MQA pt II
- Thread starter Jim Audiomisc
- Start date
DimitryZ
pfm Member
Because your post is filled with detractions, zero evidence and no subjective experience either, at least none you are willing to share. It's essentially an angry, anti-MQA diatribe. That you don't understand it as such is a separate puzzle, which I will untangle for you.
If I may read in between the lines, you had MQA, listened and probably liked it. Then you started to "pay attention," which likely means you started to notice "seemingly knowledgeable" people on the internet thought MQA is bad.
As you lacked technical knowledge to understand the arguments, (and in most cases they were flawed and biased) you decided that they must be "right." That after a while, started to enrage you that you were *swindled" with "lossy music."
Then you started chasing me around, convinced that if you insult and pester me enough, it will make you feel better. Because to you, I am an self-evident eyesore - a happy, well-adjusted person, mostly enjoying MQA and all other digital and analogue formats. Much unlike you - and that makes you angry.
I think I get you now. Your half-hour of analysis is done.
Last edited:
DimitryZ
pfm Member
I have outmost respect for Mr. Paravicini and though I haven't owned his excellent tube amplifiers, I will gladly stipulate that they are exceptional in every way technically possible. I have owned the giant Canary Audio 300B monoblocks (4 power tubes per side). Before that I owned the big Antique Sound Lab Hurricanes (8 KT88s per side). My final tube preamp was the rather special dual box, internally dual channels and PSs with tube rectification Canary 801SE. They sounded great. For my vinyl replay, there were no solid state devices in signal chain or power supplies. And I am only saying this to show I am a decades long tube audiophile. By no means a tube detractor.
However, if tube power amplifiers are directly compared, measurement by measurement to a high powered modern SOTA solid state amplifiers (like my Emotiva XPA-1 Gen 2s, for example), they will loose in every category and by some margin. And they won't be able to drive difficult speaker loads (like electromagnetic panels, for example), because they have an output transformer and can't provide the required current and damping factor. Into a right speaker, they will undoubtedly sound awesome, but will still have a bit higher distortion (usually second harmonic that makes the sound just a bit richer).
These are inconvenient facts about our hobby. We like equipment that by all measurements we have been able to think up during over a century of research and development is just not as good.
nice of you
in my post I shared my experience and thoughts about MQA, nothing really about you personally? and I asked you a simple question, why could you not bring yourself to answer it?? especially when you continue to insinuate? instead, your post is virtually all about me, rather disrespectful, rude, assuming and personal, what makes you think this is ok??
Did you forget to call for the moderators not doing their job?
)
tuga
Legal Alien
Which of the following can you refute:
• lossy compression
• apparently not authenticated
• leaky filter
• proprietary technology
• DRM potential
No arm-waving deflections please.
Jim Audiomisc
pfm Member
At present I can confirm that I have what are labelled as being in four categories.
Original Masters
Tidal Masters
MQA Tidal
Unfolded output.
In each case I have 44k and 88k examples. The unfolded include 'unfolded' and 'fully unfolded' as labelled. I've been examing their contents *thus far* as a series of binary values. Looking for patterns that may give information about the contents, etc.
I've not got anywhere near finished, nor have I with the 2L files. But there are various signs that the 2L examples and the GO ones do have some differences. That is currently a tenuous view as I've got to check various things and have to keep adapting the analysis methods as I learn more.
I assume the files you can't open are examples like
http://jcgl.orpheusweb.co.uk/temp/WaitABit.png
This is a PNG. There are countless millions of these on webpages. It opens here when I click the link in any of my browsers (Linux as well as RO.) It is a net standard filetype. It has a .png extension. I've not seen anyone else saying they can't view the files I've put up. Has anyone else had a problem with them?
Two of the later files I put up are a CSV and a zip. They contain some of the data.
As above, if anyone who wants that data can't open them, please say.
DimitryZ
pfm Member
Report me.nice of you
in my post I shared my experience and thoughts about MQA, nothing really about you personally? and I asked you a simple question, why could you not bring yourself to answer it?? especially when you continue to insinuate? instead, your post is virtually all about me, rather disrespectful, rude, assuming and personal, what makes you think this is ok??
