"Editor's Note 1: MQA ltd was sent a copy of this article several days prior to the scheduled publication date. The company requested a phone conversation, which took place earlier this week. MQA was encouraged to write a response for inclusion with the article below, but it respectfully decline to submit a formal response.
Editor's Note 2: The author of this article is writing under a pseudonym. While he is unknown to the readers, his identity has been verified by Computer Audiophile. He has no vested interest in the audio business, other than being a consumer of music.
Editor's Note 3: The technical assertions made in this article have been thoroughly checked by independent engineers, both in and out of the audio industry. To the best of our knowledge everything technical in this article is factually correct and may be duplicated at any time by anyone with the requisite skills."
I'm not surprised that virtually all music providers or transmitters will jump in. They are forever looking for some 'magic' to boost sales. DVD-A, SACD, were supposedly leaps better than CD. They had a run. Now it's this new 'thing'. I admit to being cynical about this process. "Ultra refinement" when most consumers listen on $20 'buds' ???
I've been communicating with Robert Harley from The Absolute Sound. The "MQA's unexpected twist" means that the MQA company has loosened up the requirement that the "unfolding" be done in the DAC hardware. So, if you uncheck the "MQA passthru" option on Tidal, the desktop app will "unfold" the MQA files and present an up to 94/24 file to your conventional DAC.
But apparently that will not provide you with all the benefits of MQA, specifically the reduction in digital "blur", which seems to only be able to be done in the (MQA certified) DAC.
In my case, when I've compared albums on Tidal which are offered in MQA and "regular" FLAC versions, playing thru my PS Audio DirectStream DAC, the MQA versions which are "unfolded" by Tidal sound a bit "dull" compared to the "regular" FLAC versions. I've tried that with about a half dozen albums, same result. I wonder if there's something with the FPGA design that causes that?
So, with the news that Universal has joined Warner and Tidal, I'm wondering if that's the news the DAC manufacturers were waiting for, to incorporate MQA into their future hardware designs....or, could it be that DAC chipset companies like ESS will put that feature into their future offerings?
I think this year will be make or break for MQA. I still feel MQA needs non USB Dacs in the $300- $800 range. You either have a $299 USB or $1900 price range.
From Absolute sound report on Rocky Mountain audio fest and I quote Most significant :MQA is no longer at the cheerleader stage it’s really happening Warner is on board with two other majors as well and many independent labels have signed on as well. Processors to the affordable nad to the astronomical msb are on board. Best of all A/B comparison were running all day long troughout the show. Demonstrating to most as well as me don’t know who me is some say RH. THAT ITS IMPROVEMENT TO BOTH HD AND STANDARD RESOLUTION IS SUBSTANTIAL. Why are you guys so butt hurt over this. So your telling me all these people are in on a scam and it's just away to resell music You say you have to buy all your music over again. Not true so none of you guys stream tidal 20 bucks a month. I bet some of your wire cost thousands. And a MQA dac 200 bucks. From what I’m hearing it is impressive. Why all the Debbie downers hear. If you can’t hear the dif sucks for you. I do so why bash the people that just want better sound. It’s almost like you work for dac companies that have not come on board yet. I just don’t get the better the better than thou attitude.
Got an email from Tidal this morning, they are rolling out the first MQA streaming on their desktop app (I presume they don't support it with the browser version). They are branding this "TIDAL Masters".
I have a PS Audio DirectStream DAC, which from what I read doesn't do well with MQA. So this weekend I'll make my way down to my local shop and see what MQA sounds like with the Explorer2 DAC...for $199 worth getting one, to see what all the fuss is about MQA.
Thanks for your feedback Erik. I shouldn't have made any comparison to HDCD as it was not designed to facilitate transmission by reducing file size. My only experience with HDCD is with Reference Recordings,which I've enjoyed.As I think more about this topic I foresee less need for MQA as I feel download speeds will make it less relevant.
Yes, this was a big problem when DSD first came out. The DSD versions were noticeably different in energy across the audible frequency spectrum. It can also happen in reverse. As Bob Carver advertised, a lot of early CD's were heavily compressed, not just in dynamic range but also stereo separation compared to LPs. A lot of the "bad sound" of digital could be directly attributed to really bad mastering engineering and practices.
