PDA

View Full Version : Squeezebox/External DAC - technical issues



Mark Bennett
2004-03-17, 15:45
Tonight we tried again to connect the Squeezebox up to an
external DAC with a high end sound system behind it.

In this post I'm going to focus on the ongoing technical
issues. I want to make these completely separate from any
subjective discussion about audio quality, otherwise it
will become unproductive. I'll discuss audio quality in a
separate post (see "Squeezebox/External DAC - audio quality"
thread).

At the end of the session we really have three technical
problems which would prevent/hinder the use of this
combination of equipment.

First of all, let me add some context. The only DAC we
have access to is a high-end DAC. The sort of high end where
the designers don't trust complex digital circuits....
The PLL for this DAC is a completely discrete design. It has
a very slow acquisition time, and extremely high stability.
Along with this comes an extreme intolerance for absolute
frequency error. The DAC does work perfectly well with a CD
transport (AudioLab CDM8000 MK2 - 44.1KHz) and a DAB tuner
(48KHz I think).

So on to the issues:

1) We just could not get the DAC to lock onto the output
of the squeezebox. Admittedly we only tried Flac and Ogg,
with client side decoding. In hindsight we should also
have tried WAV direct from the server, but I didn't think
this would be different. (It might not be.)

However this doesn't look like a hardware problem. When
playing high bit-rate MP3's (lame encoded, 320kb/s) the
digital output locked quickly and reliably.

To me this sounds that the digital output frequency is
probably still not quite right for uncompressed audio
streams.

2) When the Squeezebox is actually stopped there is no signal
output on the digital out. This means that the PLL loses
lock and we lost the first few seconds of music when
initially playing music from stop while the PLL locked.

By contrast the Audiolab CD transport output a silence code
all the time it was turned on, whether it was in stop or
pause. (We didn't think to try pause on the SB - should
have done in hindsight.)

This issue should affect anyone using an external DAC, is
anyone else seeing this?

3) The digital output also seemed to stop when skipping forward
or backward to next/previous tracks. This showed by the DAC
briefly flashing that it had lost input signal during the
transition.

Fortunately since the frequency didn't have time to drift
we didn't lose any music, so this is an annoyance rather than
a real problem (especially compared to 1) & 2).

Again the CD transport's behaviour was immaculate. Again, this
should probably affect most DAC's unless this one has an
untypically quick loss of input detector. Is anyone else
seeing this?

Stuart Hickinbottom
2004-03-17, 17:36
Thanks for the evaulation, very interesting. I'd be interested in your
opinion update when you have heard FLACs though it.

I have my SB connected optically to an external DAC (actually within a
not-very-new Technics digital integrated amplifier).

I can play FLACs through this arrangement - the DAC locks on successfully
at 44.1KHz.

When my SB is paused or stopped I get no digital output and the DAC loses
lock. This DAC seems to lock on very quickly, however, so I don't seem to
lose anything when I restart playing.

I also get it losing sync juring next/prev track, and also periodically
(but not continually) when fastforwarding/reversing though a track.

I don't have it losing sync between tracks when playing normally from one
track to the next.

It would be quite nice for me too if it played silence when there was no
real signal to send as my DAC has some kind of an internal relay that I
can hear clicking on and off when it's gaining and losing sync. Apart from
that, though, I think it's working well for me under normal playing
conditions.

Stuart

On Wed, 17 Mar 2004 22:45:51 +0000, Mark Bennett <mark (AT) markandliz (DOT) co.uk>
wrote:

