Originally Posted by erich6
I'm not sure what you are using as your input but assuming you are using an analog source and that you have measured the output through an analog scope and posted a screenshot then you are seeing just the effect of the EQ. The MQA filtering is supposed to incorporate the effect of the original analog to digital conversion, which I think includes a specific digital sampling approach (a nonlinear operation), as well as the digital to analog conversion at the user end. You can approximately represent this with a convolution kernel that looks similar to what you've measured here but is not exactly the same. We can argue to what extent the differences are significant (i.e., is it audibly perceptible) but mathematically they are not the same thing. And, the theory is that while the ringing frequency is inaudible (if it were to be represented by a singular tone), the impact of that ringing (it's convolution with the original continuous signal) is perceptible because it affects the timing of notes relative to each other. Now, I'll concede that the painstaking approach MQA takes to reproduce the analog signal ultimately can be invalidated by the analog speaker/headphone system being used as room acoustics, driver angle, etc. will probably swamp these effects. I am also aware of the amped-up hype and marketing...but that's normal for ANY product.
On a separate (more subjective) note, many including myself have noted the improved sound quality of Tidal Masters.... I believe this is more of a function of the actual remastering quality than the MQA encoding but if MQA is helping/encouraging studios to distribute better masters than I'm all for it!
I could have explained what I was doing a little better, sorry about that. I'm generating a sweep, sending it to my EQ(used in both instances just for consistency in the signal path to show that only the applied EQ is the reason for change in the impulse). and the output of the EQ is measured in REW(because I know I can get an impulse response easily that way and I'm lazy^_^).
and of course I'm not saying I'm doing the exact same thing as MQA does with same settings and all, but the principle is massively the same. most of the timing blablah can be simplified as sample rate and low pass filter. they explain it with marketing lingo trying to make it look like a big deal, but that's all there is to timing accuracy in digital audio. more samples increase the time accuracy, different low pass filters will create some phase shift and do a few other things.
in fact more bits also increase the timing accuracy(because sine waves), so discarding some is bad for time (ok I'm half trolling again but it's still true).
for example, take a typical 24/96 master. because unlike Meridian, the studio cares about doing nominal band limiting while keeping as much content as possible and avoiding aliasing, they tend to apply a very strong low pass close to 48khz that will ring at those freqs, but everything else is kept with great fidelity in time and amplitude. that's what the master side of MQA is supposed to "improve".
so how can they "improve" this 24/96 original master in the time domain? well they can't really pull higher sample rate data out of a hat to increase time domain accuracy. 96khz is all the data they have for this album and the hard low pass is in the signal now. so that's dead.
we're left with the idea of a new low pass filter before the low pass filter. to make a cleaner cut on the cheese, you cut it again. all of MQA aside from the compression design is based on that. they use a gentler slope for the filter, kind of like my poor example, and some extra salsa for the apodizing filter of the MQA DACs in the playback. all so that at the end of the audio band, there will be very little ringing. basically because there will be very little signal left at the ringing frequency and before. that's why they're so focused on time domain everything, because in the amplitude domain, they take no prisoner.
about making new nice sounding masters(which has nothing to do with the MQA format itself), I've already expressed my opinion. I'm all for more nice sounding music, but good masters available only in a particular format is a solution to milk the cow. choice means having all the masters available in all formats and resolutions equal or below the original. be it from the seller, or as a free convertion tool for the consumer who purchased the high resolution album. any other offer is restrictive and plays against us consumers.
also if I can only get a master in one format, I'd rather have it in one that is actually high fidelity, MQA is destructive compared to PCM.
now all this is my rant, I find the entire MQA solution to be bad for us and I would never purchase a MQA album unless everything else has disappeared(you can call me on that in 10 years if you want, it's an easy bet for me to win). but about MQA as a streaming solution, why not, those who want it can have it. those who don't, don't need to. the world is big and all is well. I just find a little strange how people would use tidal to avoid lossy formats(because let's be honest, people aren't all in love with tidal's interface....), but some of them are now enthusiastic even about listening to some 13/48 undecoded MQA on their phone. I find that strange.