[Jack-Devel] Network Audio Transmission - Quality

classic Classic list List threaded Threaded
22 messages Options
12
Reply | Threaded
Open this post in threaded view
|

[Jack-Devel] Network Audio Transmission - Quality

Thomas Brand

Hi,

I'd like to find out the impact on signal quality when using different
network tools for jack (netjack, jacktrip).

Question that come to mind are,

How can one test how good a method for audio transmission performs in
terms of equality of source and destination signal?

Is there audio material especially well suited for testing that?

How to compare source and destination signals/files?

Any hints are appreciated.

Best regards,
Thomas

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Harry van Haaren
On Sun, Jun 23, 2013 at 2:38 PM, <[hidden email]> wrote:
> I'd like to find out the impact on signal quality when using different
> network tools for jack (netjack, jacktrip).

I don't think there is *any* impact on sound quality: the audio data on both sides is binary identical: hence no loss by transmission.

So perceptually they're identical (netjack uses float as sample format, at the same samplerate: so just like an normal JACK patchbay connection, nothing is lost ). For details on encoding etc on netjack checkout this page http://trac.jackaudio.org/wiki/WalkThrough/User/NetJack
Using certain encodings / bit-depths will cause reduction in sound quality: as defaults I think its binary identical... perhaps don't quote me though, I'm not an author of any of the implementations...

> Question that come to mind are,
The questions that will show the difference between netjack, netjack2 and JackTrip are how jitter / packet loss / latency are handled / compensated for.

HTH, -Harry

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Ralf Mardorf
On Sun, 2013-06-23 at 22:14 +0100, Harry van Haaren wrote:

> On Sun, Jun 23, 2013 at 2:38 PM, <[hidden email]> wrote:
> > I'd like to find out the impact on signal quality when using
> different
> > network tools for jack (netjack, jacktrip).
>
> I don't think there is *any* impact on sound quality: the audio data
> on both sides is binary identical: hence no loss by transmission.
>
> So perceptually they're identical (netjack uses float as sample
> format, at the same samplerate: so just like an normal JACK patchbay
> connection, nothing is lost ). For details on encoding etc on netjack
> checkout this page
> http://trac.jackaudio.org/wiki/WalkThrough/User/NetJack
>
> Using certain encodings / bit-depths will cause reduction in sound
> quality: as defaults I think its binary identical... perhaps don't
> quote me though, I'm not an author of any of the implementations...
>
> > Question that come to mind are,
> The questions that will show the difference between netjack, netjack2
> and JackTrip are how jitter / packet loss / latency are handled /
> compensated for.
>
> HTH, -Harry

Audacity enables to show a spectrum. When I heard something that nobody
else was able to hear, for something that should be a 1:1 digital copy,
I could show that the spectrum did differ. It must be somewhere in the
Qtractor devel archive. It was a Jack connection from Qtractor out to
Qtractor in to mix down the recordings. However, I always wonder that
people want to measure things, when they don't hear a difference. Why?
If we can't hear it, than it's not an issue. Some time ago I read that
somebody wanted to verify by _watching_, if his music has the Katz
dynamic and not a loudness war dynamic. It's irrelevant what a meter
does tell us, important is what we're able to hear.

My advice is to use the ears. If the OP can't hear a difference there's
no need to check it with a tool. If a difference is audible where no
difference should be, than tools are needed for troubleshooting.

2 Cents,
Ralf


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Ralf Mardorf
On Sun, 2013-06-23 at 23:36 +0200, Ralf Mardorf wrote:

