MQA is Vaporware

[Lee Scoggins]

On 2/16/2019 at 4:48 PM, FredericV said:

 

Some years ago I took a 24/96 file which I considered one of the best guitar demo tracks ever, and ran that through foobar with and without DSP.

The DSP was: downsample it to 16/44.1, and upsample it back to 24/96 - both with secret rabbit code.

I was doing the A/B on the fly by switching the DSP on and off in foobar, on a set which included a big Vitus amp and a big pair of Marten speakers on some hifi show. Nobody could hear the diff. I later did another experiment with the same file, where i used sox with highest settings, to create a "cd filtered" 24/96 version, and posted this on some internet fora. Both files were 24/96 files, but one had downsampled to 16/44.1 + upsampled back to 24/96 content from the original, in order to eliminate any sound differences by the DAC.

Nobody except one hifi dealer claimed to hear the difference. One listener cheated by opening the files with an audio editor and looking at the spectrum.

The content of that file was produced by Mark, so yes I believe him as I did a similar experiment.

So why go for hi-res? Not for the sound quality of hi-res itself, but for the fact that most likely more effort was put into the production as it is geared towards those who want quality. And thus recording & better mastering leads to better sound quality ...

 

The problem here is that the A/B is contaminated by the level of quality of the upsampling/downsampling algorithm used.  

 

A better test is to make a recording and split the mic feed into two recording boxes, one recorded at 16/44 and one at 24/96, level set to same.  Use acoustic instruments to make it easier to discern the quality.  

 

Then do just basic editing on both, then you have a fair test.  You want to have the exact same mastering so I just play it back on the same devices used to record over headphones.  The only thing that has changed is resolution.  I've done this over a dozen times.  The improvement from hires is noticeable.  One particularly strong test is classical violin.  24/96 clearly captures more of the sweetness of a pricey violin than 16/44.

 

The next test, we recorded one mic split to DSD and the other to 24/176 and 24/192.  The DSD sounded more like the live performance.

[Lee Scoggins]

On 2/16/2019 at 3:40 PM, Rt66indierock said:

 

I’m willing to listen to arguments that 12-14/30ish (analog tape) is high resolution but to me high resolution starts with microphones. I need special ones to make a high-resolution recording and they weren’t around when the White Album was recorded.

 

Better is fine but at a Los Angles Audio Show seminar we were told very few studios produce high-resolution recordings.  In any case what is the point? There isn’t a commercially viable market that can distinguish a well recorded CD from a high-resolution recording.

 

Or as Mark Waldrep said yesterday MQA is a hoax and so is hi-res. Quite a change from a man whose blog is Real HD Audio and believed until recently that his High-Resolution recordings were demonstrably better than CD quality. His own readers can’t reliably tell the difference.

 

There are, in fact, many microphones that capture north of 20khz.  But that's not what you need for the benefits of hirez recording.  Hirez is really about improvements in the audible range.  Hires adds more accurate timbre of instruments and better imaging.  All that leads to a more natural presentation that sounds much more like real musicians in a real space than CD.

 

What becomes problematic is when more recent albums are done in "midrez" Pro Tools which many industry people have stupidly adopted out of laziness/deafness.  Some supposed hirez files are just upsamples on these shitty masters.

 

But there is much hirez done from either higher sampling rates or analog tape to DSD /hirez PCM transfers.  I would argue the latter are the ones that most of us here would agree sound the best.

 

Of course, there are some stunning sounding CDs due to an excellent original recording and a great mastering.  But in those case, I often find that newer hirez versions are actually the best of all...provided the tape has remained in good shape.

[Lee Scoggins]

1 hour ago, Ran said:

 

No. There is nothing wrong with Pro Tools. It's how you use it.

 

I never said that Pro Tools was the problem.  Using ProTools to so 24/48 is the problem. 

[Lee Scoggins]

1 hour ago, Ran said:

 

No. There is nothing wrong with Pro Tools. It's how you use it.

