MQA reviewed at Audiostream

[Octavius]

If you can stream HD video, you can stream DSD256.

 

I guess you can but it would take circa 8x the bandwidth to stream MQA so it would be greatly inefficient. And not all have a good enough internet connection to allow for a seamless rendition of DSD 256 or DXD or have capped data plans, want to do it on the go, etc

[master]

I guess you can but it would take circa 8x the bandwidth to stream MQA so it would be greatly inefficient. And not all have a good enough internet connection to allow for a seamless rendition of DSD 256 or DXD or have capped data plans, want to do it on the go, etc

 

Duh!!!

 

Music isn't the only thing streamed you know. YouTube requires higher b/w. Ditto for Netflix, Hulu, Amazon Prime, HBO Go, etc. And every one of these video streaming services has a lot more paying customers than any music streaming service. B/w (or it's lack) isn't the real reason music streaming is yet to turn a profit nor for the lack of HD audio streaming.

[MikeJazz]

What I do find odd is some people being firmly against MQA without listening to it properly.

 

I don't have to ear it to be against the principals in which MQA is based.

It's based on "closed ecosystem" principle that will enable MQA to have us "trapped"...together with content owners, hardware makers and streaming parties...it's just a stupid idea.

 

I hate the idea.

And I will continue to hate the idea even if the resulting sound is good.

[Octavius]

Duh!!!

 

Music isn't the only thing streamed you know. YouTube requires higher b/w. Ditto for Netflix, Hulu, Amazon Prime, HBO Go, etc. And every one of these video streaming services has a lot more paying customers than any music streaming service. B/w (or it's lack) isn't the real reason music streaming is yet to turn a profit nor for the lack of HD audio streaming.

 

Be that as it may much lower bandwidth use will certainly not hurt, will it?

[Octavius]

I don't have to ear it to be against the principals in which MQA is based.

It's based on "closed ecosystem" principle that will enable MQA to have us "trapped"...together with content owners, hardware makers and streaming parties...it's just a stupid idea.

 

I hate the idea.

And I will continue to hate the idea even if the resulting sound is good.

 

There are very successful closed ecosystems.

 

Being closed or open isn't a defining characteristic for the success of an ecosystem.

 

Unlike you my first concern is Sound Quality.

And that IME is clearly assured with MQA.

 

PS it's important to not forget that for anyone who buys an MQA capable DAC there are also the usual PCM and in some cases DSD alternatives. I just see MQA as a free bonus that can become truly important.

[master]

Be that as it may much lower bandwidth use will certainly not hurt, will it?

 

With MQA it will. It's DRM anyway you look at it.

 

I've already mentioned (in another thread) FLAC lets me do any number of things like streaming to multiple rooms with any equipment of my choosing, using any h/w or media player, streaming out of home (both lossless and lossy), etc. and a lot many other things.

 

I don't see any of that happening with MQA unless I'm willing to invest in multiple MQA DACs for starters. Now if they were a monopoly like Blu-ray and 4K Blu-ray they'd be able to force people to buy new h/w to decode the latest formats. Fortunately or unfortunately (depending on which side you are standing) they are not a monopoly and certainly don't have the deep pockets or even the clout of Sony to force something like that.

[Octavius]

With MQA it will. It's DRM anyway you look at it.

 

I've already mentioned (in another thread) FLAC lets me do any number of things like streaming to multiple rooms with any equipment of my choosing, using any h/w or media player, streaming out of home (both lossless and lossy), etc. and a lot many other things.

 

I don't see any of that happening with MQA unless I'm willing to invest in multiple MQA DACs for starters. Now if they were a monopoly like Blu-ray and 4K Blu-ray they'd be able to force people to buy new h/w to decode the latest formats. Fortunately or unfortunately (depending on which side you are standing) they are not a monopoly and certainly don't have the deep pockets or even the clout of Sony to force something like that.

 

So vote with your wallet and don't buy it. Simples.

[Cycleman]

There are very successful closed ecosystems.

 

Since you put it in the plural, name at least two of these closed ecosystems.

 

Be specific, please, and if you can, please define "very successful."

