PDA

View Full Version : Feature request: Polarity control



P Floding
2006-01-11, 20:14
Hi all!

I request a feature to control signal polarity on a per-album basis, as well as on a per-song basis. (Polarity specification for a song would override any album polarity.)

Also, an overall-polarity switch directly from the remote would also be useful for those that don't have a polarity swith on their amp. (This would be used when deciding on the correct polarity for an album or individual song.)

The background is that with digital room correction (TacT RCS 2.2x) I can now easily hear what is the correct polarity on many of my albums. And very often the default polarity is not the correct one!

My pre-amp allows polarity switching, but it is 1) cumbersome, and 2) easy to forget what you concluded was the best setting.

I might implement this feature, unless it turns out to be very difficult. However, I'd like to check if anyone here has done it already, or is working on it!

Regards

P Floding
2006-01-12, 05:10
Hi all!

I request a feature to control signal polarity on a per-album basis, as well as on a per-song basis. (Polarity specification for a song would override any album polarity.)

A overall-polarity switch directly from the remote would also be useful for those that don't have a polarity swith on their amp.

The background is that with digital room correction (TacT RCS 2.2x) I can now easily hear what is the correct polarity on many of my albums. And very often the default polarity is not the correct one!

My pre-amp allows polarity switching, but it is 1) cumbersome, and 2) easy to forget what you concluded was the best setting.

I might implement this feature, unless it turns out to be very difficult. However, I'd like to check if anyone here has done it already, or is working on it!

Regards

Any interest at all?

Anyone with a high-end setup, using an SB, ought to be interested as this would automate playback with correct polarity once you have decided on the correct one!

Comments on the feasibility of implementing this, and hints on the best approach, would also be welcome from those familiar with the various processing steps involved in getting the bits from disk to the digital/analouge output.

afblaster
2006-01-12, 05:20
I have a 'high-end' system and would like to be able to switch the signal polarity.

P Floding
2006-01-12, 05:23
I have a 'high-end' system and would like to be able to switch the signal polarity.

ok, I'm glad I'm not alone!

Michaelwagner
2006-01-12, 05:59
The "correct" way to request features like this is to file a bug report at http://bugs.slimdevices.com/ with the severity set to "enhancement".

You will need to create a userid and password to use bugzilla (the name of the bug tracking software).

If you're not sure what to do, people on this forum can help you.

This request, if I understand correctly what you're asking for, will require firmware changes, which means it's only likely to be provided for SB2/SB3 boxes (older ones are pretty much closed to development now except for unusual cases).

Only a few people have the tools to work on the firmware, so it won't get done as fast as some other feature requests.

Once you've filed the report, you can watch it's progress. It emails things to use as status changes. You can also ask other people to vote for it, and the more votes, the higher priority it is assigned.

Hope this helps.

Néstor
2006-01-12, 06:01
I created a feature request for this some time ago (bug 829), and it was later merged with this bug: 1976

It still with NEW status.

Regards
Nestor

Michaelwagner
2006-01-12, 06:09
sorry, haven't had my morning coffee yet, didn't think to check.

http://bugs.slimdevices.com/show_bug.cgi?id=829 has been closed as a dup of
http://bugs.slimdevices.com/show_bug.cgi?id=1976

You can vote for it.

If I understand correctly, a part of the problem has been fixed, there was an inversion in the signal path that made the analogue and digital signals 180° out of phase with each other. That's what's been fixed so far. The rest of it, making phase configurable channel by channel and track by track, has not been addressed yet.

So people can vote for it ...

P Floding
2006-01-12, 06:35
sorry, haven't had my morning coffee yet, didn't think to check.

http://bugs.slimdevices.com/show_bug.cgi?id=829 has been closed as a dup of
http://bugs.slimdevices.com/show_bug.cgi?id=1976

You can vote for it.

If I understand correctly, a part of the problem has been fixed, there was an inversion in the signal path that made the analogue and digital signals 180° out of phase with each other. That's what's been fixed so far. The rest of it, making phase configurable channel by channel and track by track, has not been addressed yet.

So people can vote for it ...

OK, thanks for you help!
It is not exactly what I'm proposing, but in the same realm.

I will open a new "bug" report detailing my request. (Or maybe I should just amend the existing request with the additional requests..)

Regards

Michaelwagner
2006-01-12, 06:47
Well, the existing bug was opened to solve one issue and morphed into another, so I agree it's a bit confusing.

Before you enter a bug, then, please tell us what you'd like beyond what's in the current request.

Knowing a bit more about how the slim architecture works, we can tell you if your needs are best addressed by a new bug or an add-on to the existing one.

Néstor
2006-01-12, 07:31
I think bug 829 is really close to what you are asking. Maybe we can try reopening the bug.

bludragon
2006-01-12, 09:13
829 sounds like your FR, the other bug would mean you had to manually switch for each track, which would defeat the point.

A workaround would be to use the foobar plugin here:

http://www.hydrogenaudio.org/forums/index.php?showtopic=28924

to invert the source files where required, instead of tagging them with polarity.