>
> Tonight we tried again to connect the Squeezebox up to an
> external DAC with a high end sound system behind it.
>
> In this post I'm going to focus on the ongoing technical
> issues. I want to make these completely separate from any
> subjective discussion about audio quality, otherwise it
> will become unproductive. I'll discuss audio quality in a
> separate post (see "Squeezebox/External DAC - audio quality"
> thread).
>
> At the end of the session we really have three technical
> problems which would prevent/hinder the use of this
> combination of equipment.
>
> First of all, let me add some context. The only DAC we
> have access to is a high-end DAC. The sort of high end where
> the designers don't trust complex digital circuits....
> The PLL for this DAC is a completely discrete design. It has
> a very slow acquisition time, and extremely high stability.
> Along with this comes an extreme intolerance for absolute
> frequency error. The DAC does work perfectly well with a CD
> transport (AudioLab CDM8000 MK2 - 44.1KHz) and a DAB tuner
> (48KHz I think).
>
> So on to the issues:
>
> 1) We just could not get the DAC to lock onto the output
> of the squeezebox. Admittedly we only tried Flac and Ogg,
> with client side decoding. In hindsight we should also
> have tried WAV direct from the server, but I didn't think
> this would be different. (It might not be.)
>
> However this doesn't look like a hardware problem. When
> playing high bit-rate MP3's (lame encoded, 320kb/s) the
> digital output locked quickly and reliably.
>
> To me this sounds that the digital output frequency is
> probably still not quite right for uncompressed audio
> streams. 2) When the Squeezebox is actually stopped there is no signal
> output on the digital out. This means that the PLL loses
> lock and we lost the first few seconds of music when
> initially playing music from stop while the PLL locked.
>
> By contrast the Audiolab CD transport output a silence code
> all the time it was turned on, whether it was in stop or
> pause. (We didn't think to try pause on the SB - should
> have done in hindsight.)
>
> This issue should affect anyone using an external DAC, is
> anyone else seeing this?
>
> 3) The digital output also seemed to stop when skipping forward
> or backward to next/previous tracks. This showed by the DAC
> briefly flashing that it had lost input signal during the
> transition.
>
> Fortunately since the frequency didn't have time to drift
> we didn't lose any music, so this is an annoyance rather than
> a real problem (especially compared to 1) & 2).
>
> Again the CD transport's behaviour was immaculate. Again, this
> should probably affect most DAC's unless this one has an
> untypically quick loss of input detector. Is anyone else
> seeing this?
>
>
>
>
>

dean
2004-03-17, 18:09
The lack of active S/PDIF at the beginnings and endings of tracks (and
while fast forwarding) is known and will be addressed soon. Thanks for
your patience...

-dean

On Mar 17, 2004, at 4:36 PM, Stuart Hickinbottom wrote:

> Thanks for the evaulation, very interesting. I'd be interested in your
> opinion update when you have heard FLACs though it.
>
> I have my SB connected optically to an external DAC (actually within a
> not-very-new Technics digital integrated amplifier).
>
> I can play FLACs through this arrangement - the DAC locks on
> successfully at 44.1KHz.
>
> When my SB is paused or stopped I get no digital output and the DAC
> loses lock. This DAC seems to lock on very quickly, however, so I
> don't seem to lose anything when I restart playing.
>
> I also get it losing sync juring next/prev track, and also
> periodically (but not continually) when fastforwarding/reversing
> though a track.
>
> I don't have it losing sync between tracks when playing normally from
> one track to the next.
>
> It would be quite nice for me too if it played silence when there was
> no real signal to send as my DAC has some kind of an internal relay
> that I can hear clicking on and off when it's gaining and losing sync.
> Apart from that, though, I think it's working well for me under normal
> playing conditions.
>
> Stuart
>
> On Wed, 17 Mar 2004 22:45:51 +0000, Mark Bennett
> <mark (AT) markandliz (DOT) co.uk> wrote:
>
>>
>> Tonight we tried again to connect the Squeezebox up to an
>> external DAC with a high end sound system behind it.
>>
>> In this post I'm going to focus on the ongoing technical
>> issues. I want to make these completely separate from any
>> subjective discussion about audio quality, otherwise it
>> will become unproductive. I'll discuss audio quality in a
>> separate post (see "Squeezebox/External DAC - audio quality"
>> thread).
>>
>> At the end of the session we really have three technical
>> problems which would prevent/hinder the use of this
>> combination of equipment.
>>
>> First of all, let me add some context. The only DAC we
>> have access to is a high-end DAC. The sort of high end where
>> the designers don't trust complex digital circuits....
>> The PLL for this DAC is a completely discrete design. It has
>> a very slow acquisition time, and extremely high stability.
>> Along with this comes an extreme intolerance for absolute
>> frequency error. The DAC does work perfectly well with a CD
>> transport (AudioLab CDM8000 MK2 - 44.1KHz) and a DAB tuner
>> (48KHz I think).
>>
>> So on to the issues:
>>
>> 1) We just could not get the DAC to lock onto the output
>> of the squeezebox. Admittedly we only tried Flac and Ogg,
>> with client side decoding. In hindsight we should also
>> have tried WAV direct from the server, but I didn't think
>> this would be different. (It might not be.)
>>
>> However this doesn't look like a hardware problem. When
>> playing high bit-rate MP3's (lame encoded, 320kb/s) the
>> digital output locked quickly and reliably.
>>
>> To me this sounds that the digital output frequency is
>> probably still not quite right for uncompressed audio
>> streams. 2) When the Squeezebox is actually stopped there is no
>> signal
>> output on the digital out. This means that the PLL loses
>> lock and we lost the first few seconds of music when
>> initially playing music from stop while the PLL locked.
>>
>> By contrast the Audiolab CD transport output a silence code
>> all the time it was turned on, whether it was in stop or
>> pause. (We didn't think to try pause on the SB - should
>> have done in hindsight.)
>>
>> This issue should affect anyone using an external DAC, is
>> anyone else seeing this?
>>
>> 3) The digital output also seemed to stop when skipping forward
>> or backward to next/previous tracks. This showed by the DAC
>> briefly flashing that it had lost input signal during the
>> transition.
>>
>> Fortunately since the frequency didn't have time to drift
>> we didn't lose any music, so this is an annoyance rather than
>> a real problem (especially compared to 1) & 2).
>>
>> Again the CD transport's behaviour was immaculate. Again, this
>> should probably affect most DAC's unless this one has an
>> untypically quick loss of input detector. Is anyone else
>> seeing this?
>>
>>
>>
>>
>>

