MQA - A clever stealth DRM-Trojan (CCC talk)

Still no vested interest, then?

 

And that is the true heart of the matter; DRM.

Did anyone else notice that in the presentation in the OP MQA has built in flags for data encryption?

The whole MQA delivery system is closed end to end with a pay to play.

The presentation also used a nifty term for MQA as "Freemium"; you can get partial decompression (the use of obfuscating terms like origami unfolding is really just compression) for 'free', but to get full decompression and the secret sauce filtering you have to pay by way of using MQA licensed devices.

One Format to Rule Them All.

 

I'm not sure what you are saying here so I will try to rephrase what I wrote previously.

The MQA indicator light comes on based on a flag in position bit 8 of the data stream. Information in bit 8 is deciphered by the MQA logic to tell the DAC whether or not to illuminate the blue light.

There is also a flag in the data stream to tell the DAC what sample rate to display. It must also be embedded within bit 8 to be deciphered. The sample rate that is deciphered represents the sample rate of the master file before it was encoded to MQA. It is does not represent the rate that the DAC is actually decoding.

See the experiments linked in my previous post. The experiment shows that although the MQA light was on, there was no information being unfolded, so it really wasn't hi-res. It could not be hi-res because the values to be unfolded were all zeros but the light stayed on and the display still showed a 352.8khz sampling rate. What was displayed represented the sampling rate of the master before it was encoded to MQA. It did not represent the data from the file being decoded.

MQA does not claim to retain the same resolution. They admit that the sampling frequency above 48khz is a lossy estimation of the original data. Their patent also explains that it is lossy. The base 0-48khz is a lossless sampling frequency but it reduces the bit rate from 24 bits down to 17 or 18 bits.

• It reduces the bit rate to store information in bit 8 for the blue light and I suspect the display and to decide which filters to implement.
  
• It also reduces the bit rate to use the lower bits as a storage container to recreate estimated sample rates for the range of 49 - 96khz in a lossy estimation.

 

I sure do have a pretty large library of high res FLAC files. I hope THE INDUSTRY never finds out about this!

Seriously, pretending FLAC, which is simply a lossless compression scheme for PCM audio, is a format that is in competition with other formats is completely absurd. I'd love to know of even one relevant label that has not released high res music in the FLAC format. The reason the selection of titles has been limited is because the audience for high res music is limited. They haven't held off on releasing the Britney Spears catalog in high-res because they hadn't been properly satisfied with existing formats.

 

Bandwidth is not limited. I can stream 4K HDR video over a wireless connection at home. On a bad day, I can get 33 MBps download speeds on my phone. There isn't anything special about my situation either, I have the normal performance tiers of Comcast and Tmobile. Whatever problems a person feels MQA might be solving, bandwidth is not one of them.

 

Two things here - one is what is displayed on the DAC which is an MQA indicator and the sample rate and word length of the original lossless source file that was MQA encoded, which seems fine to me. The second is an attempt to get to the bottom of what MQA encoding is and isn't - I'm not sure the two need to be conflated.

I edited my post to say "the same audible resolution" - that is, they claim that any audible benefits from the high resolution capture is retained.

I'm not convinced that the efforts to reverse engineer MQA provide any conclusive evidence of chicanery. We already knew it was a perceptual coding scheme that threw out data. The point is, have they managed to get results that match the best lossless high resolution?

Since this question resides firmly in the heavily debated grey area below the 16th bit there will never be a general consensus and for some the fact that this encoding scheme relies on it's performance in the region of a barely perceptible noise floor is enough to make it a scam by definition.

Are there any reports of MQA sounding worse than the PCM source file?

 

Practically speaking, unless Apple and Spotify begin accepting and streaming MQA, it won't get off the ground. Apple will not alter their hardware, in phones, PCs, Apple TV and streaming devices. Not ever.

 

True. It's not as if the hardware manufacturers are falling over themselves to get it incorporated either.

 

Bandwidth has value - more consumption = more money being made. Your 4K video is lossy compressed, so why not your 24/384, too?

 

The flip side of this are some claims that MQA sounds 'better', some of these from those only doing software decompression (or 'partial origami unfolding').