> On Sun, 2013-06-23 at 22:14 +0100, Harry van Haaren wrote:
> > On Sun, Jun 23, 2013 at 2:38 PM, <[hidden email]> wrote:
> > > I'd like to find out the impact on signal quality when using
> > different
> > > network tools for jack (netjack, jacktrip).
> >
> > I don't think there is *any* impact on sound quality: the audio data
> > on both sides is binary identical: hence no loss by transmission.
> >
> > So perceptually they're identical (netjack uses float as sample
> > format, at the same samplerate: so just like an normal JACK patchbay
> > connection, nothing is lost ). For details on encoding etc on netjack
> > checkout this page
> > http://trac.jackaudio.org/wiki/WalkThrough/User/NetJack
> >
> > Using certain encodings / bit-depths will cause reduction in sound
> > quality: as defaults I think its binary identical... perhaps don't
> > quote me though, I'm not an author of any of the implementations...
> >
> > > Question that come to mind are,
> > The questions that will show the difference between netjack, netjack2
> > and JackTrip are how jitter / packet loss / latency are handled /
> > compensated for.
> >
> > HTH, -Harry
>
> Audacity enables to show a spectrum. When I heard something that nobody
> else was able to hear, for something that should be a 1:1 digital copy,
> I could show that the spectrum did differ. It must be somewhere in the
> Qtractor devel archive. It was a Jack connection from Qtractor out to
> Qtractor in to mix down the recordings. However, I always wonder that
> people want to measure things, when they don't hear a difference. Why?
> If we can't hear it, than it's not an issue. Some time ago I read that
> somebody wanted to verify by _watching_, if his music has the Katz
> dynamic and not a loudness war dynamic. It's irrelevant what a meter
> does tell us, important is what we're able to hear.
>
> My advice is to use the ears. If the OP can't hear a difference there's
> no need to check it with a tool. If a difference is audible where no
> difference should be, than tools are needed for troubleshooting.
>
> 2 Cents,
> Ralf

PS: Sending the signal several times from A to B and B to A or just from
A to B and than a recording done with B again from A to B, would
increase eventually existing loss by rounding errors, jitter, bad cables
or what ever in what ever analog or digital signal chain might be
possible.


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Thomas Brand

Thanks Harry, Ralf for your suggestions.

I agree it's important to check by listening and trust the ears! It can
fool you though, just try to do the 'pepsi' test with a song as original
wave and the same as 256kbit/s mp3. You probably won't hear the difference
but it's there. :)

Another suggestion i got for testing is to route host A -> host B -> host
A, phase-invert signal and see what remains.

I learned jacktrip is lossless so if the network link is ok, the data
should be identical (except for rounding errors if using different bit
depth).

Cheers,
Thomas


> On Sun, 2013-06-23 at 23:36 +0200, Ralf Mardorf wrote:
>> On Sun, 2013-06-23 at 22:14 +0100, Harry van Haaren wrote:
>> > On Sun, Jun 23, 2013 at 2:38 PM, <[hidden email]> wrote:
>> > > I'd like to find out the impact on signal quality when using
>> > different
>> > > network tools for jack (netjack, jacktrip).
>> >
>> > I don't think there is *any* impact on sound quality: the audio data
>> > on both sides is binary identical: hence no loss by transmission.
>> >
>> > So perceptually they're identical (netjack uses float as sample
>> > format, at the same samplerate: so just like an normal JACK patchbay
>> > connection, nothing is lost ). For details on encoding etc on netjack
>> > checkout this page
>> > http://trac.jackaudio.org/wiki/WalkThrough/User/NetJack
>> >
>> > Using certain encodings / bit-depths will cause reduction in sound
>> > quality: as defaults I think its binary identical... perhaps don't
>> > quote me though, I'm not an author of any of the implementations...
>> >
>> > > Question that come to mind are,
>> > The questions that will show the difference between netjack, netjack2
>> > and JackTrip are how jitter / packet loss / latency are handled /
>> > compensated for.
>> >
>> > HTH, -Harry
>>
>> Audacity enables to show a spectrum. When I heard something that nobody
>> else was able to hear, for something that should be a 1:1 digital copy,
>> I could show that the spectrum did differ. It must be somewhere in the
>> Qtractor devel archive. It was a Jack connection from Qtractor out to
>> Qtractor in to mix down the recordings. However, I always wonder that
>> people want to measure things, when they don't hear a difference. Why?
>> If we can't hear it, than it's not an issue. Some time ago I read that
>> somebody wanted to verify by _watching_, if his music has the Katz
>> dynamic and not a loudness war dynamic. It's irrelevant what a meter
>> does tell us, important is what we're able to hear.
>>
>> My advice is to use the ears. If the OP can't hear a difference there's
>> no need to check it with a tool. If a difference is audible where no
>> difference should be, than tools are needed for troubleshooting.
>>
>> 2 Cents,
>> Ralf
>
> PS: Sending the signal several times from A to B and B to A or just from
> A to B and than a recording done with B again from A to B, would
> increase eventually existing loss by rounding errors, jitter, bad cables
> or what ever in what ever analog or digital signal chain might be
> possible.
>
>
> _______________________________________________
> Jack-Devel mailing list
> [hidden email]
> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
>