We also tend to hear any difference as improvement. Much like haircuts. :)
So the idea is to compensate for the original A/D step. This is done by measuring the blurring of the AD converter itself, and then applying a compensating algorithm to the Digital multi-track masters. As far as i know, the process for going from analog to digital is to first convert each track individually to digital. So whether or not we start from analog or digital doesn't really affect the process.
Only after multi-track masters are fully digitized can re-mastering begin.
Of course, some recordings are far simpler. A lot of great 1950-1960 orchestral recordings were multi-track, but by multi we mean 3.
Lots of great info here, thank you everyone for what has been contributed.
In my travels, one of the audio salesmen I encountered said that "remasters" oftentimes were the reason why newer versions of the same old music sounded better, especially when they could go back to the old "analog" tape. If that is true, I suspect what he's saying is that going from the original analog multi-track source to the now new original digital would make that conversion much better (with a newer analog to digital converter in the path). He also said that starting out with original multi-track digital source, to digital master could also be an improvement, but (in his opinion), not as much of an improvement as starting out with an analog multi-track source in the first place.
Thinking of MQA, I'm guess something similar applies? If the engineer can mount an original analog multi-track tape and master from there, the first analog to digital conversion would have the advantages of MQA, with the "de-blurring" happening there? Would there be some sort of similar MQA benefit when the original multi-track recording was itself a digital source?
Hi @ptss You asked a question I did not fully answer.
How is MQA superior (significantly) to HDCD
At least on spec, HDCD compresses dynamic range, in addition to providing other frequency dependent tools. That is, the equivalent of more bits in a CD.
MQA attempts to pack a 384k/24 signal into 48k/24 and improve the overall fidelity by carefully matched de-blurring filters on the AD and DA process. So it's benefits are to file/stream size as well as fidelity.
However, this is all specmanship. Sadly I have heard no improvement in MQA recordings, and don't have an HDCD capable player either so I am unable to render a subjective comparison. It kind of bugs me that there isn't an HDCD software decoder for me to use. :)
@ejr1953 That's pretty much how I understand it, based on MQA and Benchmark data.
How Warner Bros. will do this I have no idea. I know MQA provides a "generic" AD de-blur setting in case the original converter data is not available, but as I understand it, once the tracks are mixed together there's now way to de-blur that.
To avoid confusion, there are also de-blurring, or compensating steps at the DAC end too. It's easy to confuse them together. :)
The "re-mastering" comment is (to my way of thinking), key. If I understand the advantages of MQA, it's that the removal of "blur" requires that they go back to the "analog to digital" point in time and fix it at that point, so when the DAC does the "digital to analog" conversion, it's somehow "linked" to that earlier point in time...in the way the DAC does its job. It would seem to me that the only way to do that would be to "re-master" the music. I suspect that "re-mastering" is not cheap, so I wonder if the "business case" might preclude most labels from producing much MQA...especially since the majority of people of OK with MP3 on their phones, the hi-quality audio market is just a small sliver of the overall music consumption world.
Well, read Benchmark's blog on the matter. To them it IS a seriously flawed system. Their points are accurate too, but I'm not sure about audibility.
Personally I was really excited about the "digital origami" idea, when I thought it was lossless.
The good news is that if this is a good idea, then undoubtedly some one will produce an open-source version of the same idea without the flaws or licensing requirements. Personally I would love to have my music collection take 1/4th it's current size if it could be done without any loss of quality.
I hope MQA never gets off the ground. It's not that the format is flawed, just the people who created it.
The original plan was to also allow streamers to unpack it and send high res to the DAC of your choice. Since they now chosen to remove the ability Im no longer interested.
MQA is bad for music lovers! Let's let Warner bros know .
One
thing that I should point out to anyone comparing MQA, is that to do it
right re-mastering is required. Meaning you have to work your way back
to the original multi-track digital files before you can produce a 2
channel MQA product ... when I have heard significant differences what I heard seemed
due to this re-mastering and the choices taken at that time
This is a very important point. If you're comparing two versions of a recording, but they are taken from different masters, you really can't make any valid conclusion about things such as whether MQA benefited one of them. There are too many variables.
