I have that, nothing special there, apart from number of errors and mistakes.
Nothing here that would conflict with what I said. And regarding spline kernels, you get two of those in HQPlayer if you like such.
I could also refer to those and much more as basis for the work in HQPlayer… But the list of references would be in tens or hundreds of pages long.
This is all marketing material and has nothing to do with peer-reviewed scientific content.
Meanwhile, I continue to run my de-blurring for my already existing material in HQPlayer and that has output rate in MHz range (>= 128x) instead of max 8x of the MQA stuff…
I’m not going to moderate it because noting that Jussi may not be objective due to commercial interest, or that his response lacks detail, is fair comment. But I welcome Jussi’s contributions on the forum and in this thread. He doesn’t deserve the sarcasm.
When I talk about HQPlayer I try to make it clear that it’s my own talk and everybody knows where I’m coming from. I don’t try to use other people’s work as marketing material.
Using one’s own AES paper is fine for me. But I don’t like when academic research material from other people is presented such way that it seems like it is somehow directly related to the product or the product is direct result of such research.
When I post objective analysis of something, I try to make sure others can reproduce the same results independently. That’s what is largely missing from the MQA material, since they try to make it as hard as possible to objectively evaluate verify MQA’s claims.
You are free to do your own analysis and post the results. I really encourage as many people as possible to objectively evaluate real performance of MQA. And for example ask for MQA-encoded standard test tones and such.
Agreed about possible commercial interests - but this was brought up by others, not him, and he was accused without proof of his approach contradicting MQA’s, etc. So it is legitimate that he answers attacks, as a gentleman, which he has.
Thanks Jussi for your comments, agree with what you have said and your points on the technical issues (and I don’t have any ties to any company except owning Meridian Equipment and liking HQ Player and Roon that is).
Maybe you don’t need to post the full list just a few of the best references so that we other can educate ourselves would be really nice, I know I would like to read more on practical ADC and DAC issues/details (e.g. beyond the simple math).
Jussi since you have posted again the archimago blog tests I would be interested in your scientific analysis of the follow up test in the light of my comments below from last week.
Amen to that!
Actually the only thing I really care about is whether MQA Tidal streaming sounds better than HIFI Tidal streaming. If it sounds significantly better to me personally then I will push on with my plans to introduce MQA decoding capability into my main listening room set up. If not, then I won’t bother. Really simple.
I tried, thanks to 2L, several different file resolutions; CD, MQA, 192 & 352. I’m listening on a MacBook Air using a Meridian Explorer2 and Shure 535 IEMs. I cannot say I hear any difference between the higher resolution files and the MQA File. Obvious difference compared to the CD file. Could be more rig dependent too.
That would be good news for MQA - if it compares well to HiRes… I do think a lot of the major benefit is aimed at streaming as opposed to making hi-res sound ‘better’ so equal would be good.
My main take on the results is that there’s a significant reduction in SNR of the MQA version, even the decoded one, compared to hires version. Essentially the MQA version seems to be roughly equivalent of PCM sampled at 2x rate (88.2/96) and 16-bit. Distributed in such format as standard plain FLAC the file would be smaller and thus consume less bandwidth for streaming…
I have also obtained MQA capable DAC and will do some measurements with it once I find time…
I agree. I started investing in hi-Rez several years ago, but as an early adopter got let down. Not enough software. If we can get a higher adoption rate for MQA and greater availability, it is a winning scenario, particularly if the SQ is roughly equivalent.
I will first do some measurements through analog, and then later solder wires from the DAC chip’s I2S pins to a USB interface for reading the digital data.
I2S is interface for transferring PCM audio data between chips, designed by Philips. It is nowadays practically the only interface used by DAC chips for PCM. Earlier there were other similar interface protocols too that could be chosen from DAC side, but those are going away. It consists of minimum three digital signals; data, word clock and bit clock. And in most cases also fourth signal that is master clock (between 64x and 1024x of the sampling rate). Sample period is 32 bit clocks long and left/right channel samples are interleaved on the same data line. Many logic analyzers are also able to record and decode I2S.
DSD interface is simpler, it has two independent data lines, one for each channel plus a bit clock. In addition, DAC chips typically need master clock.
Then there is also I2C which is serial communications interface for inter-chip communications, also from Philips. It is typically used with DAC chips to configure lot of parameters in the DAC chip firmware. This is protocol that can be read and decoded by almost all logic analyzers since it is extremely common for lot if things varying from audio chips (DAC, ADC, volume control) to things like accelerometers/gyroscopes to thermal and moisture sensors.
.
Mr Stuart might send the codec police…