What you'll end up with here is subjective results, which I predict will heavily favor MQA.

[The Computer Audiophile]

Hi Chris. I've tried this, but a bug in the Meridian Explorer 2 (described in this post) prevents this kind of testing. That unit is a relatively inexpensive way to conduct the test. Now you're down to Mytek Brooklyn or Bluesound Node 2. What you'll end up with here is subjective results, which I predict will heavily favor MQA.

 

Hi Sam - Actually, this test must be done without an MQA decoder or MQA enabled DAC.

[Samuel T Cogley]

Hi Sam - Actually, this test must be done without an MQA decoder or MQA enabled DAC.

 

Ok, understood. When I undertook just such a test, I found the increased noise floor in the MQA-encoded material kind of made the test worthless. I respectfully submit it's not a particularly useful measurement of anything.

[The Computer Audiophile]

Ok, understood. When I undertook just such a test, I found the increased noise floor in the MQA-encoded material kind of made the test worthless. I respectfully submit it's not a particularly useful measurement of anything.

 

Ah, that's a good thing. We can correlate the measurements with the subjective results.

[Samuel T Cogley]

Ah, that's a good thing. We can correlate the measurements with the subjective results.

 

If you were getting any negative feedback from MQA about the volume of critical posts on your forums, would you tell us? I'm just trying to put your new ask in context.

[YashN]

Ah, that's a good thing. We can correlate the measurements with the subjective results.

 

I'm currently doing that with my DIY gear, nothing to do with MQA though, and it's quite fascinating.

[The Computer Audiophile]

If you were getting any negative feedback from MQA about the volume of critical posts on your forums, would you tell us? I'm just trying to put your new ask in context.

I haven't received any negative feedback from MQA.

 

I'm here for you guys. I listen to you guys not manufacturers.

 

What I'm getting at is this:

 

Given that-

 

1. MQA marketing claims that un-decoded MQA music is better sounding than standard CD rips.

 

2. Measurements done by some people appear to show un-decoded MQA is inferior to redbook.

 

3. Some people have ruled un-decoded MQA as terrible or sonically inferior or any number of things based on these measurements.

 

I think this type of test could give is some information.

[Samuel T Cogley]

I haven't received any negative feedback from MQA.

 

I'm here for you guys. I listen to you guys not manufacturers.

 

What I'm getting at is this:

 

Given that-

 

1. MQA marketing claims that un-decoded MQA music is better sounding than standard CD rips.

 

2. Measurements done by some people appear to show un-decoded MQA is inferior to redbook.

 

3. Some people have ruled un-decoded MQA as terrible or sonically inferior or any number of things based on these measurements.

 

I think this type of test could give is some information.

 

If you're doing ABX, I strongly recommend to watch the sample rate. Often times, if a DAC has to change sample rates, there will be either an audible click or a telltale pause, which makes the test the opposite of blind. If your MQA example is 48kHz, upsample the Redbook file to 48kHz before testing.

 

Like I said, the increased noise floor made the MQA file easily distinguishable for me.

[The Computer Audiophile]

If you're doing ABX, I strongly recommend to watch the sample rate. Often times, if a DAC has to change sample rates, there will be either an audible click or a telltale pause, which makes the test the opposite of blind. If your MQA example is 48kHz, upsample the Redbook file to 48kHz before testing.

 

Like I said, the increased noise floor made the MQA file easily distinguishable for me.

 

The pop or tick shouldn't matter unless it pops only for higher or only for lower sample rates. If all sample rate switches create a pop or if it's random, then it won't be an issue with only two tracks. It should pop every time.

[Samuel T Cogley]

The pop or tick shouldn't matter unless it pops only for higher or only for lower sample rates. If all sample rate switches create a pop or if it's random, then it won't be an issue with only two tracks. It should pop every time.

 

With all due respect, it doesn't sound like you've done a lot of mixed sample rate ABX.

 

If it doesn't click or pop, you got the same file as last time, invalid test. The score after the test is the whole point of ABX. If you get A two times in a row, or B two times in a row, you'll know. When it does click or pop, you'll know you got A. Or vice versa.