Pat Farrell
2004-03-17, 20:49
At 05:45 PM 3/17/2004, Mark Bennett wrote:
>Tonight we tried again to connect the Squeezebox up to an
>external DAC with a high end sound system behind it.
>1) We just could not get the DAC to lock onto the output
> of the squeezebox. Admittedly we only tried Flac and Ogg,
> with client side decoding.

What cabling did you use? toslink? SPDIF?
All my music is FLAC, works mostly great.

>When the Squeezebox is actually stopped there is no signal
> output on the digital out. This means that the PLL loses...
> This issue should affect anyone using an external DAC, is
> anyone else seeing this?

Not on my setup, Benchmark DAC-1


> should probably affect most DAC's unless this one has an
> untypically quick loss of input detector. Is anyone else
> seeing this?

I found the coax SPdif connection to be flakey. But there were threads
on it being fixed in a near future release so I have not been
scientific about it.

I just use the TOSlink optical connection.
I get occasional dropouts, but in general the sound quality
is wonderful, and the dropouts are not frequent enough to drive me nuts,
altho they are noticable and I'm tempted to hold my breath until
the new firmware /server ware fixes it.

I would not be happy with the coax SPdif, but Toslink is at least OK.

Pat

Mark Bennett
2004-03-18, 00:06
So, apart from the FLAC experience, it sounds like you're
seeing the same problems as us. I should have mentioned that
we also do not see any problems when transitioning from
the end of one track to the start of another.

I suspect that the FLAC locking is due to very fussy DAC, so
I expect that it will work for most people. We're trying to
find the specs for the DAC, but it is now obsolete, and
my friend isn't sure where it's instruction manual is...

Stuart Hickinbottom wrote:
> Thanks for the evaulation, very interesting. I'd be interested in your
> opinion update when you have heard FLACs though it.
>
> I have my SB connected optically to an external DAC (actually within a
> not-very-new Technics digital integrated amplifier).
>
> I can play FLACs through this arrangement - the DAC locks on
> successfully at 44.1KHz.
>
> When my SB is paused or stopped I get no digital output and the DAC
> loses lock. This DAC seems to lock on very quickly, however, so I don't
> seem to lose anything when I restart playing.
>
> I also get it losing sync juring next/prev track, and also periodically
> (but not continually) when fastforwarding/reversing though a track.
>
> I don't have it losing sync between tracks when playing normally from
> one track to the next.
>
> It would be quite nice for me too if it played silence when there was
> no real signal to send as my DAC has some kind of an internal relay
> that I can hear clicking on and off when it's gaining and losing sync.
> Apart from that, though, I think it's working well for me under normal
> playing conditions.
>
> Stuart
>

Mark Bennett
2004-03-18, 00:11
Thanks for the info on this. It's good to know these are recognised
problems that are being worked on. I can afford to be patient if I
know something will happen - I don't think the CFO (wife) will
sanction the expense on an external DAC for a few months!