Did you forget to call for the moderators not doing their job?
DimitryZ
pfm Member
LOSSY compression:
My sense is what is coloquially referred as such, is the rather carefully chosen "leaving behind" totally musically unrelated portions of traditionally ultrasonic noise loaded hires files.
Unauthenticated.
Music business complaint. Whomever holds the copyright has to sign the release. Maybe not the right person, someone who wasn't present, especially if the musicians are dead.
Leaky filter.
Just a deragatory term used by anti-MQA folks. Many people prefer these sonically and set their DAC filter settings to them.
Proprietary tech.
Zero problems with that. You and I use such every day, but you only complain about MQA.
DRM potential.
Whatever it had at inception has been publically disavowed by MQA "Only crypto in MQA is for authentication." DRM has long been abandoned by the music industry, but audiophiles continue to scare small children with it.
Werner
pfm Member
tuga
Legal Alien
Compression is used to solve bandwidth limitations.
FLAC is more effective at compressing: a 88.2kHz/19-bit FLAC is smaller than the equivalent MQA (only 60kHz/17-bit).
And if are selling the definitive master recording from the studio then the compression cannot be lossy.
The following is MQA propaganda:
"Devices or apps with MQA decoding capability can fully ‘unfold’ the MQA file and reveal the original master resolution.
They will also authenticate the file to guarantee that it is the definitive master recording from the studio.
MQA’s philosophy is based on the three pillars of quality, authentication and convenience.
This ensures that the artist-approved version of a song is delivered conveniently and efficiently, and in the highest quality, from the studio to the music fan."
I disagree. For me it's an end user compaint.
MQA is promissing the definitive master recording from the studio but that definitive master has been crippled and apparently it's not even guaranteed by the studio.
How many?
The filter is innefective. The adjective "leaky" is quite illustrative of the issue in question.
I would say that in this thread only you don't complain about MQA. But you don't complain so much and so loudly that it looks like you are many and we are only a few...
You have to trust MQA.
DimitryZ
pfm Member
To reiterate, as far as we know:
1. "Original Master": Original files GO uploaded to the MQA encoder.
2. "Tidal Master": Files published by Tidal that presumably have been MQA encoded.
3. "MQA Tidal Unfolded Output": 2. above that have been run through some form of MQA decoding.
What is puzzling is why Tidal "raw" content would be available in 44 and 88 rates? MQA is a 44KHz container and doesn't exist in the wild in 88KHz form.
Further, what is meant by "fully unfolded"? MQA doesn't use this terminology. Tidal will unfold MQA container to frequency of the original, say, 88KHz (we will leave the strange cases of 44KHz MQAed original files out for now). This is send to the DAC, along with a filter setting and upsampling commands (as mansr has shown). This is LPCM and can be digitally captured.
What can't be digitally captured is the final rendered and uosampled LPCM that is sent to the DAC's D/A stage. It can perhaps can be captured from the analog outputs into an ADC, but that's a rather crude process.
Perhaps you can reach out to GO, researcher to "researcher," and get him to reveal his actual methodology, which despite all the flash and graphs shown on YouTube, he has not done.
I can open the file in your last post, but not the ones in your zipped file. The zipped file opens fine in Windows, but the files have an extension of "file" and won't open. When changed to a "png" extension they still don't open.
Can you post these files with a native "png" extension without being zipped?
DimitryZ
pfm Member
Totally unnecessary "TO ME"
Jim Audiomisc
pfm Member
What term woulf you prefer? The filers - by design - generate anharmonic aliasing. Which in other contexts would be regarded as distortion. I agree 'leaky' isn't really an ideal term to indicate this, but I'm not sure what would be better.
Jim Audiomisc
pfm Member
In addition the the three you list they include MQA encoded. i.e. nominally two types as per Tidal described. One MQA the other 'Tidal Master' according to GO. I don't use Tidal so can't comment on that.
AIUI GO sent in *two* versions. He generated an 88k source file with content covering its full range. He also generated a 44k nominal equivalent which is the same up to its Nyquist limit but omits the higher components of the 88k source. So he sent in a 44k and an 88k version. Then got 'Tidal maste' and 'Tidal MQA' versions back. Then 'decoded' these using the two processes that MQA offer. One expands the 44k to 88k. The other expands further on that. This is what I think he means from what he says.
