The web is on fire, for MQA supposedly uses upsampled music for a number of albums.
Time to investigate.
One way to check whether a file is upsampled or not is to shoot a spectrum using a technique
called fast fourier transform or FFT.
Any computer can do that, provided you have the right software.
For MQA you do need to be able to record the analogue signal back into the computer at
a higher sample rate, 96 kHz will be sufficient.
Since I have measurement equipment that is able to do that, I used that.
I used Roon to select and play the music, connected the Mytek Brooklyn MQA DAC over
USB to a Mac Mini that I used as endpoint and fed the output of the Mytek
to the Audio Precision 2522 measurement set.
Using Roon, I selected Owner of a Lonely Heart by Yes from the Tidal Master page and played it.
On the display of the DAC both the MQA indicator and 192 kHz was shown.
At 1 minute 30 I started the Audio Precision to take 128 samples
over 40 seconds and calculate the average.
This became visible on the screen: a sharp roll off at 20 kHz.
Would that be the proof that they used a normal 44.1 kHz recording and upsampled it?
Let's check another 192 kHz MQA file, a live Gary Clark Junior recording from 2016.
Here the energy slowly rolls off as you would expect for a 192 kHz recording.
Not so with the Yes recording.
So I compared it to the cd-rip by overlaying a second measurement, taken the same way.
Again - and here to be expected - there was a sharp roll-off that, for the first part,
was identical to the MQA 192 kilohertz version.
At about 23 kHz there is a kind of blurb that might be due to the reconstruction filter
that is different for 44,1 kHz, but I am not sure.
I also have a 96 kHz recording, ripped from the DVD-Audio version, if I remember well.
So let's measure that one too.
And again the same roll-off for the first part.
So the DVD-A was a rip-off too?
OK, my last resort, the SACD.
I don't have an SACD player anymore but I have all my SACD's ripped in Roon.
So I can also play the SACD version through the same setup.
And again there is a sharp roll off above 20 kHz, this time even deeper than the other
three measurements, as you can see in close up here.
Again I expect it to be due to different reconstruction filters, but that is not of great importance now.
The question is what happened.
The 90125 album started off as an album by two ex members of Yes,
since the band had split up in 1981.
They worked on the album at several places and with a growing number of band members
to finally issue the album under the Yes name again.
The album name is the catalogue number it got at Atco Records, which is an Atlantic
Records sub division.
Given the period the album was produced in, it must have been recorded on analogue multitrack.
The first digital Multitrack in the US and Europe was the Sony PCM-3324 that was introduced
in 1982 so the recordings were already on their way.
It is unlikely they redid the album and it is even more unlikely Atco would have payed
for the extra cost of hiring a PCM-3324.
What you might not know is that analogue tape recorders are also limited in bandwidth and
since 20 kHz was seen as the highest frequency people can hear, tape recorders were designed
to do no more than that.
Or sometimes a tat more for commercial reasons.
So we have a multitrack tape that is limited to about 20 kHz.
But it gets even worse…
The 90125 album is filled with the output of early keyboards
like the Synclavier and Fairlight.
I know little about these devices but I do know that, depending on the incarnation and
choices that had to be made, they could be limited to even 5 kHz bandwidth.
Don't forget that it was the beginning of the computer outside of big companies.
The first IBM PC was introduced in 1981 and was only able to display characters on a monochrome
- often green - screen.
The Fairlight costed as much as what I had payed - well, mortgaged - for my house.
Successful musicians bought these for it challenged their creativity.
I once read that all guitar solos on Owner of a Lonely Heart were sampled and played
back notes from one of these machines.
That might play a role in the limited frequency response too, although I don't expect it
causes the sharp drop at 20 K.
When all is recorded on the multitrack recorder, that tape is played back and mixed to a stereo
program on a 2-track recorder.
In this time period that could have been a digital stereo recorder, the first professional
machine, the Sony PCM-1600, was introduced in 1978.
This was a analogue to digital and digital to analogue converter that registered the
bits stored in a video signal on a modified U-Matic video cassette recorder.
One year earlier Sony introduced digital recording to the consumer the same way: the PCM-F1 adapter
that digitises the sound and used a Betamax recorder.
These were also used in studio's since they were far less expensive
than the professional variant
In both cases the master tape would have been a 16 bit 44.1 kHz digital registration
on U-Matic or Betamax.
In those days digital was seen as lossless so it could well be that an analogue copy
of this digital recording was stored as master or second master.
Since there hardly is any equipment left to playback those digital recordings, the analogue
tape might have been the 'original' master for the record company.
Analogue tape recorders are far more durable than U-matic machines, let alone Betamax VCR's.
And even if the multitrack was down mixed to an analogue machine, chances are there
was no information above 20 kHz due to the music passing through an analogue recorder twice.
There might be recordings from that period that go beyond 20 kHz since some use very
high tape speeds, like 76 cm per second - 30 inch per second imperial.
If that is done for both the multitrack and two track machines,
there can be spectrum above 20 kHz.
That would go at the expense of one octave on the low end - Patrons can watch the Patreon
Special I made called "Why analogue recordings have their limits too".
See the link in the notes.
Again it is not likely that this happened with 90125 since recording at double tape
speed would also double the tape cost from probably around $ 100 per 30 minute reel to
15 minutes on the same reel for the multitrack.
Don't forget that the multitrack is the sketchbook to a rock musician and that to
produce the 45 minutes album, they probably have used a great number of reels.
Also keep in mind that the album didn't even start of as a Yes album.
But again, these are all speculations.
Stating that the MQA version of 90125 was upsampled from a CD master is difficult to proof,
as is the opposite.
But there are indications that MQA does want the MQA label to be used for 'fair trade' recordings.
On Tidal many MQA titles are only 44.1 or 48 kHz, the sound improvement here should
come from the time smearing reduction MQA is claiming to offer.
So why should they issue Ed Sheeran DeLuxe at 44.1 kHz and 090125 at 192 kHz?
If 90125 has been recorded fully in analogue it does make sense to use MQA 192 since then
no or limited time smearing is added while digitising.
MQA even claims it can compensate for time smearing that occurred in the analogue tape recorder.
On High End Munich 2017, coming may, I will try to find out more on this subject.
So subscribe to this channel, my newsletter or to the Facebook or twitter page.
Even better subscribe to my Patreon channel for only one dollar a month and also watch
my Patreon Special videos while supporting me to do my work.
Thank you for watching and see you the next show or on theHBproject.com.
And whatever you do, enjoy the music.
For more infomation >> TÉLÉCHARGER DES CC - TUTO SIMS4 - Duration: 29:29. 
For more infomation >> Tutorial: Instalar retropie 4.2 en Raspberry Pi 1, 2, 3, zero alternativa a la Nintendo nes mini - Duration: 11:45. 
For more infomation >> Symbolik - Les trains (Sweetberry) - Duration: 9:33. 
Không có nhận xét nào:
Đăng nhận xét