 

Although to be more fair to engineers, a good number of them are being ordered to do this by the record labels they work for.

[Lee Scoggins]

2 hours ago, The Computer Audiophile said:

I could show you recordings that would blow your mind and are better than anything I've ever heard, that were also done on a ProTools rig. Blaming the tool doesn't make sense. It's all about the people involved.

 

You are not reading what I wrote above in responding to Ron and taking my words out of context.  It's not the app that is the problem, it is the implementation.  If you talk to someone like Joe Palmaccio, he will tell you that you need 24/96 to be effective.  Mark Waldrep would say the same thing repeatedly, at least up to the last year.

[Lee Scoggins]

2 hours ago, FredericV said:

1. So you are saying sox and secret rabbit code are contaminating the quality? We won best of show in Munich 2017 from 2 hifi press outlets, and sox upsampling was active > 1000 others brands and hundreds of competing rooms. Oh and we did this in 3.5 years we are on the market.

If nobody can hear the difference between 24/96 and a sox filtered version, why do we care?

 

 

2. Editing in 16/44.1 is stupid if you can record in 24 bit and have more headroom to work with. It's like shooting in JPEG with a pro camera if you can shoot RAW, and then use the 14 bit / color headroom for grading to output an 8 bit / color final result. Editing in 8 bit will limit what you can do in POST. Beginners shoot in JPEG.

So it again shows you are technologically illiterate.

Editing in 24 bit with dithering to 16 bit for the distribution file can still lead to 120dB of dynamic range. Just like a 4 color CMYK printer can output the illusion of millions of colors by distributing the error.

https://en.wikipedia.org/wiki/Dynamic_range
 

Now combine 1 and 2.

 

It's not your algorithm I am judging.  It's that the split-mic feed approach provides a better test and there is nothing interfering with the format differences.

[Lee Scoggins]

3 hours ago, Paul R said:

 

That would be 20/48k in a lossless format. 

 

 

Historically speaking, true high resolution has meant 24/88.2 or higher.

[Lee Scoggins]

1 minute ago, mansr said:

Your implication is that Pro Tools can't handle high-res formats, and that's patently false.

 

False.  I did not say that at all.  I know Pro Tools can handle hirez formats.  I have worked on pro recordings with it. The implication was that many in the studio are 1. either doing 24/44 or 24/48 out of following a standard or laziness, or 2. because the labels sometimes request that.

 

Many more seasoned engineers are arguing for all recordings to be done in at least 24/96. 

[Lee Scoggins]

3 hours ago, mansr said:

This is what you said:

 

If that's not implying that Pro Tools can't do high-res, I don't know what would, nor what you could possibly be intending to say. Perhaps you should just put that shovel down before the hole you're in gets any deeper.

 

Typo on my part. It should say "midrez on Pro Tools."  

[Lee Scoggins]

2 hours ago, Paul R said:

 

Well, in 1982, 16/44.1K was absolutely high resolution, and the promise was perfect sound forever. We all know how that worked out, but - there are still significant numbers of people today who will tell you that 16/44.1 is *it*, and it just doesn't get any better. Some people have the same opinion concerning vinyl. Or Reel to Reel tape.

 

As Chris pointed out, the skill of the person making the recording and mastering the final product seems to count a heck of a lot more than actual resolution, or technique, or equipment, or even the software. My personal opinion is that technically, the best sounding recordings of the past few years came from 24/192K recordings. Such as the title from Soundkeeper Recordings below. If it were only available in MQA, I would buy it without any hesitation at all. 

 

Today, I would possibly say DSD/DXD recordings technically have the best sound. So to me, that is probably the bar for "high resolution." 

 

It's kind of an individual thing though. The "best sounding recordings" to me are more often good recordings of music I really love, and I don't care all that much about the format. Well, except I will choose the format that sounds the best to me in regards to an individual recording, be it MP3 or Quad DSD. 

