MQA alternative - open source

I, and some others, have concerns about the 'closed' nature of MQA and the
way it might give big companies a level of control over what makers and
users of hardware and music may do. So we have been considering the
possibility of an open and free alternative.

This has been discussed on the 'pink fish' forum and I've released a 'bit
freezer' program as a part of this for experiment purposes. If anyone is
interested I'd suggest looking at the relevant forum threads. But the basic
idea is to have a system that people may prefer and will cost less in terms
of money *and* in terms of limitation of your choice.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

How is bit freezing different from simply converting 24-bit FLAC/ALAC
files to 16 bits (with appropriate dithering)?

Julf

In article , Johan Helsingius
wrote:
How is bit freezing different from simply converting 24-bit FLAC/ALAC
files to 16 bits (with appropriate dithering)?

Just to check: Have you already looked at the webpages I wrote on this a
while ago? If not, they do give some of the background, etc, which may
help.

The bitfreezing is discussed and exampled on
http://www.audiomisc.co.uk/MQA/cool/bitfreezing.html

but these two pages preceeded it to look at MQA and get this into context

http://www.audiomisc.co.uk/MQA/origa...reAndBack.html
http://www.audiomisc.co.uk/MQA/bits/Stacking.html

So reading those first might be useful.

More specifically, bitfreezing lets you choose how many bits per sample to
'freeze'. Simply converting down to 16 bits nominally means losing 8 bits
per sample if you start from 24 bit. That may be too much or too little for
a given recording.

But with bitfreezing you can remove, say, 6 bits per sample. Hence you can
tailor this optimally for the amount of excess noise bits in the given
input material. i.e. remove only over-specified noise.

Indeed, in some cases you may find that high sample rate material has a
wideband noise floor distinctly *above* the 16 bit level. So might choose
to freeze some of the LS bits in a 16bit recording.

Beyond that, it comes down to being a case-by-case decision.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

On 05/09/2016 17:12, Jim Lesurf wrote:
The bitfreezing is discussed and exampled on
http://www.audiomisc.co.uk/MQA/cool/bitfreezing.html

I hadn't. Interesting!

FYI you have a typo:

However there are some special cases where can’t make use of FLAC
for what the MQA patents call ‘legacy’ reasons. So let’s now look
briefly at those...

You need "we" after "where".

I remain happy with my LPs digitised at 88k2/24 and downsampled to
44k1/16 after click reduction. As good as I think I'll get.

Andy

Just to check: Have you already looked at the webpages I wrote on this a
while ago?

Yes, I had looked at them, thanks!

More specifically, bitfreezing lets you choose how many bits per sample to
'freeze'. Simply converting down to 16 bits nominally means losing 8 bits
per sample if you start from 24 bit. That may be too much or too little for
a given recording.

OK, so 24-16 (possibly dithered) truncation is a special case of
bitfreezing - bitfreezing being a general N-M (possibly dithered)
truncation?

If so, then yes, it probably gives you all the benefits of MQA, without
any patent/licensing issues.

Julf

In article , Johan Helsingius
wrote:
Just to check: Have you already looked at the webpages I wrote on this
a while ago?

Yes, I had looked at them, thanks!

More specifically, bitfreezing lets you choose how many bits per
sample to 'freeze'. Simply converting down to 16 bits nominally means
losing 8 bits per sample if you start from 24 bit. That may be too
much or too little for a given recording.

OK, so 24-16 (possibly dithered) truncation is a special case of
bitfreezing - bitfreezing being a general N-M (possibly dithered)
truncation?

If so, then yes, it probably gives you all the benefits of MQA, without
any patent/licensing issues.

Yes. It also avoids the problems inherent in MQA 'Origami' which tends to
spray anharmonic distortions into the results. No need to even downsample.

So far as I can tell, it works as well, or better, and is free and open.
Anyone who wishs can use it - as is, or modified to suit.

FWIW I also have my doubts about the MQA 'doctrine' which takes it as
'axiomatic' that human hearing "beats Fourier" and that very high "time
resolution" is required. So I'm looking into this at present and may add
another page to deal with it.

My personal suspicion is that the main reason companies may adopt MQA is
because they think it might given them a new IPR control mechanism and
platform. Adnd let them re-sell the same old content to us all, yet again.
Money for old rope. For them, 'sound quality' may be the bait they can use
to hook users.