Michaelwagner
2006-01-12, 09:35
the other bug would mean you had to manually switch for each track, which would defeat the point.
No, the bug has been partially worked on. One part of the problem was that the digital outputs were reversed from the analogue. That part has been fixed. The other part, which says, wait for a configuration option, would require more firmware support, and then it could be implemented on the server on whatever granularity is required, including track by track.

P Floding
2006-01-12, 12:38
I think bug 829 is really close to what you are asking. Maybe we can try reopening the bug.

Yes, it is very similar. Just a bit less specific. It is not really a duplicate of the other bug report either.

I don't agree with the assessment, in that request, that only "sensitive" people would be able to make use of this feature. Rather that a system's resolution has to be pretty good to make use of it. A headphone based system that easily could benefit might not cost very much at all.

Who knows, in the future maybe even public polarity information on various CDs could be made available for the benefit of those who don't know what to listen for, but might still enjoy the improvement in sound quality? (Is the CD info database format extensible?)

With a high resolution system (a phase coherent one) the difference between correct and incorrect polarity on some recordings are very dramatic. It is the difference from listening to a hifi, to "being there"!

As the SB is taken up more and more by audiophiles, who typically set up serveral SBs in their homes (and are habitual upgraders), it may be a good idéa to cater to them by including a feature such as this. The SB has a good following at least amongst TacT users -as the TacT room correction system (pre amp) has digital inputs.

Regards

kdf
2006-01-12, 12:50
Quoting P Floding <P.Floding.21jecn (AT) no-mx (DOT) forums.slimdevices.com>:

>
> Néstor Wrote:
>> I think bug 829 is really close to what you are asking. Maybe we can try
>> reopening the bug.
>
> Yes, it is very similar. Just a bit less specific. It is not really a
> duplicate of the other bug report either.

you can choose to dilute the requests by making your own, having your
name in lights as it were, and being very specific.

OR, you can choose to add yoru comments to shape the existing
enahancement request. The currently open request asks for a user
toggle. Having it automatic is not that far off from that. Replaygain
handling already uses code to automatically determine album/track so
were there ever to be a polarity setting in the db (my mind hurts
thinking of the percentage of users who'd care), it could use the same
logic.

-kdf

Michaelwagner
2006-01-12, 13:12
With a high resolution system (a phase coherent one) the difference between correct and incorrect polarity on some recordings are very dramatic.
I have one of those test setup CDs that plays honks and chirps at different frequencies, plays into the left and right channel so you can see if you reversed your speaker wires, etc.

It has one section where the voice says
"This section is recorded in phase" and then
"This section is recorded out of phase"

Even with pretty lousy equipment (DJ system in an auditorium for a dance kind of equipment) I can hear the difference. So I don't think you have to be all that high-end to hear a phasing problem.

The thing I don't get is, do people really release recordings with the phasing wrong?

Alex Twisleton-Wykeham-Fiennes
2006-01-12, 13:30
On Thu 12 January 2006 20:12, Michaelwagner wrote:
> P Floding Wrote:
> > With a high resolution system (a phase coherent one) the difference
> > between correct and incorrect polarity on some recordings are very
> > dramatic.
>
> I have one of those test setup CDs that plays honks and chirps at
> different frequencies, plays into the left and right channel so you can
> see if you reversed your speaker wires, etc.
>
> It has one section where the voice says
> "This section is recorded in phase" and then
> "This section is recorded out of phase"
>
> Even with pretty lousy equipment (DJ system in an auditorium for a
> dance kind of equipment) I can hear the difference. So I don't think
> you have to be all that high-end to hear a phasing problem.
>
> The thing I don't get is, do people really release recordings with the
> phasing wrong?

I agree that signals being out of phase between channels is blindingly
obvious, but I believe P Floding is talking about absolute system phase, ie
switching the entire phase of the system.

I've come across this mainly when working with live sound systems when micing
transient sources like drums. Ideally you want your initial wavefront from
the drum source to be reproduced by a "push" with the speaker cones in the
system thereby producing a much better reproduction of the original transient
than you would get by all your cones moving away from the listener and trying
to reproduce the effect of the drums by "sucking".

I wasn't sure about the application of this to global phase on recordings -
most modern mastering is so horribly multi-band compressed and limited that I
dread to think what has happened to the original phase coherency of the mix,
but I did a bit of googling and the name which kept coming up was Charles
Wood who released a book named The Wood Effect. Interesting page about it
here:-

http://www.positive-feedback.com/Issue1/cjwoodeffect.htm

Makes quite a lot of good points, but also points out that the majority of
cross overs in the majority of speakers make such a pigs ear of phase
coherency that you're kind of fighting a losing battle...

Anyway, enough woffling - I'm off to go and swap my speaker terminals around
and try and decide if I like it better or worse!