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Rigg-16
On Mon, Jun 24, 2013 at 01:17:31AM +0200, [hidden email] wrote:
> I agree it's important to check by listening and trust the ears! It can
> fool you though, just try to do the 'pepsi' test with a song as original
> wave and the same as 256kbit/s mp3. You probably won't hear the difference
> but it's there. :)

If I couldn't hear a difference between an original .wav file and a 256kb/s
mp3 I'd repair my monitoring equipment. In my experience mp3 encoding at
256kb/s can even change the perceived balance between elements in a mix,
sometimes requiring a remix if that is the main release format.

Back on topic, a simple null test is a good way of showing differences
between two audio files, particularly if your monitoring equipment has
insufficient resolution to reveal them when played normally. If there is
any difference over the network, something is broken.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Paul Davis



On Mon, Jun 24, 2013 at 4:27 AM, John Rigg <[hidden email]> wrote:
On Mon, Jun 24, 2013 at 01:17:31AM +0200, [hidden email] wrote:
> I agree it's important to check by listening and trust the ears! It can
> fool you though, just try to do the 'pepsi' test with a song as original
> wave and the same as 256kbit/s mp3. You probably won't hear the difference
> but it's there. :)

If I couldn't hear a difference between an original .wav file and a 256kb/s
mp3 I'd repair my monitoring equipment.

.. and join the small though not empty set of people who are capable of doing this. double blind testing of 256kbps mp3s and uncompressed digital audio generally reveal a failure to be able to differentiate, despite what is said so often in mainstream and technical publications. 

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Matthew Robbetts-2
Is the set really not empty, with a good encoder? I am surprised, frankly.


On 24 June 2013 13:17, Paul Davis <[hidden email]> wrote:



On Mon, Jun 24, 2013 at 4:27 AM, John Rigg <[hidden email]> wrote:
On Mon, Jun 24, 2013 at 01:17:31AM +0200, [hidden email] wrote:
> I agree it's important to check by listening and trust the ears! It can
> fool you though, just try to do the 'pepsi' test with a song as original
> wave and the same as 256kbit/s mp3. You probably won't hear the difference
> but it's there. :)

If I couldn't hear a difference between an original .wav file and a 256kb/s
mp3 I'd repair my monitoring equipment.

.. and join the small though not empty set of people who are capable of doing this. double blind testing of 256kbps mp3s and uncompressed digital audio generally reveal a failure to be able to differentiate, despite what is said so often in mainstream and technical publications. 

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org



_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Adrian Knoth
On Mon, Jun 24, 2013 at 02:06:42PM +0100, Matthew Robbetts wrote:


> >> If I couldn't hear a difference between an original .wav file and a
> >> 256kb/s mp3 I'd repair my monitoring equipment.
> > .. and join the small though not empty set of people who are capable of
> > doing this. double blind testing of 256kbps mp3s and uncompressed digital
> > audio generally reveal a failure to be able to differentiate, despite what
> > is said so often in mainstream and technical publications.
> Is the set really not empty, with a good encoder? I am surprised, frankly.

Generally speaking, if you're hearing impaired, your ear no longer
functions as anticipated by the psychoacoustic model. That said, the
encoder will produce something which sounds different to what you're
used to, thus making such files distinguishable.