My friend will hold off purchasing SB until the issues are resolved.

dean blackketter wrote:

> The lack of active S/PDIF at the beginnings and endings of tracks (and
> while fast forwarding) is known and will be addressed soon. Thanks for
> your patience...
>
> -dean

Mark Bennett
2004-03-18, 00:14
Pat Farrell wrote:

> At 05:45 PM 3/17/2004, Mark Bennett wrote:
>
>> Tonight we tried again to connect the Squeezebox up to an
>> external DAC with a high end sound system behind it.
>> 1) We just could not get the DAC to lock onto the output
>> of the squeezebox. Admittedly we only tried Flac and Ogg,
>> with client side decoding.
>
>
> What cabling did you use? toslink? SPDIF?
> All my music is FLAC, works mostly great.

We only tried the Toslink. I guess we should have tried the
S/PDIF, but only had limited time to get things sorted. Since
MP3 was working it seemed pretty unlikely that the Flac/toslink
combo was a problem.

>> When the Squeezebox is actually stopped there is no signal
>> output on the digital out. This means that the PLL loses...
>> This issue should affect anyone using an external DAC, is
>> anyone else seeing this?
>
>
> Not on my setup, Benchmark DAC-1

This is interesting, especially since I have my eye on the
same DAC at ome point in the future:-)

Mark Bennett
2004-03-18, 16:27
One extra piece of information on this. I found out today that
this DAC only has an input frequency error tolerance of 100ppm.
For a normal 44.1kHz signal this basically means 4Hz out in
either direction.

Obviously this is very tight, so the clock setting in the chip
might only need to be one count away from ideal to upset this
DAC. Unfortunately I don't think we have any lab equipment at
work which would allow me to measure to this degree of
accuracy, so I don't know how I can help further.

I guess the answer is to get access to a less fussy DAC, but
this isn't going to happen for months, so this would be the
end of the route as regards further testing.....

Mark Bennett wrote:
> So, apart from the FLAC experience, it sounds like you're
> seeing the same problems as us. I should have mentioned that
> we also do not see any problems when transitioning from
> the end of one track to the start of another.
>
> I suspect that the FLAC locking is due to very fussy DAC, so
> I expect that it will work for most people. We're trying to
> find the specs for the DAC, but it is now obsolete, and
> my friend isn't sure where it's instruction manual is...
>
> Stuart Hickinbottom wrote:
>
>> Thanks for the evaulation, very interesting. I'd be interested in
>> your opinion update when you have heard FLACs though it.
>>
>> I have my SB connected optically to an external DAC (actually within
>> a not-very-new Technics digital integrated amplifier).
>>
>> I can play FLACs through this arrangement - the DAC locks on
>> successfully at 44.1KHz.
>>
>> When my SB is paused or stopped I get no digital output and the DAC
>> loses lock. This DAC seems to lock on very quickly, however, so I
>> don't seem to lose anything when I restart playing.
>>
>> I also get it losing sync juring next/prev track, and also
>> periodically (but not continually) when fastforwarding/reversing
>> though a track.
>>
>> I don't have it losing sync between tracks when playing normally from
>> one track to the next.
>>
>> It would be quite nice for me too if it played silence when there was
>> no real signal to send as my DAC has some kind of an internal relay
>> that I can hear clicking on and off when it's gaining and losing
>> sync. Apart from that, though, I think it's working well for me under
>> normal playing conditions.
>>
>> Stuart
>>
>
>
>

T
2004-03-19, 01:47
> One extra piece of information on this. I found out today that
> this DAC only has an input frequency error tolerance of 100ppm.
> For a normal 44.1kHz signal this basically means 4Hz out in
> either direction.
>
> Obviously this is very tight,

Actually, thats very loose. Too loose for professional use, anyway.