Alex

JJZolx
2006-01-12, 13:41
The thing I don't get is, do people really release recordings with the phasing wrong?
I've read that as many as half of all recordings are out of phase, so apparently it's not much of a priority to the recording industry. And also probably why many audiophile preamps include a phase switch. See the following:

http://db.audioasylum.com/cgi/m.mpl?forum=amp&n=85970

pfarrell
2006-01-12, 13:55
Alex Twisleton-Wykeham-Fiennes wrote:
> On Thu 12 January 2006 20:12, Michaelwagner wrote:
>>The thing I don't get is, do people really release recordings with the
>>phasing wrong?
>
> I agree that signals being out of phase between channels is blindingly
> obvious, but I believe P Floding is talking about absolute system phase, ie
> switching the entire phase of the system.

The feature request only makes sense after your system, and its speakers
are right.

> I've come across this mainly when working with live sound systems when micing
> transient sources like drums. Ideally you want your initial wavefront from
> the drum source to be reproduced by a "push" with the speaker cones in the
> system

Right, a kick drum should start out towards the listener. The speaker
cone should move forward.

As Alex wrote, after all the EQ and effects that get done to most
modern recordings, the phase gets all wacko. It is
impossible (almost? totally?) to do EQ without introducing phase artifacts.

As a plugin or effect, it is pretty easy, just flip the data.
I forget the details of PCM value encoding, I think they are
just 0 to 2^16-1, so you have to arithmetically change it.
Still just one byte operation per 16 bit sample.

--
Pat
http://www.pfarrell.com/music/slimserver/slimsoftware.html

Mark Lanctot
2006-01-12, 14:02
I believe what the OP referred to here is what is
called "absolute polarity".

This is outside of how your speakers are wired and
assumes that speakers need to be moving out to
produce certain sounds, not in.

From what I have read at AVS Forum, it's pretty
controversial. Only high-end equipment has
controls to alter polarity, and even then, only
certain people claim to hear a difference.

It's a fundamental idea affecting how speakers
work. We think that to reproduce a drum, for
example, the speaker cone needs to move the same
way the drum head moves. But since the drum head
moves down, should the speaker cone move in or
out? That is the concept of absolute polarity.

Speaker phase, i.e. wiring, has very direct and
very measurable effects if the speakers are wired
unalike since it creates constructive and
destructive interference when sound waves from one
speaker meet sound waves from the other. Absolute
polarity assumes your speakers are wired in phase
but that sound can only fundamentally be produced
one way by speakers.

As I said, it's controversial. There are many who
believe it is a myth.

http://www.avsforum.com/avs-vb/showthread.php?t=595226&highlight=absolute+polarity

....one myth proponent in that thread is Ethan
Winer, who is quite an expert.

Michaelwagner wrote:
> P Floding Wrote:
>
>>With a high resolution system (a phase coherent one)
the difference
>>between correct and incorrect polarity on some
recordings are very
>>dramatic.
>
> I have one of those test setup CDs that plays honks
and chirps at
> different frequencies, plays into the left and right
channel so you can
> see if you reversed your speaker wires, etc.
>
> It has one section where the voice says
> "This section is recorded in phase" and then
> "This section is recorded out of phase"
>
> Even with pretty lousy equipment (DJ system in an
auditorium for a
> dance kind of equipment) I can hear the difference.
So I don't think
> you have to be all that high-end to hear a phasing
problem.
>
> The thing I don't get is, do people really release
recordings with the
> phasing wrong?
>
>

--
___________________________________


Mark Lanctot
___________________________________

JJZolx
2006-01-12, 14:06
I've added the following comment to bug/request #1976. Hopefully it would cover all needs in this issue:


Ideally, this should _not_ be a simple server-wide preference. It should be
implemented as configurable in several places. Also, despite the bug summary,
this should apply to both the analog and digital output streams simulatneously.

1) As a player pref to compensate for any polarity inversion in the playback
system used with the player.

2) A tag should be designated that can be used to invert polarity on a
track-by-track basis and should be stored permanently in the tracks table.
This should probably be combined with the option in number 1 so that two
inversions add up to none.

3) A means of inverting the polarity should be implemented via the remote UI so
that a listener could invert it on the fly from the listening chair. This
would simply shift phase back and forth regardless of the combination of 1 and
2.
If it's important to you, then vote for the bug. Don't just CC yourself. By voting you automatically get CC'd with all new comments and changes to the bug.

http://bugs.slimdevices.com/show_bug.cgi?id=1976

P Floding
2006-01-12, 18:57
I've added the following comment to bug/request #1976. Hopefully it would cover all needs in this issue:


If it's important to you, then vote for the bug. Don't just CC yourself. By voting you automatically get CC'd with all new comments and changes to the bug.

http://bugs.slimdevices.com/show_bug.cgi?id=1976

JJZolx, thanks for your addition! I will vote for it, of course!

Regarding some of the other comments above:

Yes, I'm talking about absolute phase, not phase error between left and right channel -if you can't hear that, you have a seriously poor system, or hearing, as that totally defocusses voices, for example!