So it's certainly the exception, not the rule.

--
mail: [hidden email]   http://adi.thur.de        PGP/GPG: key via keyserver

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Matthew Robbetts-2
Ah, good idea. I know one person who developed tinnitus and said he became more sensitive to such things after. I've no idea if tinnitus specifically is a plausible cause, but I like the explanation overall!


On 24 June 2013 15:05, Adrian Knoth <[hidden email]> wrote:
On Mon, Jun 24, 2013 at 02:06:42PM +0100, Matthew Robbetts wrote:


> >> If I couldn't hear a difference between an original .wav file and a
> >> 256kb/s mp3 I'd repair my monitoring equipment.
> > .. and join the small though not empty set of people who are capable of
> > doing this. double blind testing of 256kbps mp3s and uncompressed digital
> > audio generally reveal a failure to be able to differentiate, despite what
> > is said so often in mainstream and technical publications.
> Is the set really not empty, with a good encoder? I am surprised, frankly.

Generally speaking, if you're hearing impaired, your ear no longer
functions as anticipated by the psychoacoustic model. That said, the
encoder will produce something which sounds different to what you're
used to, thus making such files distinguishable.


So it's certainly the exception, not the rule.

--
mail: [hidden email]       http://adi.thur.de      PGP/GPG: key via keyserver

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Ralf Mardorf
In reply to this post by Adrian Knoth
On Mon, 2013-06-24 at 16:05 +0200, Adrian Knoth wrote:
> Generally speaking, if you're hearing impaired, your ear no longer
> functions as anticipated by the psychoacoustic model. That said, the
> encoder will produce something which sounds different to what you're
> used to, thus making such files distinguishable.

That's interesting. IOW in this context impaired would also be for
healthy hearing organs, nerves etc. but e.g. a different
working/filtering brain, so even a touch of Asperger, synaesthesia and
savantism etc. pp. would make MP3s not working anymore?
What does a small band frequency gap for one ear e.g. cause?

Many of the kids are hearing impaired, I always thought that this is the
reason, that they can't hear a difference. It's vice versa? If you hear
"less good" or "better" or better lets call it "different" than
averaged, that might be the reason, if you can't stand MP3s?

If so, that would explain a lot for me.

Regards,
Ralf


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Thomas Brand
In reply to this post by Matthew Robbetts-2

Hi,
well it depends on encoding quality too..
Looking at uncompressed and mp3 side by side in baudline, one of the most
obvious differences is that mp3 cuts off at say 18khz. We can hear up to
20khz it's said. When you try with closed earphones, and someone plays a
sine at 16khz (say in jaaa), i would be surprised if anyone can
consciously hear it (easy to test). But some people have very sensitive
ears indeed.
Greetings,
Thomas

> Ah, good idea. I know one person who developed tinnitus and said he became
> more sensitive to such things after. I've no idea if tinnitus specifically
> is a plausible cause, but I like the explanation overall!
>
>
> On 24 June 2013 15:05, Adrian Knoth <[hidden email]> wrote:
>
>> On Mon, Jun 24, 2013 at 02:06:42PM +0100, Matthew Robbetts wrote:
>>
>>
>> > >> If I couldn't hear a difference between an original .wav file and a
>> > >> 256kb/s mp3 I'd repair my monitoring equipment.
>> > > .. and join the small though not empty set of people who are capable
>> of
>> > > doing this. double blind testing of 256kbps mp3s and uncompressed
>> digital
>> > > audio generally reveal a failure to be able to differentiate,
>> despite
>> what
>> > > is said so often in mainstream and technical publications.
>> > Is the set really not empty, with a good encoder? I am surprised,
>> frankly.
>>
>> Generally speaking, if you're hearing impaired, your ear no longer
>> functions as anticipated by the psychoacoustic model. That said, the
>> encoder will produce something which sounds different to what you're
>> used to, thus making such files distinguishable.
>>
>>
>> So it's certainly the exception, not the rule.
>>
>> --
>> mail: [hidden email]       http://adi.thur.de      PGP/GPG: key via
>> keyserver
>>
>> _______________________________________________
>> Jack-Devel mailing list
>> [hidden email]
>> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
>>