But my basic view is that no-one actually needs MQA if they simply want
lower stream rates and smaller files for 'high resolution' audio. There are
alternatives. Bitfreezing is one. Based simply on realising that a lot of
the 'content' of high rate files may simply be over-specified noise bits!
These pad the files/streams to no useful audible purpose.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

In article ,
Vir Campestris wrote:
On 05/09/2016 17:12, Jim Lesurf wrote:
The bitfreezing is discussed and exampled on
http://www.audiomisc.co.uk/MQA/cool/bitfreezing.html

I hadn't. Interesting!

FYI you have a typo:

However there are some special cases where cant make use of FLAC
for what the MQA patents call 'legacy reasons. So lets now look
briefly at those...

You need "we" after "where".

Noted. :-)

I remain happy with my LPs digitised at 88k2/24 and downsampled to
44k1/16 after click reduction. As good as I think I'll get.

I've tended to use 96k/24 and then leave it as such after declicking. Saves
the bother of then resampling, and pushes any DAC reconstruction problems
up well above 20kHz. As things stand I have plently of disc space. But I
realise this is wasteful.

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

FWIW I also have my doubts about the MQA 'doctrine' which takes it as
'axiomatic' that human hearing "beats Fourier" and that very high "time
resolution" is required. So I'm looking into this at present and may add
another page to deal with it.

Great! Seems most of the audio press has swallowed the MQA doctrine
hook, line and sinker...

My personal suspicion is that the main reason companies may adopt MQA is
because they think it might given them a new IPR control mechanism and
platform. Adnd let them re-sell the same old content to us all, yet again.
Money for old rope. For them, 'sound quality' may be the bait they can use
to hook users.

I agree.

But my basic view is that no-one actually needs MQA if they simply want
lower stream rates and smaller files for 'high resolution' audio. There are
alternatives. Bitfreezing is one. Based simply on realising that a lot of
the 'content' of high rate files may simply be over-specified noise bits!
These pad the files/streams to no useful audible purpose.

+1. Too bad nobody has the guts to actually state that in the
mainstream audio press.

In article , Johan Helsingius
wrote:
But my basic view is that no-one actually needs MQA if they simply
want lower stream rates and smaller files for 'high resolution' audio.
There are alternatives. Bitfreezing is one. Based simply on realising
that a lot of the 'content' of high rate files may simply be
over-specified noise bits! These pad the files/streams to no useful
audible purpose.

+1. Too bad nobody has the guts to actually state that in the mainstream
audio press.

Sadly, they tend to focus on "Well [I think] it sounds good" without even
asking if it sounds the same as the source material.

FWIW I tend to approach this by writing about the technical aspects which
may tend to 'undercut' claims made by companies, etc.

e.g. My column in the current (October 2016) issue of 'Hi Fi News' explains
bitfreezing and how it can cut away the 'fat' of over-specified noise to
get smaller files/streams. But doesn't mention MQA at all. People can then
draw their own conclusions... :-)

One good outcome of this has been that it made me, and some others, realise
just how much such 'fat' there is in many high rez files/streams! In some
ways this reminds me of the retail practice of putting things into a much
bigger box than actually needed for the content. Makes the item more
impressive at point-of-sale and tends to squeeze competitors off the shelf.
"Wow! A big file. I'm getting a lot for my money!" :-)

However sometimes you may get the same meat as in the past, but put into a
bigger bun. ;-

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

Sadly, they tend to focus on "Well [I think] it sounds good" without even
asking if it sounds the same as the source material.

Ah, yes, the "audio is art, not science" school of thought.

e.g. My column in the current (October 2016) issue of 'Hi Fi News' explains
bitfreezing and how it can cut away the 'fat' of over-specified noise to
get smaller files/streams. But doesn't mention MQA at all. People can then
draw their own conclusions... :-)

That column is what made me look up your longer write-up. Hi Fi News
is the only Hi-Fi magazine I still subscribe to, mainly because of
stuff like the spectrograms of "hi-res" recordings, but with almost
every issue lately I been having the feeling that they are on a
rapidly slipping slope to pure subjectivism - and considered cancelling
my subscription.

One good outcome of this has been that it made me, and some others, realise
just how much such 'fat' there is in many high rez files/streams! In some
ways this reminds me of the retail practice of putting things into a much
bigger box than actually needed for the content. Makes the item more
impressive at point-of-sale and tends to squeeze competitors off the shelf.
"Wow! A big file. I'm getting a lot for my money!" :-)

However sometimes you may get the same meat as in the past, but put into a
bigger bun. ;-

Indeed. But bigger numbers must be better, right?

Julf