About the comment on instruments being pointed downwards: It doesn't matter for this issue at all, since the microphone would be positioned in the direction the listener would be, and hence the microphone will receive the polarity that the listener would receive.

For close-micked multimicrophone recordings, polarity switching may not make any difference. However, I have noticed that even on some studio recordings it can still make a difference. Perhaps all tracks were recorded correctly, but some late stage in the mixing/mastering inverts the ready-mixed signal inadvertently.

It is true that you system has to be pretty good to hear the difference. I had difficulty hearing it before i got the TacT RCS. I imagine headphones would do the job at a lower cost, however.

Alex Twisleton-Wykeham-Fiennes
2006-01-13, 02:36
On Fri 13 January 2006 01:57, P Floding wrote:
> JJZolx Wrote:
> > I've added the following comment to bug/request #1976. Hopefully it
> > would cover all needs in this issue:
> >
> >
> > If it's important to you, then vote for the bug. Don't just CC
> > yourself. By voting you automatically get CC'd with all new comments
> > and changes to the bug.
> >
> > http://bugs.slimdevices.com/show_bug.cgi?id=1976
>
> JJZolx, thanks for your addition! I will vote for it, of course!
>
> Regarding some of the other comments above:
>
> Yes, I'm talking about absolute phase, not phase error between left and
> right channel -if you can't heat that, you have a seriously poor system,
> or hearing, as that totally defocusses voices, for example!

There is a school of thought that says that the fact that the inversion of a
some steady state non-symmetric signals being audibly different (which
appears to be the most common proof of absolute phase being important) is
proof that loudspeakers are not quite as linear and symetrical as they should
be so maybe with a "perfect system" it would be a different story? (I jest
slightly)

> About the comment on instruments being pointed downwards: It doesn't
> matter for this issue at all, since the microphone would be positioned
> in the direction the listener would be, and hence the microphone will
> receive the polarity that the listener would receive.
>
> For close-micked multimicrophone recordings, polarity switching may not
> make any difference. However, I have noticed that even on some studio
> recordings it can still make a difference. Perhaps all tracks were
> recorded correctly, but some late stage in the mixing/mastering inverts
> the ready-mixed signal inadvertently.

I would be amazed if any multi-microphone recording (even so-called
time-aligned multi-mic classical recordings) came anywhere close enough to be
phase coherent in themselves so any improvement would be due to chance
psycho-accoustic affects rather than any "absolute correctness".

> It is true that you system has to be pretty good to hear the
> difference. I had difficulty hearing it before i got the TacT RCS. I
> imagine headphones would do the job at a lower cost, however.

Just out of interest - are you switching your phase before you go into the
TacT or after (ie on the room corrected signal)? Do you have an option on
this? I would be very interested to know if there is a difference on your
system. I could imagine that the room specific processing that the TacT
performs on the digital signal would be much more sensitive to absolute phase
than the signal from the source that is being fed into the processor.

Alex

P Floding
2006-01-13, 19:35
On Fri 13 January 2006 01:57, P Floding wrote:
> JJZolx Wrote:
> > I've added the following comment to bug/request #1976. Hopefully it
> > would cover all needs in this issue:
> >
> >
> > If it's important to you, then vote for the bug. Don't just CC
> > yourself. By voting you automatically get CC'd with all new comments
> > and changes to the bug.
> >
> > http://bugs.slimdevices.com/show_bug.cgi?id=1976
>
> JJZolx, thanks for your addition! I will vote for it, of course!
>
> Regarding some of the other comments above:
>
> Yes, I'm talking about absolute phase, not phase error between left and
> right channel -if you can't heat that, you have a seriously poor system,
> or hearing, as that totally defocusses voices, for example!

There is a school of thought that says that the fact that the inversion of a
some steady state non-symmetric signals being audibly different (which
appears to be the most common proof of absolute phase being important) is
proof that loudspeakers are not quite as linear and symetrical as they should
be so maybe with a "perfect system" it would be a different story? (I jest
slightly)


Sure. However, we only have loudspeakers, and if they sound better with correct polarity, then that is what we should use!

Personally, I believe, based on basic physics, that an initial transient that pushes rather than pulls would sound more realistic. I have seen more sophisticated theoretical discussions on this subject, having to do with doppler effect, etc, if I remember correctly. (Since air is a compressable medium, the air does not behave exactly identically in the compression phase as in the expansion phase when "vibrating" to a sound. I'm admittedly on thin ice now, regarding my own knowledge in this area.)




> About the comment on instruments being pointed downwards: It doesn't
> matter for this issue at all, since the microphone would be positioned
> in the direction the listener would be, and hence the microphone will
> receive the polarity that the listener would receive.
>
> For close-micked multimicrophone recordings, polarity switching may not
> make any difference. However, I have noticed that even on some studio
> recordings it can still make a difference. Perhaps all tracks were
> recorded correctly, but some late stage in the mixing/mastering inverts
> the ready-mixed signal inadvertently.