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Rigg-16
In reply to this post by Adrian Knoth
On Mon, Jun 24, 2013 at 04:05:00PM +0200, Adrian Knoth wrote:

> On Mon, Jun 24, 2013 at 02:06:42PM +0100, Matthew Robbetts wrote:
>
>
> > >> If I couldn't hear a difference between an original .wav file and a
> > >> 256kb/s mp3 I'd repair my monitoring equipment.
> > > .. and join the small though not empty set of people who are capable of
> > > doing this. double blind testing of 256kbps mp3s and uncompressed digital
> > > audio generally reveal a failure to be able to differentiate, despite what
> > > is said so often in mainstream and technical publications.
> > Is the set really not empty, with a good encoder? I am surprised, frankly.
>
> Generally speaking, if you're hearing impaired, your ear no longer
> functions as anticipated by the psychoacoustic model. That said, the
> encoder will produce something which sounds different to what you're
> used to, thus making such files distinguishable.

Interesting theory. I can definitely hear the difference quite clearly between even
a 320kb/s mp3 and the original .wav file. That's using lame on the highest quality
setting, so it isn't as if I'm using a broken encoder. It becomes less distinguishable
if the playback system is bad enough, but it doesn't take esoteric equipment to reveal
it.

Most of the other audio engineers and many musicians I work with can hear the same
thing. That's quite interesting as it covers many styles of music like classical and
acoustic folk, not just loud stuff.

I'm more inclined to believe that those who use critical listening skills in their
work can develop the ability to hear these things through constant practice.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Paul Davis



On Mon, Jun 24, 2013 at 2:54 PM, John Rigg <[hidden email]> wrote:

Interesting theory. I can definitely hear the difference quite clearly between even
a 320kb/s mp3 and the original .wav file.

Have you ever actually tried this double blind? I know at least a dozen folks who SWORE that they could tell the difference, and indeed, in non-blind tests, they could accurately identify certain "artifacts" in the mp3. in double blind testing, they all lost the ability to do better than 50% ....
 
I'm more inclined to believe that those who use critical listening skills in their
work can develop the ability to hear these things through constant practice.

There is no evidence that acoustic perception can be trained There may be specific characteristics of compressed (or even just digitally sampled audio) that one can learn to look for. The problem is that every time some "golden eared" listener attempts to explan what they are, double blind testing suggests that they are wrong.

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Rigg-16
On Mon, Jun 24, 2013 at 03:45:10PM -0400, Paul Davis wrote:
> There is no evidence that acoustic perception can be trained There may be
> specific characteristics of compressed (or even just digitally sampled
> audio) that one can learn to look for. The problem is that every time some
> "golden eared" listener attempts to explan what they are, double blind
> testing suggests that they are wrong.

I haven't done a formal double blind test, but I regularly test colleagues
(and vice versa) by playing mp3s and wav files without telling the listener
which is which. The success rate with certain listeners is close to 100%.

Double blind tests prove that a particular set of listeners in a particular
set of circumstances can or can't hear something. The circumstances are
important. I know that if I'm tired or stressed I find it more difficult to
concentrate well enough to hear subtle differences, for example.

I agree that physical perception of sounds can not be trained, but I do
believe that the ability to distinguish certain elements in what we are
hearing can be improved by training.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Emmas
On 24/06/2013 21:17, John Rigg wrote:
>
> I haven't done a formal double blind test, but I regularly test colleagues
> (and vice versa) by playing mp3s and wav files without telling the listener
> which is which. The success rate with certain listeners is close to 100%.
>

I must admit, I'm with John on this one.  If a person can be taught to
play a musical instrument (and especially if they can be taught to be a
sound mixer) it stands to reason that they can be taught to hear subtle
nuances in sound.  In both cases, it goes with the territory.