 

Dave, who is regularly confounded by the general and inexact usage of language

[Octavius]

Since you put it in the plural, name at least two of these closed ecosystems.

 

Be specific, please, and if you can, please define "very successful."

 

Dave, who is regularly confounded by the general and inexact usage of language

 

I can rephrase if you want for the sake of being rigourous. There is at least one very successful closed ecosystem in the consumer space.

 

I'm sure you'll know the one I'm thinking about.

 

PS Sorry if my inexactness confounded you somehow.

 

PPS I truly like your signature!

[Miska]

I heard MQA files in a non MQA DAC and they didn't sound worse to my ears than the redbook version.

 

At least based on objective measurement and analysis it doesn't seem to be able to match RedBook.

 

Please note that MQA can go up to 24/352.8, so a 24/192 DAC would not be capable of taking full advantage of MQA.

 

Again, based on objective analysis of decoded MQA it is severely lossy compared to original 352.8/24 content. The original 2L content contains frequencies up to 50+ kHz while decoded MQA output reaches about 30 kHz and has noise floor higher than the original content. So there's certainly nothing to be gained there.

 

Perhaps more importantly my knowing about Schiit's previous failure with DSD made their conservative stance somewhat predictable to me.

 

I don't care much about Schiit's failure or non-failure. I make my own independent objective measurements of the performance.

[Miska]

About MQA I think its big advantage is having made Hires streaming feasible. I'm hoping Tidal or some other streaming service include MQA in their lossless offer at no additional cost.

 

1) MQA is not lossless, there is no problem seeing severe cut of high frequency content*

2) It is less efficient for streaming than standard FLAC of equivalent resolution

 

 

*) For example for the 2L encodes, it loses about 20 kHz worth of high frequency content, which IMO is quite a lot

[Miska]

I'm rooting for MQA as its SQ is IMO at least on pair with DXD or DSD 256 and it's technically possible to stream it.

 

Certainly not anywhere near that. Based on my measurements, it is practically roughly equivalent of 60 kHz PCM with 17-bit resolution. DSD64 does better than that!

[Octavius]

Certainly not anywhere near that. Based on my measurements, it is practically roughly equivalent of 60 kHz PCM with 17-bit resolution. DSD64 does better than that!

 

Your measurements are certainly not capturing the whole thing.

 

Have you had the chance to do a proper listening comparison with a good MQA DAC in a good setup?

[Octavius]

1) MQA is not lossless, there is no problem seeing severe cut of high frequency content*

2) It is less efficient for streaming than standard FLAC of equivalent resolution

 

 

*) For example for the 2L encodes, it loses about 20 kHz worth of high frequency content, which IMO is quite a lot

 

Regarding 1) and MQA's measurements in general have you seen this?

 

Inside MQA | Stereophile.com

 

Regarding 2) why do you say that?

[Miska]

Your measurements are certainly not capturing the whole thing.

 

What is the thing that it is not capturing?

 

From time domain and temporal resolution point of view, DSD is much better. DXD aka 352.8/24 was created to retain most of DSD64's performance while reducing amount of required processing resources, and also due to other practical reasons such as one byte of DSD data resulting in one sample of DXD data (making it easier for audio editor to handle the content due to avoiding sub-byte/byte-crossing operations).

 

Do you have objective technical data to demonstrate something I'm not capturing and definition how I can reproduce the results in my lab?

 

Have you had the chance to do a proper listening comparison with a good MQA DAC in a good setup?

 

Yes, but listening impressions are personal and subjective. I much prefer true hires upsampled to DSD256/DSD512.

[Octavius]

What is the thing that it is not capturing?

 

From time domain and temporal resolution point of view, DSD is much better. DXD aka 352.8/24 was created to retain most of DSD64's performance while reducing amount of required processing resources, and also due to other practical reasons such as one byte of DSD data resulting in one sample of DXD data (making it easier for audio editor to handle the content due to avoiding sub-byte/byte-crossing operations).

 

Do you have objective technical data to demonstrate something I'm not capturing and definition how I can reproduce the results in my lab?

 

 

 

Yes, but listening impressions are personal and subjective. I much prefer true hires upsampled to DSD256/DSD512.