I would be amazed if any multi-microphone recording (even so-called
time-aligned multi-mic classical recordings) came anywhere close enough to be
phase coherent in themselves so any improvement would be due to chance
psycho-accoustic affects rather than any "absolute correctness".


I'm not so sure. If each channel of a multi-mic recording is in correct phase, then the result is simply a superpositioning of these channels. Sure, individual distances (time and volume), and low level ambient information shared between the channels would be a bit of a mish-mash, but the fact would remain that the leading edge of transients from all channels (from the instruments dominating each channel) would be in the correct direction.




> It is true that you system has to be pretty good to hear the
> difference. I had difficulty hearing it before i got the TacT RCS. I
> imagine headphones would do the job at a lower cost, however.

Just out of interest - are you switching your phase before you go into the
TacT or after (ie on the room corrected signal)? Do you have an option on
this? I would be very interested to know if there is a difference on your
system. I could imagine that the room specific processing that the TacT
performs on the digital signal would be much more sensitive to absolute phase
than the signal from the source that is being fed into the processor.

Alex

That is an interesting point. However, if you regard the TacT pre, the poweramp, and the speakers as a "black box" (let's say, a digitally corrected active loudspeaker), then this question would be pointless. If the corrected system gives much better performance with a input signal of the correct polarity, then that is what we should use!

I do understand your point though. Is the digital room correction doing something that exagerates polarity differences, more than a perfect phase coherent speaker in an ideal listening room would have done? I don't know. I do know that people with very good systems, but no digital room correction, report the same type of audible differences between listening to correct or incorrect polarity. (I should say that "correct" here means, best sounding! I have yet to verify that the best sound corresponds to any recording's correct absolute polarity. Something that would be an interesting project! Correlating phase-data for various CDs with other audiophiles might be a good start.)

Also, I do not know if the polarity inversion of the TacT is done before or after the digital room correction processing.

tom permutt
2006-01-13, 19:57
If you were playing the CD live, I can see why you would want such a switch. But you're not. You're ripping it and playing it later. So, if you think the waveform is upside down, why don't you just flip it over before you send it to the Squeezebox?

P Floding
2006-01-13, 20:11
If you were playing the CD live, I can see why you would want such a switch. But you're not. You're ripping it and playing it later. So, if you think the waveform is upside down, why don't you just flip it over before you send it to the Squeezebox?

Firstly, to allow the data extracted from the CD to stay in unaltered form. Perhaps to later re-create a lost CD, or for any other of a number of reasons that you do not want to alter the original data.

Secondly, to establish the correct polarity one has to listen, and to be able to switch polarity while listening. Even if you have the facility to swith polarity while listening, it would be very cumbersome to go back and alter the entire ripped contents. (Perhaps only to find out later that some songs on that album needed a different polarity setting.)

In many ways it makes much more sense to do polarity switching as a processing stage.

Michaelwagner
2006-01-13, 20:35
I think Tom's question amounts to questioning where the "flipping" should be done, in the squeezebox or in the server.

The problem is you have to essentially decode a stereo signal in order to flip the phase of one side. The squeezebox is already doing this decoding, so adding a bit to each channel saying if the signal should be flipped is much "cheaper" in resources than taking the signal apart on the server, flipping one side, putting it all back together, then having the squeezebox do the decoding yet again.

If that wasn't Tom's point, then just ignore what I wrote. ...

tom permutt
2006-01-14, 04:10
I think Tom's question amounts to questioning where the "flipping" should be done, in the squeezebox or in the server.

The problem is you have to essentially decode a stereo signal in order to flip the phase of one side. The squeezebox is already doing this decoding, so adding a bit to each channel saying if the signal should be flipped is much "cheaper" in resources than taking the signal apart on the server, flipping one side, putting it all back together, then having the squeezebox do the decoding yet again.

If that wasn't Tom's point, then just ignore what I wrote. ...

Mainly I was trying to make sure I understood mathematically what we were talking about. Thanks, P Floding and Michaelwagner, for clarifying that.

I can see that it would be convenient to be able to listen to both versions (faster than by switching between two versions of the same file) until you had it figured out. I don't get the point, though, about "unaltered" data. Nobody said you had to throw away the original, and the transformation is losslessly invertible anyway.

So, yes, you could do it in the server, or before the server. It's a trivial calculation, it doesn't need to be done in real time, and CPU cycles on the box and on the computer don't cost the same, so the comparison of "resources" may not be fair.

Aside from the practicalities, which don't seem insurmountable either way, I am wondering if we are being swayed by a psychological distinction between seemingly clean and dirty ways of producing identical bits at the DAC and identical voltages at the speaker terminals.

P Floding
2006-01-14, 06:09
I think Tom's question amounts to questioning where the "flipping" should be done, in the squeezebox or in the server.

The problem is you have to essentially decode a stereo signal in order to flip the phase of one side. The squeezebox is already doing this decoding, so adding a bit to each channel saying if the signal should be flipped is much "cheaper" in resources than taking the signal apart on the server, flipping one side, putting it all back together, then having the squeezebox do the decoding yet again.