Many years ago when I worked on AMS Neve's AudioFile DAW, our R&D guys
came up with a new design to improve the phase response of its
brick-wall filters, after countless complaints from BBC sound mixers.  
Subtleties in phase are notoriously difficult to detect and the guys
were pretty confident that no-one would be able to hear the difference,
even though the phase response was measurably more linear - but I could
hear it.  In testing, I could tell with 100% reliability whether the old
filter or the new filter was being used.  The tests were (literally)
double-blind in the sense that I was forced to wear a blindfold AND to
wear headphones, so that I couldn't be influenced by knowing which
system was playing - nor by the direction or acoustic characteristics of
the sound coming from different loudspeakers.  What's interesting is
that out of an R&D dept of over 40 people, I was the only one who could
hear the difference!

Of course, I was a lot younger back then.  Another important factor is
that one's hearing changes markedly with age.  At any given time there's
probably only a small percentage of the population who are even
physically able to hear such nuances, even if they want to.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

Ralf Mardorf
On Tue, 2013-06-25 at 14:59 +0100, John Emmas wrote:

> On 24/06/2013 21:17, John Rigg wrote:
> >
> > I haven't done a formal double blind test, but I regularly test colleagues
> > (and vice versa) by playing mp3s and wav files without telling the listener
> > which is which. The success rate with certain listeners is close to 100%.
> >
>
> I must admit, I'm with John on this one.  If a person can be taught to
> play a musical instrument (and especially if they can be taught to be a
> sound mixer) it stands to reason that they can be taught to hear subtle
> nuances in sound.  In both cases, it goes with the territory.
>
> Many years ago when I worked on AMS Neve's AudioFile DAW, our R&D guys
> came up with a new design to improve the phase response of its
> brick-wall filters, after countless complaints from BBC sound mixers.  
> Subtleties in phase are notoriously difficult to detect and the guys
> were pretty confident that no-one would be able to hear the difference,
> even though the phase response was measurably more linear - but I could
> hear it.  In testing, I could tell with 100% reliability whether the old
> filter or the new filter was being used.  The tests were (literally)
> double-blind in the sense that I was forced to wear a blindfold AND to
> wear headphones, so that I couldn't be influenced by knowing which
> system was playing - nor by the direction or acoustic characteristics of
> the sound coming from different loudspeakers.  What's interesting is
> that out of an R&D dept of over 40 people, I was the only one who could
> hear the difference!
>
> Of course, I was a lot younger back then.  Another important factor is
> that one's hearing changes markedly with age.  At any given time there's
> probably only a small percentage of the population who are even
> physically able to hear such nuances, even if they want to.

That's a useless discussion we had several times. I have given up to
talk about this and many other issues. Btw. it's not really important
regarding to MP3, but regarding to bugs we have for Linux audio. Issues
are audible for engineers even with aged and less good ears, but the
consensuses for Linux audio is, that it's impossible that something is
as it shouldn't be. When using Linux audio one has to accept this
policy ;). People might be reputable audio engineers, but that doesn't
mean that people of the audio Linux community have the same impression
of those community members, but IMO this isn't important, no pro-audio
employer does take care about the opinions from Linux audio community
members.

However, consensus for MP3 is that there are no audible differences.
Many, even amateurs are able to hear differences, but this is ignored.
Instead of talking about this again and again, we could send us links to
mail archives were everybody of us already made statements.


_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Rigg-16
In reply to this post by John Emmas
On Tue, Jun 25, 2013 at 02:59:10PM +0100, John Emmas wrote:

> Many years ago when I worked on AMS Neve's AudioFile DAW, our R&D guys  
> came up with a new design to improve the phase response of its  
> brick-wall filters, after countless complaints from BBC sound mixers.  
> Subtleties in phase are notoriously difficult to detect and the guys  
> were pretty confident that no-one would be able to hear the difference,  
> even though the phase response was measurably more linear - but I could  
> hear it.  In testing, I could tell with 100% reliability whether the old  
> filter or the new filter was being used.  The tests were (literally)  
> double-blind in the sense that I was forced to wear a blindfold AND to  
> wear headphones, so that I couldn't be influenced by knowing which  
> system was playing - nor by the direction or acoustic characteristics of  
> the sound coming from different loudspeakers.  What's interesting is  
> that out of an R&D dept of over 40 people, I was the only one who could  
> hear the difference!

