Is DEQX a game changer?

I see that my recent post may have confused a few people so I will attempt to clarify.

Most of my listening is with vinyl rather than digital and I prefer a very natural analogue sound whereas (to me) a lot of digitally reproduced music has a slight glare to treble which sounds unnatural. Different DACs can either exaggerate or lessen this. As a result I’ve auditioned and used a variety of these over the years until I found something that, in my opinion was as close to the pure ’analogue’ sound I was after. That is what inputs to the HDP-5 via balanced analogue connection.

Therefore I prefer to use the balanced XLR input where digital sources (CD, FLAC etc) are already processed so in effect the processor sees everything as analogue. Processing of digital inputs to the DAC in the HDP-5 is certainly a step up from the HDP-3 - both are very good. It’s just that I am rather picky & have a setup (using the external DAC) where it is pretty difficult to tell the difference between a vinyl or digital copy of the same album. Using any digital inputs into the DEQX DAC, treble is a little more ’brittle’ (I just did a swap to make sure I could qualify this remark - the DEQX processors are so good that we are talking small degrees here & I continue to be astounded by the life-like realism that the system produces).

note: from the DEQX product overview of the HDP-5... "provides transparent analogue pass-through.......Analogue inputs utilize Cirrus Logic’s reference ‘professional’ ADC to provide absolute transparency for analogue sources such as vinyl preamps and surround receivers"....I agree with that.

For sure, in theory removing the sequence that I use (external DAC =>A/D =>D/A =>out) should benefit from reduced processing. From living with DEQX & comparative listening for over 4 years, that isn’t the case for me.

drewan77, 

Sorry if I missed it, but have you tried DEQX digital out --> external DAC --> preamp (if volume control is needed)? If so, how did that compare to the setup you are currently using ( external DAC =>A/D =>D/A =>out)? 

WARNING, I’m about to bring up the controversial subject of MQA. I have been following MQA with great interest these last couple years. I haven’t had the opportunity to hear it myself, but have read as much as possible about the technical aspects as well as subjective sound quality from likes of Stereophile, AudioStream, Digital Audio Review, The Absolute Sound, and others. I’ve been crossing my fingers for the major labels to adopt the process and for content to start streaming in (pun intended). I feel MQA could be the next big thing in hifi.

I also feel that DSP ala DEQX has major potential for improving hifi. But here’s the rub, we might not be able to have both. According to Benchmark, “MQA requires a lossless transmission system from the file source to the final D/A converter. Benign DSP processes such as a digital volume control (used in moderation) immediately defeat the MQA decoding. The same is true for digital crossovers, digital EQ, and room correction. The MQA stream will be corrupted if any of these common processes are encountered. These common forms of digital processing will shut down the MQA decoder and revert the system to a 44.1 or 48 kHz sample rate and an effective bit depth of 13 to 15 bits.” https://benchmarkmedia.com/blogs/application_notes/163302855-is-mqa-doa

I haven’t heard any hifi journalists/reviewers mention this. I’d feel more confident in the truth of this claim if it were echoed by another trusted and knowledgeable source. But it seems to make sense. If true…that sucks. Both MQA and DEQX are trying to get to better sound through, among other things, timing accuracy. It’s too bad both technologies can’t play nice together.

@blang11

We have a couple of other threads re:MQA going on, but yes, it’s true, but ONLY if you are EQ’ing in the digital domain. Using something like a miniDSP or DiracLive which does AD -> EQ -> DA is not affected! Except that if you do that, you are going to do away with the fairy dust MQA sprinkled. More below.

The reason this happens is that the extra octaves are being hidden in the lowest bits of the signal. So, change any of those bits and you ruin the ability to decode the signal. Just like if you took a ZIP file and started messing with the lowest 2 bits of every byte.

Even if you include your DSP in more advanced DAC’s (of which I know of like 2) It’s not at all clear that the alleged benefits of MQA’s digital decoding filters would live through it.  Like doing this in your player:  MQA-> PCM -> EQ -> MQA ouptut filters -> D/A converter

Then you take the reason the filters allegedly work. They claim less time smearing than conventional digital filters. Well, if you stick a miniDSP (or equivalent) your going to be putting your sound through another A/D, D/A signal without any of the fairy dust present.

To my mind this whole house of mirrors starts to unravel when you start asking what happens with multi-track sources, and when the MQA encoding filters gets applied.

By the way, I went to 2L.com and listened to a few MQA encoded tracks. I could not tell a difference between the 96/24 and the MQA which claimed to be doing 384k/24.

From a practical perspective, this whole thing is looking more and more like Enron or "Dark Energy Generators." I’ve yet to either hear a difference, and I have asked, repeatedly, to be given access to a pair of demo tracks which would demonstrate this alleged superiority.

In addition, what I did not know, is that the whole encode/decode system is lossy. Based on the origami charts I had assumed it was lossless. Not so,

Best,


Erik

@blang11

Another thing that is not clear is how MQA encoding plays with regular studio work.

The way it should to work is that the MQA encoding process should take your A/D converers into account and compensate for it’s shortcomings. Well, what does that mean if you are mixing several tracks, or doing level changes, EQ, compression, or expansion? Does MQA somehow magically know how to compensate for these algorithms? Mind you, this is all work done before we even encode the signal. 

It is possible some of this could be overcome, like with DSD. What’s not clear to me is that there’s any benefit at all to MQA encoding. I remain unimpressed.

Best,


Erik