If that wasn't Tom's point, then just ignore what I wrote. ...

Agree, apart from the "one channel" part. Having a flag for each channel is fine, for maximum flexibility. However the problem with incorrect polarity is not a problem where just one channel is out of phase!

If one channel is out of phase with regard to the other channel, then the result is a massive sound degradation capable of making you feel sick after a while! In addition base information would cancel out if you use a single sub woofer, leaving you with no sub base at all. I have never heard of any CD being released with such a gross error.

The sound degradation due to both channels being inverted is much more subtle, and can be hard or impossible to hear on a simpler system.

P Floding
2006-01-14, 06:19
Mainly I was trying to make sure I understood mathematically what we were talking about. Thanks, P Floding and Michaelwagner, for clarifying that.

I can see that it would be convenient to be able to listen to both versions (faster than by switching between two versions of the same file) until you had it figured out. I don't get the point, though, about "unaltered" data. Nobody said you had to throw away the original, and the transformation is losslessly invertible anyway.

So, yes, you could do it in the server, or before the server. It's a trivial calculation, it doesn't need to be done in real time, and CPU cycles on the box and on the computer don't cost the same, so the comparison of "resources" may not be fair.

Aside from the practicalities, which don't seem insurmountable either way, I am wondering if we are being swayed by a psychological distinction between seemingly clean and dirty ways of producing identical bits at the DAC and identical voltages at the speaker terminals.

Mucking about with the large data files on the hard disk is something that I think should be avoided.

Flipping the data on the server side, in real time as the data is being sent to the SB, is fine as a solution, provided that it is practical to implement. This is the solution I first imagined, before it was mentioned that a firmware toggle would be easier (?).

Michaelwagner
2006-01-14, 07:50
I think we should let the boffins worry about the implementation details (where, on which platform, it is cheapest to do the transformation) and concentrate instead on the important user elements of how it should work.

It seems like one bit per channel for the music, and one bit per channel for the player (in case for some reason the room is wired badly) would work. That way, if you get an inversion for the song, and an inversion for the room, they would cancel. That was the suggestion made in the feature request, and it seems feasible and sensible.

Any thoughts?

P Floding
2006-01-14, 09:51
I think we should let the boffins worry about the implementation details (where, on which platform, it is cheapest to do the transformation) and concentrate instead on the important user elements of how it should work.

It seems like one bit per channel for the music, and one bit per channel for the player (in case for some reason the room is wired badly) would work. That way, if you get an inversion for the song, and an inversion for the room, they would cancel. That was the suggestion made in the feature request, and it seems feasible and sensible.

Any thoughts?

I agree. (Noting that you are indeed talking "a bit" about implementation ;-) )

Anyway, implementation matters insofar that anyone could implement server side changes (Slimserver is open source, right?), wheras if anything at all needs to be done in firmware the picture is different.

Michaelwagner
2006-01-14, 12:01
I agree with what you've said, but I do think it's appropriate to get the "externals" "right" before delving into internals.

Otherwise one tends to wind up with "it was easy to implement" rather than "it works right" implementations.

tom permutt
2006-01-14, 12:04
One way of looking at it is, we have a new, lossless file format for representing what we want to hear. Instead of, say, FLAC, we have possibly-upside-down FLAC, in which the correct waveform is represented either directly in FLAC or as its inverse, and the file also is tagged to say which it is. So, you don't even have to hack the server if you don't want to: you just need a plug-in transcoder for yet another file format.

P Floding
2006-01-14, 12:39
One way of looking at it is, we have a new, lossless file format for representing what we want to hear. Instead of, say, FLAC, we have possibly-upside-down FLAC, in which the correct waveform is represented either directly in FLAC or as its inverse, and the file also is tagged to say which it is. So, you don't even have to hack the server if you don't want to: you just need a plug-in transcoder for yet another file format.

Elegant, if you only care about FLAC. I assume this proposed solution would be a bit messy if you wanted to support polarity flipping on all formats?

In addition, transcoding removes some features, such as FF/RW, I believe? Also, would it not require flagging the song in some sligtly messy way to invoke this "transcoding" -perhaps alter file name extension? (Pardon me if I'm wrong, as I haven't explored Slimserver much yet.)

Alex Twisleton-Wykeham-Fiennes
2006-01-15, 13:39
<snip>

> > There is a school of thought that says that the fact that the inversion
> > of a
> > some steady state non-symmetric signals being audibly different (which
> >
> > appears to be the most common proof of absolute phase being important)
> > is
> > proof that loudspeakers are not quite as linear and symetrical as they
> > should
> > be so maybe with a "perfect system" it would be a different story? (I
> > jest
> > slightly)
>
> Sure. However, we only have loudspeakers, and if they sound better with
> correct polarity, then that is what we should use!

I would be interested to know whether or not the results of people's polarity
decisions based on listening match with other people's conclusions. I wonder
whether or not there would be a difference of opinion with different speaker
technologies (ie do electrostatic panel speakers behave the same way "upside
down" as conventional cones? my gut reaction is not - they won't necessarily
be correct but they will almost definitely be different).