Since the whole folding/unfolding blah blah is just MQA's claimed 'audibly lossless' compression, there is no way this can sound better than the original source file. The claims that MQA 'lite' sounds better is pointing to the source file being different than what is used for a given comparison.

I think it bears repeating (again) that MQA is really two things: the 'audibly lossless' compression AND the temporal filtering/correction 'stuff'.

To date, has there been any true apples to apples comparison?

 

Good for you but anecdotal at best.

Really....everybody? Wow, that for sure isn't a bias just like when I drive to work I always come to the conclusion that everybody in the country own a car and get out at 07:00 each morning. Unless you tell us that they managed to drag into the demo people who had no idea what MQA is it's pretty much like every worthless clinical trial where the population is pre-selected to emphasis the benefit of the drug tested.

 

Consistent with the theme of solving problems that haven't existed in a decade, tell me more about this metered internet concept you speak of .

 

Because there is plenty of bandwidth that 24/192 does not need to be lossy. So why make it lossy?

 

The Relative Cost of Bandwidth Around the World

 

Tell me again about the audio bandwidth problem MQA is trying to solve because neither CISCO nor myself are able to find it.

 

The idea behind MQA is that it sounds just as good (better if you believe the hype) with a lower bitrate, which is more efficient. Service providers don't want to waste bandwidth unnecessarily, not in a capitalist system.

 

When I used the example that if it's possible to stream 4K video there certainly are no bars to streaming smaller file sizes of high res audio, you made the pointless statement that 4K video is compressed. There's no equivalency to be made here, we're just talking the requirements of streaming files of different sizes.

Now, fully knowing the only component relevant here is what consumers are paying for their internet, you attempt to explain inherent costs of internet infrastructure .

I have to say, this has all the signs of a person who wants to win an argument on the internet vs. actually make sense.

 

Netflix streams in 4k, Sky broadcasts in 4k for some sport content, the BBC, Prime and others have streamed HD for years. Bandwidth is not a problem here Mal...

 

What we have here is failure to communicate...

I'm not arguing that the network speed can't cope with lossless hi-res audio - what I am saying is that if the same consumer requirement can be fulfilled with lower bandwidth then that is of value to the service provider.

 

I'm sorry. The $200 a month I pay Comcast is going to have to be enough value for them. This is before we get to the part where we're creating yet another fake benefit of MQA. Please, let's step away from the crazy bandwidth ledge here. It's been a non-starter from the start and there is literally no argument to be made in this space.

 

Sorry, not seeing the benefit to audiophiles here. MQA is a solution still in search of a problem to solve.

 

@Mal I think you are confusing me with someone who care about the service provider bottom line, or are you implying that the real customers of MQA are the service providers?

 

But aren't you thrilled about the selfish benefits to the music industry and the ISP hegemony? I generally approach all decisions I make in my life around "but what would be the best for Comcast?", after all.

 

Sorry if this has already been covered. I use Tidal Hifi for streaming, through either a Schiit Modi 2 Uber DAC (non-MQA) or an Audioquest Dragonfly Red (MQA compatible). I haven't noticed much if any difference between streaming MQA "Masters" vs. regular Tidal Hifi files, which are 16/44 lossless. Maybe my system isn't resolving enough, but Tidal Hifi already sounds really good to me.

However, I do appreciate that many of the Tidal Masters use more recent masterings that were presumably developed for HDTracks and other similar hi-res stores. For example, if I want to stream something on the Blue Note record label, the standard 16/44 Tidal Hifi version is usually the poor sounding RVG mastering, many of which are also subject to UMG's poorly implemented watermarking. However, many of the same Blue Note albums are available as MQA Masters, and they use the more recent remastered versions (I think done by Alan Yoshida) that you can buy from HDTracks. These sound WAY better, and do not seem to include any audible watermarking. I don't really care if they're MQA or not - I'd be happy enough to have the same masterings available to stream in 16/44.

I guess this means I'm not necessarily pro-MQA or anti MQA, but I do want my streaming to use the best possible remasters available.

 

The average consumer will not be aware of this and believe that they are getting the full original sample rate delivered. This seems to be misleading at best and deliberately deceptive at worst, which seems to describe both Meridian and the MQA process. I see no value in MQA and will not support it in any way.