 

-Paul 

 

When we were doing the early dvd-audio recordings at Chesky in the mid-90s, 24/88 was considered hirez and it was really 24/96 back then and you only had Chesky's ironically-named "Super Audio Discs" and Classic Records "DAD" discs that would play on existing DVD players to output a 24/96 signal.

[Lee Scoggins]

37 minutes ago, Sonicularity said:

 

I don't see how 16/44.1 can be compared to vinyl or reel-to-reel when it comes to fidelity.  You are conflating inferior audio formats to rational people with a format that can be audibly transparent for playback of the recording.  You can't get better than transparent, but that does not mean that every Redbook product was made without flaws, only that there is no reason to believe it could not sound identical to any supposedly superior HiRes format.

 

 

 

A vinyl record can hold a 50khz signal which translates to 100khz sampling rate.  So vinyl is roughly around DVD-Audio quality which is what I usually hear as well.  Reel to reel is higher fidelity at 15 ips and still higher at 30 ips.  Analog tape is capable of truly outstanding fidelity.  

 

Redbook can sound great with a great recording and a great mastering.  But given the same recording on hirez and great mastering it will sound better still.

[Lee Scoggins]

Just now, Sonicularity said:

 

Why would any human care about 50kHz or 100kHz?  I have not seen any reliable evidence to suggest what you claim about HiRes being better.   I did say rational people, my bad.

 

There are several AES studies that found people could hear the improvements of hirez music.  The sonic advantages are pretty obvious to me.

[Lee Scoggins]

21 minutes ago, Sonicularity said:

 

Let's see these and we can create a new topic for discussion.  You still have not provided any reliable evidence.  

 

https://www.researchgate.net/publication/257068631_Sampling_Rate_Discrimination_441_kHz_vs_882_kHz

[Lee Scoggins]

This one speaks to how sound engineers prefer higher bit rates:

 

https://www.researchgate.net/publication/257068576_Subjective_Evaluation_of_MP3_Compression_for_Different_Musical_Genres

[Lee Scoggins]

Meta-study:

 

http://www.aes.org/press/?ID=362

[Lee Scoggins]

Study showing non-musicians can hear the difference.

 

https://secure.aes.org/forum/pubs/conferences/?elib=19650

[Lee Scoggins]

Perception of formats through headphones.

 

https://secure.aes.org/forum/pubs/conventions/?elib=18228

[Lee Scoggins]

50 minutes ago, Sonicularity said:

 

Why would any human care about 50kHz or 100kHz? 

 

Don't confuse audible range with sampling rate.  The latter has to do with resolution and that impacts the 20-20khz audible range.  Hirez also allows for better filters in digital.

[Lee Scoggins]

10 minutes ago, Paul R said:

 

Well I thought it was pretty obvious, but it did have to be explained for some reason. 

 

That MQA will never be the only digital format is just as obvious I would think.  Regardless of how much it is or is not accepted by the labels.

 

Tidal was MQA’s  great hope for dominance, and that is not working out all that well for them. The auto industry is their next hope, same as it was for sat radio.

 

We do live in interesting times.

 

In my opinion, the streaming services are the best hope for MQA due to the volume of customers and how royalties can be arranged.  If MQA catches hold with one of those then adoption would become wider and likely create a more viable business.

[Lee Scoggins]

8 minutes ago, mansr said:

You say that as if it were a good thing.

 

For those of us who like the sound improvement, it is a good thing!

[Lee Scoggins]

2 hours ago, Jud said:

 

Lee, you have a history (well OK, one example - LHL crowdfunding) of being oblivious to the dangers others point out in the financial machinations of the audio and music industry.  The danger here is the potential to rather easily cut off the supply of non-MQA RedBook and hi res for those of us who prefer it.

 

Jud,

 

This is misleading at best.  I wrote an article five years ago (March 2014) talking about the advantages of the direct to consumer model and its advantages.