<snip>

> > I would be amazed if any multi-microphone recording (even so-called
> > time-aligned multi-mic classical recordings) came anywhere close enough
> > to be
> > phase coherent in themselves so any improvement would be due to chance
> >
> > psycho-accoustic affects rather than any "absolute correctness".
>
> I'm not so sure. If each channel of a multi-mic recording is in correct
> phase, then the result is simply a superpositioning of these channels.
> Sure, individual distances (time and volume), and low level ambient
> information shared between the channels would be a bit of a mish-mash,
> but the fact would remain that the leading edge of transients from all
> channels (from the instruments dominating each channel) would be in the
> correct direction.

I've spent quite a lot of time in the past being paid by people to mic up and
amplify / record concerts and while you are correct that the phase
information for each microphone would indeed be correct, the absolute
butchering of the soundstage and any background reverberation when using
multiple microphones. Of course, this is quite often preferrable to the
problems of working with a stereo pair in terms of balance and gain
(especially in a live situation) so often it is the only way to go, but in
terms of discussing a recording being really "correct", I think that this
only really applies to single point stereo pair recordings.

> Alex Twisleton-Wykeham-Fiennes Wrote:
> > > It is true that you system has to be pretty good to hear the
> > > difference. I had difficulty hearing it before i got the TacT RCS. I
> > > imagine headphones would do the job at a lower cost, however.
> >
> > Just out of interest - are you switching your phase before you go into
> > the
> > TacT or after (ie on the room corrected signal)? Do you have an option
> > on
> > this? I would be very interested to know if there is a difference on
> > your
> > system. I could imagine that the room specific processing that the
> > TacT
> > performs on the digital signal would be much more sensitive to absolute
> > phase
> > than the signal from the source that is being fed into the processor.
> >
> > Alex
>
> That is an interesting point. However, if you regard the TacT pre, the
> poweramp, and the speakers as a "black box" (let's say, a digitally
> corrected active loudspeaker), then this question would be pointless.
> If the corrected system gives much better performance with a input
> signal of the correct polarity, then that is what we should use!
>
> I do understand your point though. Is the digital room correction doing
> something that exagerates polarity differences, more than a perfect
> phase coherent speaker in an ideal listening room would have done? I
> don't know. I do know that people with very good systems, but no
> digital room correction, report the same type of audible differences
> between listening to correct or incorrect polarity. (I should say that
> "correct" here means, best sounding! I have yet to verify that the best
> sound corresponds to any recording's correct absolute polarity.
> Something that would be an interesting project! Correlating phase-data
> for various CDs with other audiophiles might be a good start.)
>
> Also, I do not know if the polarity inversion of the TacT is done
> before or after the digital room correction processing.

The key point here is that if the polarity is post TacT processing then adding
the functionality to the Squeezebox will probably not have the same effect.
However, it should be a quite easy effect to implement (although a little bit
of thought would need to be performed as to whether or not it is a global
switch for the output stage or a per-track decoding option, especially with
cross-feeds) and it would be an interesting thing to play around with.

Alex

Robert57
2006-11-18, 12:13
I thought I would revisit this polarity (or phase) toggle feature request, now that the relevant Bug 1976 (Polarity Control for SB2) is coded "Fixed" and "Resolved." Richard Titmuss noted on 9/18/06 that this feature is included in Slimserver v. 6.5, but is only usable for the Transporter. I have not seen any further discussion and was hoping this is still something the Slim team, or a savvy developer, can work on.

I'd like to ask again, can we get this polarity feature for the SB2 and SB3? For those of us with a phase inverting preamp, or with a mod that inverts phase, this would be very helpful feature. I'd be delighted just to have a simple toggle for general polarity, and leave for another day the song -by-song (or album-specific) polarity control. Since Firmware 15 reversed the phase, perhaps this option is best added in a new firmware version.

Do we need to initiate another feature request, or are there insurmountable problems?

Thanks.

Rob

MrSinatra
2006-11-18, 13:26
p floding,

i am curious, can u tell just by casually listening if a song needs polarity switched?

would u be able to tell in a double blind test situation?

it sounds to me like this is a very hard to detect issue, and one only a handful of people could make use of.

it seems to me though, that someone else had it right earlier, why not invert the file itself? the process would be reversable.

also, is there no program that can scan and detect it?

meaning, surely there are tell tale signs... wouldn't a program that detects this state be a better way to tell if its in this state? and perhaps an online DB for comparison. then flip the individual file. why process it each and every time u play it?

do u know of any popular recording that some of us may have that is reversed? i'd like to try to tell myself.

DSK
2007-01-11, 02:34
I thought I would revisit this polarity (or phase) toggle feature request, now that the relevant Bug 1976 (Polarity Control for SB2) is coded "Fixed" and "Resolved." Richard Titmuss noted on 9/18/06 that this feature is included in Slimserver v. 6.5, but is only usable for the Transporter. I have not seen any further discussion and was hoping this is still something the Slim team, or a savvy developer, can work on.