[mansr]

I haven't received any negative feedback from MQA.

 

I'm here for you guys. I listen to you guys not manufacturers.

 

What I'm getting at is this:

 

Given that-

 

1. MQA marketing claims that un-decoded MQA music is better sounding than standard CD rips.

 

2. Measurements done by some people appear to show un-decoded MQA is inferior to redbook.

 

3. Some people have ruled un-decoded MQA as terrible or sonically inferior or any number of things based on these measurements.

 

I think this type of test could give is some information.

MQA undecoded has a resolution of 15 bits. This is an indisputable fact. If this is indistinguishable from, let alone better than, standard CD audio, why do we need high-res at all?

[The Computer Audiophile]

MQA undecoded has a resolution of 15 bits. This is an indisputable fact. If this is indistinguishable from, let alone better than, standard CD audio, why do we need high-res at all?

 

This is 100% beside the point.

 

Please take the test to see if you can determine a sonic difference between the two. If you can't, dare I suggest your complaining about 15 bits is misplaced.

[The Computer Audiophile]

With all due respect, it doesn't sound like you've done a lot of mixed sample rate ABX.

 

If it doesn't click or pop, you got the same file as last time, invalid test. The score after the test is the whole point of ABX. If you get A two times in a row, or B two times in a row, you'll know. When it does click or pop, you'll know you got A. Or vice versa.

I haven't done it in a long time.

You're totally right. Hopefully the pop can be controlled or it's random. If not, it's a no-go.

 

Thanks for the follow up.

[mansr]

This is 100% beside the point.

 

Please take the test to see if you can determine a sonic difference between the two. If you can't, dare I suggest your complaining about 15 bits is misplaced.

 

It is not beside the point at all. If 15 bits at 44.1 kHz is so damn great, what the hell do we need high-res for?

[The Computer Audiophile]

It is not beside the point at all. If 15 bits at 44.1 kHz is so damn great, what the hell do we need high-res for?

 

Perhaps we don't.

 

You take the test yet?

[mansr]

Perhaps we don't.

 

You take the test yet?

No, but would anyone care about my results if I did?

[The Computer Audiophile]

No, but would anyone care about my results if I did?

 

Yes, because you have a big issue with MQA being 15 bit and this could really bolster your argument that it's inferior to redboook

[mansr]

Yes, because you have a big issue with MQA being 15 bit and this could really bolster your argument that it's inferior to redboook

Why would I need to bolster that argument? It's a simple matter of fact. Whether or not the difference is audible to you depends on your system and your ears.

[The Computer Audiophile]

Why would I need to bolster that argument? It's a simple matter of fact. Whether or not the difference is audible to you depends on your system and your ears.

If it isn't audible, then why get so uptight about 15 bits versus 16?

[STC]

If it isn't audible..........

 

 

+1 If it isn't audible, it doesn't matter.

[jmudrick]

+1 If it isn't audible, it doesn't matter.

How MQA sounds seems a distraction for some here.

 

Sent from my Nexus 5X using Tapatalk

[bcwang]

Don't worry so much about sample rate changes. A lot of the 2L mqa tracks are based on 44khz sample rate since dxd is 352khz. So compare that to the cd file and it won't change sample rates.

[agladstone]

I've been listening to the MQA Playlists that I've made in Tidal to listen to on my MQA DAC at home in my car (in my car I use an audioquest dragonfly v 1.2) and I've compared many of the albums to the non MQA versions in my car, and to me, the MQA versions seem to sound better (most of them).

So, just some personal experience/ input in regards to this thread if it helps at all?

[STC]

How MQA sounds seems a distraction for some here.

 

 

It is not a distraction to me. But I am very curious to know why it sounds different. Is it possible for 1 bit difference to contribute higher noise floor? Does noise floor have some kind of effect to our preference? Are we now coming to an agreement that 15 bit is enough? Was Philips correct to say that 14 bit over Sony's 16 bit was enough?