Basically I can summarize in three points:
Lossy compression to save bandwidth (we don’t have a bandwidth issue).
‘De-blurring’ or improvement in impulse response at the expense of aliasing effects and IM distortion at high frequency due to use of leaky filters in both ADCs and DACs.
Hardware locked in ‘DRM’ requires purchase of MQA DAC to get the best performance.
Another day, another MQA thread…
Why can’t I take a nice shiny 24/192k master and put it through a ‘slow leaky filter with IM etc’ and get the same or better sound as MQA? Obviously, the resulting audio would have more bits than an MQA file (as MQA is lossy) but why wouldn’t this work and has anyone reverse engineered this type of filter?
Yes, it will be even better because the native 192k sampling is 2x compared to MQA core at 96kHz. The aliasing effect will reduce significantly. The aliasing happens at 10k and 15kHz in the audio range is at -40dB which softer part of music exists when using leaky MQA filter.
For IM distortion it is dependent on the type of digital filter used. Linear type don’t exhibit increased in distortion at high frequencies.
With high res the need for filters almost dissapears so don’t do anything to it all is the best solution.
Bob needed a bandwidth problem and “deblurring” to be able to lock and license the solution.
The reason I like the idea of MQA is that it looks at the characteristics of the AD conversion and unwinds them in the DA. Original deficiencies taken out. Doesn’t this work? I use filters for coffee…that’s also a leaky filter 
how about the main principle?
That topic is debatable. When is it being done and to what scale? Do you think the mass conversions at the studios actually has someone matching each album and track to the appropriate de-blurring or it’s just a general MQA setting? How could deblurring possibly be applied to multi track digital recordings that could have multiple different ADCs involved?
As for the filter, if someone likes the sound of it, Ayre uses something very similar and doesn’t require licensing and unfolding etc. Or you can use SOX to get pretty close.
I though it was the AD characteristics of the original recording. Yes I thought it was being done using the details of the original AD for each recording, isn’t that the theory? I get that there are cracks but the general principle…
You can’t take away the deficiencies without creating a new set of problems. There’s always a compromise that need to weigh in. In this case they emphasise more on impulse response and compromise the aliasing effects and IM distortion in the audible range in favour of ‘de-blurring’
When aliasing effects and IM distortion happens in the audible range, our ears are likely able to perceive them. Moreover, this type of effects can manifest itself with the music. This result? It may change the tonal quality of the music to some extent. That’s why many people were quick to point out it is effect of the ‘de-blurring’ not the aliasing and IM distortion in the audible range?
Difference does not necessary equate to be better.
But the Ayre filter manages to sound good irrespective of any dubious ADC matching process with anything put through it (according to reviews, I don’t own one).
Ayre built the DAC in the Pono and it sounds fantastic. I think they re-used the tech for their Codex. They are on my short list for an upcoming DAC refresh.
Are you thinking QX-5 Twenty?
Yep.
/10char
Nice! Same here. Saving for it.
There is also the new one QX-8 priced at $5,500 loaded, coming in May, but I doubt it will be as good as the “5”
Ayre uses minimum phase slow roll-off filter to eliminate pre as well reducing post ringing effects that makes it more natural sounding. There’s compromise of using this type of filter as oppose to linear ones; mainly it introduce group delay, increased distortion at high frequencies and aliasing effects get reflected back into the audio range. Most of the off selves DACs, especially latest ESS Sabre and AKM have wide variety of filters options; which include both linear and minimum phase to suite different listening conditions.
More reading…
And interestingly, it only uses quite a ‘cheap’ DAC the ESS 9018, as used by Oppo (RIP) and others in mobile solutions. Pono and what happened to it is actually bang on topic. Their distribution company was bought by Apple and immediately without warning shut down. Neil Young apologised for being forced to charge high prices by the record companies for HD music. With MQA remasters, they now get to sell it all again…
I have no doubt that the 24/192 full resolution music will be pulled from sale and gone forever once MQA takes hold.
This is indeed a sad scenario, that’s what the music industry is going through, they want to be in control, it is like ‘end to end’ DRM. They are fearful of giving their ‘crown jewel’ to the consumers, so before MQA takes over, better save up and buy those lossless Hi-Res masters downloads.
That would be my advice, too.
Puts hand up
I’ll happily take the first unfold in Roon, effectively giving me access to 18/96 kHz material without charge, but I’m not sold on the whole A/D to D/A “deblurring”. In particular I don’t buy the rationale that it can’t be done in software that knows what DAC it is connected to but must be located close to the hardware in the DAC. I suspect that is a commercial restraint and not a technical one.
If I were looking for a new DAC I would want well implemented and isolated Ethernet connectivity, a good clock, discrete R2R ladders for PCM and DSD and a power supply who’s purity would shame an angel (UltraCapacitor banks if possible). In other words a Holo Spring incorporating as much of a microRendu, XSPDIF-2 and LPS-1 as necessary to get from Ethernet to R2R. I wouldn’t want any oversampling or filtering, I can do that externally. In other words, the DAC would do Digital (Ethernet) to Analog conversion and absolutely nothing else, all the DSP grunt work being done in my server.
The idea of doing Network Attached DAC or ‘NADAC’ is pretty amazing. I know Merging Technologies have one that do just that.
Pretty cool but out of reach for most audiophiles. Hope that more manufacturers can come out more affordable ‘NADAC’.
Is there 18bits after the 1st unfold?
I ask because I don’t know but have seen random postings of 14bits and 15bits around the place. I don’t know those people’s sources though.