I'd like to ask again, can we get this polarity feature for the SB2 and SB3? For those of us with a phase inverting preamp, or with a mod that inverts phase, this would be very helpful feature. I'd be delighted just to have a simple toggle for general polarity, and leave for another day the song -by-song (or album-specific) polarity control. Since Firmware 15 reversed the phase, perhaps this option is best added in a new firmware version.

Do we need to initiate another feature request, or are there insurmountable problems?

Thanks.

Rob

Rob, I agree completely. Many audiophiles have their SB's modified so that the analog signal goes out directly from the DAC. This inverts polarity. As a result, many of us are still sitting on 6.2.1 with V15 firmware (that also inverted polarity) and remedied the problem. However, this means that we don't get all the bug fixes. We can't just remedy the problem by swapping the speaker wires at the terminals as this inverts polarity for ALL sources in the system. Also, the vast majority of pre-amps do not have a polarity inversion switch.

So a simple software switch "Invert polarity - Yes/No" via SlimServer interface would at least get us back to having all sources in correct polarity. The ability to switch polarity on the fly via the remote control would be nice but could wait until later.

DSK
2007-01-11, 02:39
After the above post, I went to Bug 1976 (mentioned at bottom of page 2 of this thread) to vote for it .... but it was already closed.

The bug was opened with specific reference to SB2, so I don't understand why a fix was introduced for the Transporter only and the bug was closed?

peter
2007-01-11, 02:52
DSK wrote:
> After the above post, I went to Bug 1976 (mentioned at bottom of page 2
> of this thread) to vote for it .... but it was already closed.
>
> The bug was opened with specific reference to SB2, so I don't
> understand why a fix was introduced for the Transporter only and the
> bug was closed?
>

What's the problem with the polarity? As I understand it polarity
doesn't matter as long as all the speakers agree on it. Or do
audiophiles disagree with this?

Regards,
Peter

chevvies
2009-08-01, 15:52
Was there a fix for this?

Phil Leigh
2009-08-01, 23:35
Was there a fix for this?

What do you think is/was broken?

The firmware is (correctly) non polarity-inverting on all platforms and has been for ages.

If you mean "has the feature request to implement remote polarity switching or automated switching based on a database tag been implemented?" - then, no.

Given that detection of absolute polarity is only possible by humans in certain well defined circumstances ref AES:Greiner + Melton (e.g. certain solo acoustic instruments) AND that very little attention is paid to this when manufacturing media or indeed audio gear in general except in studios where channel polarity is a critical issue...(and also a creative tool)

...I wouldn't hold my breath!

I can remotely invert polarity on my TACT. I have never used this feature since I do not possess any music which sounds different with +/- absolute polarity...

chevvies
2009-08-02, 01:27
Apologies for the vagueness, maybe I am misunderstanding the thread, it was the only one I could find that seemed to be addressing my problem. My issue is that when connecting my SB3 to my DAC via co-axial digital the stereo channels are reversed, therefore the soundstage is the wrong way around, and I don't know how to correct them.

Could it be an issue with the DAC? Thanks.

Phil Leigh
2009-08-02, 02:51
Apologies for the vagueness, maybe I am misunderstanding the thread, it was the only one I could find that seemed to be addressing my problem. My issue is that when connecting my SB3 to my DAC via co-axial digital the stereo channels are reversed, therefore the soundstage is the wrong way around, and I don't know how to correct them.

Could it be an issue with the DAC? Thanks.

Oh - sorry I thought you were talking about absolute polarity :-)
L/R inversion... is your DAC seperate from your amp? - if so just swap the DAC-AMP cables over...
If not, then swap your speaker cables over at the amp end.

What version of SC are you using?

chevvies
2009-08-02, 04:53
Yes, the amp and DAC are separate, I will try and reverse the digital co-ax cable, at work at the moment, I think SC should be the latest version, I'll check that when I get back too, thanks.

Changing the speaker end would make everything else wrong, so hopefully your first will work.

Phil Leigh
2009-08-02, 04:56
Yes, the amp and DAC are separate, I will try and reverse the digital co-ax cable, at work at the moment, I think SC should be the latest version, I'll check that when I get back too, thanks.

Changing the speaker end would make everything else wrong, so hopefully your first will work.

No - not the digital coax!!! - leave that alone. The two cables between your DAC and AMP need to be swapped over!

chevvies
2009-08-02, 04:58
I got it, sorry, thanks.:)

chevvies
2009-08-02, 12:57
Thanks, I double checked that cables were the right way around and they were so I presume it's the DAC at fault after all! :)

Phil Leigh
2009-08-03, 07:22
Thanks, I double checked that cables were the right way around and they were so I presume it's the DAC at fault after all! :)

So... did you swap the cables over?

chevvies
2009-08-03, 08:51
Yes thanks, and now the soundstage is correct but the red positive phono is now of course in the black socket.