 

http://thehighfidelityreport.com/death-of-a-salesman-lh-geek-out-campaigns/

 

Unfortunately I used LH Labs as an example of how consumer could save money.   They were financially healthy at the time and delivering product and well regarded for their flagship DAC.  Two years later things started to change but there was no way for anyone to foresee that.  The good news is that the  business value of a direct to consumer model has held up well.  Massdrop, PS Audio Sprout, and other examples are evidence of that.

 

As for cutting off non-MQA supply, there appears to be no danger of that.  We have Qobuz successfully launching and non-MQA source material is plentiful.

[Lee Scoggins]

1 hour ago, Jud said:

 

How is it you knew this?

 

Even Larry's very complimentary bio for a talk he gave at a business seminar in Taiwan said the crowdfunding campaigns had only brought the company near to break-even.  (I've linked it in the original LHL non-delivery thread if you'd like to read it.)  If you're a relatively small firm that has received $4.8 million in advance payments (the amount Crunchbase reports coming in from the campaigns - also linked in the same thread) for products you haven't yet spent the development, manufacturing and distribution funds for, and it only gets you near break-even, how financially healthy are you?

 

I was just going from the amount of money they raised at that point in time.  As a private firm, there was no way to get at financial statements.

[Lee Scoggins]

2 hours ago, Jud said:

 

In your article you talked about people foreseeing bad things happening because of LHL getting ahead of themselves with the crowdfunding campaigns, and dismissed it as sour grapes from competitors who'd be swept aside by, as you stated in your title, the "Death of [their old-fashioned] Sales Model."

 

No, it wasn't two years after 2014 that "bad things started to happen."  Most of the crowdfunding had ended by then (some spilled over into 2015), and two years later people didn't have any product.  That means a lot of bad things happened in between the money being collected and two years later, which is certainly enough time to deliver some pretty ordinary types of products - DACs, DAPs, and such.  Pono, no one's idea of a well organized business, managed to do it just fine with their eponymous DAP.

 

Sorry Lee, I don't think I'm being at all misleading by stating that people in the business were saying LHL was headed for a fall and you incorrectly dismissed their concerns.

 

I don't recall reading about major concerns in early 2014.  In fact, a group of us in Atlanta were receiving shipments of Geek Pulses (the local audio club was one of the biggest initial orders with 26 units).  I think it was our local President John was the first to tell me about it. I thought the business approach was interesting and I was in the market for a less expensive DAC so I put in an order thinking $200 or so was a reasonable sum to risk.  Of course Gavin went nuts and every month or two sent out an email suggesting yet another upgrade.  I had around $800 into the game and received the Pulse Infinity with "naked resistors".  If there were credible concerns then that I was aware of, then I would have chosen another company to discuss.  

 

Also, there were sour grapes from retailers I knew at the time and from Stereophile who was probably protecting their advertising base to some extent.  At the time, two groups of business were threatened by this new model: the distributor and the retailer.  Fortunately, now we have manufacturers happily supporting kickstarter campaigns and sites like Massdrop.

 

So I made a decision to write about LH in the article because:

 

1.  It was interesting from a strategy consulting viewpoint, ie. new business model with advantages.

2.  It would be an opinion piece that my publisher Chris thought made sense.  He encouraged me to write about it.

3.  The company appeared to be real in that I had met Larry and Gavin at RMAF and Atlanta Axpona and they seemed to be legit.

4.  They had already started shipping units to our local audio club.  Everybody received their units on the initial order in fact so no red flags there.

 

It was reasonable decision based on what we knew at the time.  It's easy to second guess this decsion five years later.

[Lee Scoggins]

2 hours ago, Jud said:

for products you haven't yet spent the development, manufacturing and distribution funds for, and it only gets you near break-even, how financially healthy are you?

 

This isn't accurate.  Larry and Gavin were showing the Geek Pulse boards at RMAF already.  The Pulse had been fairly far ahead in development which is another factor in my own personal decision to send in money.  Of course the next campaign for the Wave was a disaster.

[Lee Scoggins]

6 minutes ago, crenca said:

 

This is off topic - there is another thread about this, please post all this detail there.

 

Talk to @Jud, he brought it up.