One thing that I should point out to anyone comparing MQA, is that to do it right re-mastering is required. Meaning you have to work your way back to the original multi-track digital files before you can produce a 2 channel MQA product. Becnhmark Media has a lot more on this.
In any event, when I have heard significant differences what I heard seemed due to this re-mastering and the choices taken at that time. If you've ever heard good re-masters you know what I'm talking about.
HDCD is the perfect analogy. HDCD is a pretty complicated format as well. Microsoft bought HDCD but then did nothing at all with it. It's a shame because I would love to have a software scanner to go through all my FLAC files, and decode any HDCD content.
MQA is not involved in DRM either. There is no playback, copy or encryption restrictions on the files or the decoding. Just like HDCD however, you must have compatible hardware. MQA IS attempting to be an authoritative standard so when your MQA light goes up you know you downloaded the file without additional alteration/compression done by the download service.
HDCD was a recording engineer's toolbox. It was up to the engineering team to decide which features, and when and how much they would take advantage of. This is a little different from MQA as everything is done by the hardware vendors. The recording engineers don't really have any control over MQA besides turning it on or not.
Reference Recordings uses HDCD (developed by Keith Johnson, also of Spectral fame,and Pflash Pflaumer who is the technical wizard of Berkely Alphac Dacs) which is good enough that Microsoft bought it. How is MQA superior (significantly) to HDCD. At least HDCD was not involved in DRM.
You bring up a very good point. MQA isn't just competing against high resolution data files, it's competing against the latest crop of DAC's. Over the past 5 years or so the performance of DACs with Redbook has markedly improved while the performance at super-high res has remained more or less constant.
Aside from the technical details, I'm guessing that if MQA doesn't gain "traction" in the marketplace by the end of next year, it'll go the way of SACD/DSD. I'm reading that Warner Music is working on converting many of the albums in their catalog to MQA, I hope that's accurate. But I'm also reading that Tidal has put "on hold" their plans to stream MQA. Not the greatest news for the format. I must say, my AIFF files (most 44.1/16, some higher res) decoded by my PS Audio DirectStream DAC sounds pretty darn good! It would be a "hard sell" for me to abandon that DAC, especially if I needed a "microscope and tweezers" to find the improvements that MQA offers.
@tomcy6 No worries. The truth is MQA is a very ambitious encoding format and it's easy to misread how parts of it fit together. I went a long time thinking it was lossless, until Benchmark set me straight.
Yeah Erik, I think I misread the Stereophile measurements completely. I now believe all three filters are without MQA so I deleted all my posts so that I don't spread misinformation.
I take it back, I have a good analogy for the SR filters. The SR Filter on the Brooklyn is like wearing the wrong glasses. Everything sounds like it's in the wrong place and disconnected. That's just me though, make yourself happy!
Interesting. However, I’m afraid my experiences correlate exactly with Paul McGowan at PS Audio as well as most members of the SFAS who went to the local demo. Again, if we have to carefully look for "improvements" it’s not worth buying. If this were a Dolby B or dbx demo in the time of magnetic tape it would have been no contest. We would all be wowed. MQA? Meh, and double meh!
Just to add, since my system uses custom speakers and amps, I used AKG K712 headphones for critical listening with MQA sources. I compared native 24/96 to MQA. I didn’t care to evalutate an MQA file with/without MQA decoding, that comparison doesn’t mean anything to me.
Hi @tomcy6 What may not be clear is that the minimum phase filter IS selectable without MQA. Let me try to summarize the behavior (which jives with the text of the review):
- With MQA detection set OFF you have 3 filters to choose from, including minimum phase (MPH). The other two are conventional Fast Rolloff (FR) and Slow Rolloff (SR) - With MQA detection set ON the Brooklyn forces the filter to be minimum phase, whether or not the input is MQA encoded.
There’s a little bit about that behavior in the body of the review, and I’m glad I read it because I was going nuts trying to get to the filter selection menu. :)
My only point is that the specific digital filter in the Brooklyn is independent of MQA. Personally my hierarchy (though speakers) is:
FR > MPH > SR
I find FR and MPH very close, with FR having more clearly delineated attacks. SR is just too funky to listen to. Maybe it’s because my tweeters go to 30kHz or 40kHz and I’m getting more effects in the audible range. Don’t know, but don’t like it, I can’t quite explain the weird discomfort.