I've not yet done much with his 'unfolded' files, as I'm intially working on the others and the 2L examples.
Actually capture from the DAC output can be done quite well if you have a decent ADC and calibrate its behaviour, etc. FWIW I use the Benchmark one which seems pretty good. But I have various others.
The files in the zip are all essentially text or csv. Can you not filetype them on a Windos box or drag them to text editor? Afraid I stopped bothering with Windows a couple of decades ago as I found it too clumsy to use. I can re-do the zip though. If so, do Doze users also need a C program to have a .c extension?
FWIW I would like to converse with GO. But I don't want to have to join the ASR forum as well as this one as it would take up more of my time and TBH it isn't as easy for me to read/use as PFM. Tony's layout is excellent and I can use it easily. ASR has more cruft and a lower percentage of screen area devoted to the 'body text', and is, for me, harder to read.
Jim Audiomisc
pfm Member
Following on from the above it may be worth my mentioning that I've recently been testing and using a 3rd Gen Scarlett 2i2 ADC+DAC and given its price, etc, this is quite impressive. It certainly isn't as good as a Benchmark or similar, but is far cheaper. So used with care, can make decent measurements. Not as flat-response or low-noise at HF, but not bad given the budget price.
Jim Audiomisc
pfm Member
OK, in case it helps I've just modified the contents of the zip at
http://jcgl.orpheusweb.co.uk/temp/IQResults.zip
I've added filetype extensions to the files and also added a short text file saying what the other files contain.
The data is essentially all CSV. The program is a 'C' one.
Sorry for any confusion. I'm used to being able to DnD any file into a general file editor and it will show it as 'text' regardess. The graphs I put up were all pngs so should be fine anyway.
http://jcgl.orpheusweb.co.uk/temp/IQResults.zip
I've added filetype extensions to the files and also added a short text file saying what the other files contain.
The data is essentially all CSV. The program is a 'C' one.
Sorry for any confusion. I'm used to being able to DnD any file into a general file editor and it will show it as 'text' regardess. The graphs I put up were all pngs so should be fine anyway.
how playful! don’t flatter yourself
DimitryZ
pfm Member
FLAC is way less effective as the recording frequency increases. It is grossly less effective at the 352 DXD frequency, which is loaded with random ultrasonic noise - the current darling of the "stupid lossless" crowd.
Your issues with MQA advertising should be addressed to MQA PR department, not me.
All digital reconstruction filters are a set of compromises. The filter most "effective" (I assume you mean "in the frequency domain" - and it's good to be specific if you are going to forest into technical side) is the original brick wall filter, but it's has poor time domain performance.
There are a few others that are agnostic or even positive about MQA, but they are understandably reluctant to be exposed to the vitriol thrown at anyone who dares "not to hate MQA, together with the rest of us."
You have to trust you ears.
DimitryZ
pfm Member
All filters have traditional signal processing discipline designations, that should be used instead of coloquial terms. My understanding is that the filter long favored by Meridian and one most likely used in MQA encoding (likely with coefficient variations specific to the recording chain used) is a form of the apodizing filter.
DimitryZ
pfm Member
MQA offers one decoding process, which it calls unfolding. This has now long been available in the Tidal software clients on various platforms. The second process is rendering (specific reconstruction filter instructions and original file bitrate upsampling instructions). Mansr has made this quite plain. Feeding the encoder 44KHz files seem irrelevant to me, but since MQA themselves seem to do this, it's not out of bounds. Regardless of what bitrate (44 or 88) was fed into the encoder, the resulting MQA container is 44KHz, by definition, it seems, as it's required to be CD compatible. If the 44KHz MQA container is upsampled by GO to 88KHz that would seem to be a wrong thing to do entirely.
Perhaps you can reach out to GO directly. If I was doing this work I would want to know exactly how these files were handled and if they went through an ADC, what equipment he used, etc. But that's me.
Perhaps someone who is a member at ASR (paging mansr) can invite GO to join PFM.