 

About measurements please see my follow up post #39.

 

What was the DAC and the rest of the setup that you used to do the comparison?

[Miska]

Regarding 1) and MQA's measurements in general have you seen this?

 

Inside MQA | Stereophile.com

 

Yes I have, and it is incorrectly comparing 24-bit versions to the MQA ones which certainly don't have 24-bit worth of dynamic range. And the source recording doesn't have much notable amount of content above 20 kHz so higher sampling rates are mostly wasted. When you make equivalent bandwidth and bit depth reductions and produce a standard FLAC it is smaller than the MQA version.

 

192/18 content created from the 352.8/24 master, fits into smaller FLAC than the MQA version - you can see content reaching about 56 kHz:

 

Then decoded MQA version:

 

You can notice content disappearing in noise by about 30 kHz and also severe image/alias distortion in the 60 - 90 kHz band. Both recorded from analog output of the same MQA capable Meridian DAC.

 

I have more results than this and I will make more recordings and analysis, no problem.

 

Regarding 2) why do you say that?

 

Because creating a standard FLAC with equivalent resolution to the MQA results in a smaller file. This is largely because when MQA is put into FLAC container, for FLAC encoder the MQA origami stuff is uncompressible noise.

[Octavius]

Yes I have, and it is incorrectly comparing 24-bit versions to the MQA ones which certainly don't have 24-bit worth of dynamic range. And the source recording doesn't have much notable amount of content above 20 kHz so higher sampling rates are mostly wasted. When you make equivalent bandwidth and bit depth reductions and produce a standard FLAC it is smaller than the MQA version.

 

192/18 content created from the 352.8/24 master, fits into smaller FLAC than the MQA version - you can see content reaching about 56 kHz:

[ATTACH=CONFIG]26548[/ATTACH]

 

Then decoded MQA version:

[ATTACH=CONFIG]26549[/ATTACH]

 

You can notice content disappearing in noise by about 30 kHz and also severe image/alias distortion in the 60 - 90 kHz band. Both recorded from analog output of the same MQA capable Meridian DAC.

 

I have more results than this and I will make more recordings and analysis, no problem.

 

 

 

Because creating a standard FLAC with equivalent resolution to the MQA results in a smaller file. This is largely because when MQA is put into FLAC container, for FLAC encoder the MQA origami stuff is uncompressible noise.

 

Sorry for being insistent but don't the measurements made by Stereophile and the one provided by Bob Stuart show that the Origami works as advertised and that the actual resolution is the same as 24/88.2 and not 17/60 as your measurements led you to believe?

[Miska]

Sorry for being insistent but don't the measurements made by Stereophile show that the Origami works as advertised and that the actual resolution is the same as 24/88.2 and not 17/60 as your measurements led you to believe?

 

No, because the source content Stereophile has doesn't really have notable output above 30 kHz. And because the source content is nowhere near dynamic range of 24-bit, so comparing 24-bit source file size to MQA source file size is completely moot.

 

Plus I want to be able to reproduce the measurement results. I don't need or want to trust anybody else.

 

Wonder why they didn't post MQA version of the Fig.1 for comparison?

 

One of the major problems with MQA is that I cannot encode my own test signals with MQA for measurements. They are prohibiting such on purpose. That's why Stereophile can also encode only signal approved by MQA and not anything they please on their own.

 

With any other normal codec, be it competing ISO/IEC standard MPEG-4 SLS or similar, you can get full reference source code for the encoder and decoder plus full description of the process and encode/decode any measurement signals you please to verify the functionality without restrictions. But not with MQA! It's a secret black box and you need to ask them to encode something for you.

[Octavius]

No, because the source content Stereophile has doesn't really have notable output above 30 kHz. And because the source content is nowhere near dynamic range of 24-bit, so comparing 24-bit source file size to MQA source file size is completely moot.

 

Plus I want to be able to reproduce the measurement results. I don't need or want to trust anybody else.

 

Wonder why they didn't post MQA version of the Fig.1 for comparison?