In any event, I encourage everyone to trust your ears and buy only those things that are worth it for you. :)
There is some recent MQA discussion on audio asylum. They did do an hi-rez/MQA demo of the same song at RMAF with an emphasis on reducing temporal distortion. The result of the demo was, of course, forgone, and the difference was quite large.
Even though the news/info is disappointing with regards to hopes for improved fidelity; thank you for passing along the info you learned from your observations.
Well, I know the answer about sound quality already. :)
If I have to look for an "improvement" with a microscope and tweezers it's not worth it to me. :)
Remember Dolby A, B and C? That was worth paying money for. Differences were clear to anyone. MQA.... meh. I can't see myself on my deathbed going "OH, if only MQA had survived, my audio life would have been so much better."
MQA does in fact require the use the apodizing filter. What MQA adds is decoding of the high resolution data as well as their claims to time domain fixes. MQA attempts to compress 384k data (in a lossy way) into a 48kHz signal.
Think of MQA more like FLAC vs. MP3 sort of kind of. :)
The best independent critique I know if how it works is at Benchmark media’s site.
In the case of the Brooklyn, when you have MQA detection and decoding enabled it will force you to use the apodizing filter. If you disable MQA you can pick from 3 different filter types.
PS Audio said they were not at all impressed by it. Benchmark has a long blog post on the matter and just how convoluted the entire process really is.
Meh, the Brooklyn is a very good dac with or without MQA, but I was looking forward to finding yet another veil lifted... or some other ridiculous thing like that. Nothing.
The whole promise of MQA to labels was to give them a way NOT to release music in high-res equivalent to studio master, and still get some improvement over CD quality just enough to make people buy the same music again. This is actually good that it didn't fly- otherwise we won't see any more high-res releases. As to apodizong filter with no pre-ringing, that type of filter is available in some DACs independently of MQA.
The big question will be if the market goes for MQA. Without a large enough catalog of music, I don't see the DAC makers investing in making their units MQA-compatible.
The MQA folks have said that non-MQA-compatible DACs will be able to play back their files (as they are in a PCM wrapper), but from the few articles I've read on the subject, it seems that FPGA-based DAC manufacturers report that the resulting sound quality is not good.
I guess time will tell. But it seems that there was a real flurry of excitement about MQA earlier in 2016, which seems to have calmed down.
I used a Mytek Brooklyn. Yes, MQA was properly identified when played. My sources came from 2l.no
In my case I comapred 96/24 to MQA and I honestly could hear no difference at all. Not a bad thing considering MQA is a lossy and complicated format to reconstruct. Still, it has no value for me at all.
The Brooklyn IS a very good DAC, and to my ears, plays Redbook much better than most DACs from even 5 years ago. MQA is currently disabled so I can play with other filters. There is an odd bit of interface troubles. If you enable MQA you can't change filters. :-)
Hi Eric Would you be able too describe the decoder you were using or heard and if the recorded material was identified by the decoder/player as being MQA files. I was hoping that the MQA format would result in overall better recordings, not that I am anxious to start over or replace a lot of my current media. Even though my best source is my Turntable, I am not a advocate of any one particular media. I believe that all source equipment are recording dependent. I do believe that if all steps in the analogue process are of the highest caliber, that vinyl has the advantage, but with this in mind, I have a lot of poorly recorded LP's and superbly recorded CD's / DSD/SACD. I was disappointed to hear your observation of MQA files, I was hoping that process would improve the overall sonic quality of recordings and be a reliable source for quality recordings. I guess I will have to listen for myself, but do value your opinion
Tidal sale is over, Apple was just conducting competitive research to see if they posed a threat to iTunes. It didn't so they had no reason to buy it.
Tidal hasn't done anything on MQA since then.
Stereophile ran an article online a month or so ago explaining that it was hibernating, but that Warner (or some other label) was busy converting their entire back-catalog to MQA.
I've listened to several MQA recordings and find them no better or worse than redbook. Claims of it improving anything in a recording are, as far as I am concerned snake oil and mysticism.
You must have a verified phone number and physical address in order to post in the Audiogon Forums. Please return to Audiogon.com and complete this step. If you have any questions please contact Support.