Tom

Mark Bennett
2004-03-19, 02:21
Tom,

I think you might have mis-understood me. I'm not talking about
jitter here, I'm talking about absolute frequency error. You can
have 0ppm jitter (theoretically) and still have 100pm (or more)
absolute frequency error.

Having a tolerance of just +/-4Hz on whether the DAC will see the
input signal or not is tiny. I bet that most professional DAC's
are much more tolerant that this. One well respected professional
DAC - the Benchmark DAC1 (keeps cropping up doesn't it?) seems
to be able to lock to *any* sample rate between 28KHz and 198KHz.
This is a massive tolerance to frequency error in terms of
ability to lock.

Of course, 100ppm of jitter would not be acceptable to professional
users, nor most audiophile domestic users, but this isn't what
I'm talking about.

100ppm of frequency error is equivalent to a 5 minute track being,
at most 30ms longer (or shorter) than it should be. I would be
staggered if anyone really through they could hear this make an
audible difference to played music.

Of course, 100ppm of jitter would not be acceptable to professional
users, nor most audiophile domestic users, but this isn't what
I'm talking about.

Cheers,
Mark.

T wrote:
>>One extra piece of information on this. I found out today that
>>this DAC only has an input frequency error tolerance of 100ppm.
>>For a normal 44.1kHz signal this basically means 4Hz out in
>>either direction.
>>
>>Obviously this is very tight,
>
>
> Actually, thats very loose. Too loose for professional use, anyway.
>
> Tom
>
>

T
2004-03-19, 06:21
> I think you might have mis-understood me. I'm not talking about
> jitter here, I'm talking about absolute frequency error. You can
> have 0ppm jitter (theoretically) and still have 100pm (or more)
> absolute frequency error.

No, it didn't misunderstand.

"The AES have recommended use of D.A.R.S. and have defined a "Grade 1"
system as one having a long term frequency accuracy of better than +/-1 part
per million. "Grade 2" D.A.R.S. (e.g. the Amek DMS) have a long term
frequency accuracy of better than +/-10 parts per million. Grade 2
references are considered entirely adequate for most practical studio
installations. "

As I said, 100 ppm is a lot considering that 10ppm is considered "adequate
for most practical studio installations".

However, I agree that any good DAC should tolerate worse. But, there is no
valid reason for even a consumer grade piece of audio source equipment to
have a frequency error of 100ppm or more.

Tom

seanadams
2004-03-19, 08:19
I measured 44100.00 KHz for several squeezeboxen using a frequency
counter... that's the best precision I can measure with this counter,
and it's within your tolerance requirement.


On Mar 18, 2004, at 3:27 PM, Mark Bennett wrote:

> One extra piece of information on this. I found out today that
> this DAC only has an input frequency error tolerance of 100ppm.
> For a normal 44.1kHz signal this basically means 4Hz out in
> either direction.
>
> Obviously this is very tight, so the clock setting in the chip
> might only need to be one count away from ideal to upset this
> DAC. Unfortunately I don't think we have any lab equipment at
> work which would allow me to measure to this degree of
> accuracy, so I don't know how I can help further.
>
> I guess the answer is to get access to a less fussy DAC, but
> this isn't going to happen for months, so this would be the
> end of the route as regards further testing.....
>
> Mark Bennett wrote:
>> So, apart from the FLAC experience, it sounds like you're
>> seeing the same problems as us. I should have mentioned that
>> we also do not see any problems when transitioning from
>> the end of one track to the start of another.
>> I suspect that the FLAC locking is due to very fussy DAC, so
>> I expect that it will work for most people. We're trying to
>> find the specs for the DAC, but it is now obsolete, and
>> my friend isn't sure where it's instruction manual is...
>> Stuart Hickinbottom wrote:
>>> Thanks for the evaulation, very interesting. I'd be interested in
>>> your opinion update when you have heard FLACs though it.
>>>
>>> I have my SB connected optically to an external DAC (actually within
>>> a not-very-new Technics digital integrated amplifier).
>>>
>>> I can play FLACs through this arrangement - the DAC locks on
>>> successfully at 44.1KHz.
>>>
>>> When my SB is paused or stopped I get no digital output and the DAC
>>> loses lock. This DAC seems to lock on very quickly, however, so I
>>> don't seem to lose anything when I restart playing.
>>>
>>> I also get it losing sync juring next/prev track, and also
>>> periodically (but not continually) when fastforwarding/reversing
>>> though a track.
>>>
>>> I don't have it losing sync between tracks when playing normally
>>> from one track to the next.
>>>
>>> It would be quite nice for me too if it played silence when there
>>> was no real signal to send as my DAC has some kind of an internal
>>> relay that I can hear clicking on and off when it's gaining and
>>> losing sync. Apart from that, though, I think it's working well for
>>> me under normal playing conditions.
>>>
>>> Stuart
>>>
>>