 

One of the major problems with MQA is that I cannot encode my own test signals with MQA for measurements. They are prohibiting such on purpose. That's why Stereophile can also encode only signal approved by MQA and not anything they please on their own.

 

With any other normal codec, be it competing ISO/IEC standard MPEG-4 SLS or similar, you can get full reference source code for the encoder and decoder plus full description of the process and encode/decode any measurement signals you please to verify the functionality without restrictions. But not with MQA! It's a secret black box and you need to ask them to encode something for you.

 

I can surely see that that can be a downside for you.

 

For someone like me, who just cares about SQ, it being a black box is not relevant.

 

About measurements I really just care about what my ears can do. In that department MQA has passed with flying colours.

[Miska]

For someone like me, who just cares about SQ, it being a black box is not relevant.

 

You are not worried about future accessibility of your content investment? For example 10 years from now?

 

About measurements I really just care about what my ears can do. In that department MQA has passed with flying colours.

 

That's your personal subjective choice, which is complete fine. But it cannot be generalized and I'm not sure (based on information I have) how exhaustive it is within all available possibilities.

 

Did you compare downloading original 352.8/24 recordings upsampled to DSD512 for example?

 

I personally want to run digital room correction which makes much bigger difference than MQA and MQA is doing it's best to prevent such.

[master]

For someone like me, who just cares about SQ, it being a black box is not relevant.

 

You are not convincing in the least bit. At least not in terms of SQ.

 

You certainly are entitled to prefer MQA, nothing wrong about. Just don't make it sound like its the best thing since sliced bread especially for SQ.

 

PS: Anyone gung ho on SQ, it usually begins and ends with vinyl... maybe some DSD. Certainly not a lossy format like MQA.

 

Just my 2 cents in the vain hope you'll at least put forth some valid arguments... instead of simply arguing for arguments sake.

[Octavius]

You are not worried about future accessibility of your content investment? For example 10 years from now?

 

 

 

That's your personal subjective choice, which is complete fine. But it cannot be generalized and I'm not sure (based on information I have) how exhaustive it is within all available possibilities.

 

Did you compare downloading original 352.8/24 recordings upsampled to DSD512 for example?

 

Not really. I'm waiting for a streaming service to take up MQA (Tidal seems to be in the pole position) and besides that I only have a few albums in MQA format.

I look at MQA as an almost free bonus should it become a real success.

 

Never heard DXD upsampled to DSD 512. Nor am I too keen to do it as I'm not in interested in spending what would be needed in terms of software and hardware resources.

[Octavius]

You are not convincing in the least bit. At least not in terms of SQ.

 

You certainly are entitled to prefer MQA, nothing wrong about. Just don't make it sound like its the best thing since sliced bread especially for SQ.

 

PS: Anyone gung ho on SQ, it usually begins and ends with vinyl... maybe some DSD. Certainly not a lossy format like MQA.

 

Just my 2 cents in the vain hope you'll at least put forth some valid arguments... instead of simply arguing for arguments sake.

 

Hey everyone's entitled to their opinion.

 

Mine is truly positive regarding MQA's SQ based on what I've heard for myself.

 

Yours is negative. Fine.

 

But don't go on saying I'll perhaps end up with vinyl. I certainly understand the fascination but it's clearly not for me.

 

About DSD it sounds great but so does hi resolution PCM ...

[Miska]

I look at MQA as an almost free bonus should it become a real success.

 

Where is the free bonus part? They want license fees for encoding, streaming and decoding.

 

Instead standard FLAC is truly free and achieves the same or better result. No license fees involved and anybody can independently implement encoder or decoder. To me that sounds like a future-proof solution to store content.

 

Never heard DXD upsampled to DSD 512. Nor am I too keen to do it as I'm not in interested in spending what would be needed in terms of software and hardware resources.

 

I already have a computer and DACs. I anyway keep my computers up to date and my computers provide much more processing capability per Euro than I have seen in any DAC. I'm not interested spending extra money to buy a new DAC to get something like MQA support especially when it means limiting possibilities on doing digital room correction and stuff. But I obtained MQA capable DAC just for the fun of measuring it, otherwise I wouldn't bother because it is technically sub-par.