And it was quite likely a difference in the level of some side effect like
pre-echo that you could hear, rather than the difference in phase response
itself. That's what makes it so hard to convince those that don't hear the
problem. The theory says something is inaudible, but they aren't looking
at what else is going on.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Emmas
On 25 Jun 2013, at 20:51, John Rigg wrote:

> On Tue, Jun 25, 2013 at 02:59:10PM +0100, John Emmas wrote:
>> Many years ago when I worked on AMS Neve's AudioFile DAW, our R&D guys  
>> came up with a new design to improve the phase response of its  
>> brick-wall filters, after countless complaints from BBC sound mixers.  
>> Subtleties in phase are notoriously difficult to detect and the guys  
>> were pretty confident that no-one would be able to hear the difference,  
>> even though the phase response was measurably more linear - but I could  
>> hear it.  In testing, I could tell with 100% reliability whether the old  
>> filter or the new filter was being used.  The tests were (literally)  
>> double-blind in the sense that I was forced to wear a blindfold AND to  
>> wear headphones, so that I couldn't be influenced by knowing which  
>> system was playing - nor by the direction or acoustic characteristics of  
>> the sound coming from different loudspeakers.  What's interesting is  
>> that out of an R&D dept of over 40 people, I was the only one who could  
>> hear the difference!
>
> And it was quite likely a difference in the level of some side effect like
> pre-echo that you could hear, rather than the difference in phase response
> itself. That's what makes it so hard to convince those that don't hear the
> problem. The theory says something is inaudible, but they aren't looking
> at what else is going on.
>

Exactly.  Too many people believe that if the frequency response is adequate to cover the range of human hearing and the dynamic range is similarly adequate, that's enough - whereas in fact, things can be audible without them being related to either of those things.  In this particular case the new filters gave a much more "defined" stereo image, instead of the slightly "muddy" image produced by our original filters.  I found it so easy to detect that I was genuinely surprised at how few others could hear it.

John
_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
Reply | Threaded
Open this post in threaded view
|

Re: Network Audio Transmission - Quality

John Emmas
On 26/06/2013 05:36, John Emmas wrote:
In this particular case the new filters gave a much more "defined" stereo image, instead of the slightly "muddy" image produced by our original filters.  I found it so easy to detect that I was genuinely surprised at how few others could hear it.


In fact, to add to that...  it's entirely possible to hear things for which there is (literally) no mechanical or scientific explanation.  For example, a famous TV company where I used to be a sound mixer, kitted out all its studios with some new loudspeakers.  For me, the image had no "depth".  It was as if the sound was mono and coming from a single point source.  Of course, technically, there's nothing in a stereo signal that can convey depth - and yet I always perceive depth when listening to stereo material.  Out of all the sound mixers, only me and one other person could hear the lack of depth.  Nobody else could detect it.

In truth, the perception of depth was probably inside our brains - but something about those speakers wasn't stimulating whatever it was that produced that effect in us.  What this proves is that psychoacoustics are very real.  Hearing isn't (simply) a mechanical transmission process.  To quote from Wikipedia on the subject of psychoacoustics:-

" Hearing is not a purely mechanical phenomenon of wave propagation, but is also a sensory and perceptual event; in other words, when a person hears something, that something arrives at the ear as a mechanical sound wave traveling through the air, but within the ear it is transformed into neural action potentials. These nerve pulses then travel to the brain where they are perceived. Hence, in many problems in acoustics, such as for audio processing, it is advantageous to take into account not just the mechanics of the environment, but also the fact that both the ear and the brain are involved in a person’s listening experience. "

How often that gets forgotten!

John

_______________________________________________
Jack-Devel mailing list
[hidden email]
http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org
12