Ron Thigpen
2004-03-19, 09:12
Sean Adams wrote:

> I measured 44100.00 KHz for several squeezeboxen using a frequency
> counter... that's the best precision I can measure with this counter,
> and it's within your tolerance requirement.

i believe you intended to type "44100.00 Hz".

--rt

seanadams
2004-03-19, 09:15
Heh.. yes. Unless I accidentally hit the turbo switch.

On Mar 19, 2004, at 8:12 AM, Ron Thigpen wrote:

> Sean Adams wrote:
>
>> I measured 44100.00 KHz for several squeezeboxen using a frequency
>> counter... that's the best precision I can measure with this counter,
>> and it's within your tolerance requirement.
>
> i believe you intended to type "44100.00 Hz".
>
> --rt
>

Mark Bennett
2004-03-19, 14:14
OK, so I believe that it's not frequency error....

What else could be causing a failure to work with an
external DAC when the analogue outputs are working
correctly?

Also any ideas on why MP3 works perfectly, but any raw
data stream doesn't?

Is there anything else that we should be trying to
identify the cause of problem?

Sean Adams wrote:
>
> I measured 44100.00 KHz for several squeezeboxen using a frequency
> counter... that's the best precision I can measure with this counter,
> and it's within your tolerance requirement.
>
>
> On Mar 18, 2004, at 3:27 PM, Mark Bennett wrote:
>
>> One extra piece of information on this. I found out today that
>> this DAC only has an input frequency error tolerance of 100ppm.
>> For a normal 44.1kHz signal this basically means 4Hz out in
>> either direction.
>>
>> Obviously this is very tight, so the clock setting in the chip
>> might only need to be one count away from ideal to upset this
>> DAC. Unfortunately I don't think we have any lab equipment at
>> work which would allow me to measure to this degree of
>> accuracy, so I don't know how I can help further.
>>
>> I guess the answer is to get access to a less fussy DAC, but
>> this isn't going to happen for months, so this would be the
>> end of the route as regards further testing.....
>>
>> Mark Bennett wrote:
>>
>>> So, apart from the FLAC experience, it sounds like you're
>>> seeing the same problems as us. I should have mentioned that
>>> we also do not see any problems when transitioning from
>>> the end of one track to the start of another.
>>> I suspect that the FLAC locking is due to very fussy DAC, so
>>> I expect that it will work for most people. We're trying to
>>> find the specs for the DAC, but it is now obsolete, and
>>> my friend isn't sure where it's instruction manual is...
>>> Stuart Hickinbottom wrote:
>>>
>>>> Thanks for the evaulation, very interesting. I'd be interested in
>>>> your opinion update when you have heard FLACs though it.
>>>>
>>>> I have my SB connected optically to an external DAC (actually within
>>>> a not-very-new Technics digital integrated amplifier).
>>>>
>>>> I can play FLACs through this arrangement - the DAC locks on
>>>> successfully at 44.1KHz.
>>>>
>>>> When my SB is paused or stopped I get no digital output and the DAC
>>>> loses lock. This DAC seems to lock on very quickly, however, so I
>>>> don't seem to lose anything when I restart playing.
>>>>
>>>> I also get it losing sync juring next/prev track, and also
>>>> periodically (but not continually) when fastforwarding/reversing
>>>> though a track.
>>>>
>>>> I don't have it losing sync between tracks when playing normally
>>>> from one track to the next.
>>>>
>>>> It would be quite nice for me too if it played silence when there
>>>> was no real signal to send as my DAC has some kind of an internal
>>>> relay that I can hear clicking on and off when it's gaining and
>>>> losing sync. Apart from that, though, I think it's working well for
>>>> me under normal playing conditions.
>>>>
>>>> Stuart
>>>>
>>>

seanadams
2004-03-19, 15:14
Yeah - I'm a little perplexed on that one myself. If you weren't
across the pond I'd offer to drop by with our analyzer and take a
look....

Has anyone else seen this problem with this DAC?



On Mar 19, 2004, at 1:14 PM, Mark Bennett wrote:

> OK, so I believe that it's not frequency error....
>
> What else could be causing a failure to work with an
> external DAC when the analogue outputs are working
> correctly?
>
> Also any ideas on why MP3 works perfectly, but any raw
> data stream doesn't?
>
> Is there anything else that we should be trying to
> identify the cause of problem?
>
> Sean Adams wrote:
>> I measured 44100.00 KHz for several squeezeboxen using a frequency
>> counter... that's the best precision I can measure with this counter,
>> and it's within your tolerance requirement.
>> On Mar 18, 2004, at 3:27 PM, Mark Bennett wrote:
>>> One extra piece of information on this. I found out today that
>>> this DAC only has an input frequency error tolerance of 100ppm.
>>> For a normal 44.1kHz signal this basically means 4Hz out in
>>> either direction.
>>>
>>> Obviously this is very tight, so the clock setting in the chip
>>> might only need to be one count away from ideal to upset this
>>> DAC. Unfortunately I don't think we have any lab equipment at
>>> work which would allow me to measure to this degree of
>>> accuracy, so I don't know how I can help further.
>>>
>>> I guess the answer is to get access to a less fussy DAC, but
>>> this isn't going to happen for months, so this would be the
>>> end of the route as regards further testing.....
>>>
>>> Mark Bennett wrote:
>>>
>>>> So, apart from the FLAC experience, it sounds like you're
>>>> seeing the same problems as us. I should have mentioned that
>>>> we also do not see any problems when transitioning from
>>>> the end of one track to the start of another.
>>>> I suspect that the FLAC locking is due to very fussy DAC, so
>>>> I expect that it will work for most people. We're trying to
>>>> find the specs for the DAC, but it is now obsolete, and
>>>> my friend isn't sure where it's instruction manual is...
>>>> Stuart Hickinbottom wrote:
>>>>
>>>>> Thanks for the evaulation, very interesting. I'd be interested in
>>>>> your opinion update when you have heard FLACs though it.
>>>>>
>>>>> I have my SB connected optically to an external DAC (actually
>>>>> within a not-very-new Technics digital integrated amplifier).
>>>>>
>>>>> I can play FLACs through this arrangement - the DAC locks on
>>>>> successfully at 44.1KHz.
>>>>>
>>>>> When my SB is paused or stopped I get no digital output and the
>>>>> DAC loses lock. This DAC seems to lock on very quickly, however,
>>>>> so I don't seem to lose anything when I restart playing.
>>>>>
>>>>> I also get it losing sync juring next/prev track, and also
>>>>> periodically (but not continually) when fastforwarding/reversing
>>>>> though a track.
>>>>>
>>>>> I don't have it losing sync between tracks when playing normally
>>>>> from one track to the next.
>>>>>
>>>>> It would be quite nice for me too if it played silence when there
>>>>> was no real signal to send as my DAC has some kind of an internal
>>>>> relay that I can hear clicking on and off when it's gaining and
>>>>> losing sync. Apart from that, though, I think it's working well
>>>>> for me under normal playing conditions.
>>>>>
>>>>> Stuart
>>>>>
>>>>

Pat Farrell
2004-03-19, 15:27
At 05:14 PM 3/19/2004, Sean Adams <sadams (AT) slimdevices (DOT) com> wrote:
>Yeah - I'm a little perplexed on that one myself. If you weren't across
>the pond I'd offer to drop by with our analyzer and take a look....

Would probably be cheaper to FexEx or DHL the DAC across the pond,
they don't expect chocolate biscuits.

Pat Farrell
2004-03-19, 15:44
At 04:14 PM 3/19/2004, Mark Bennett wrote:
>What else could be causing a failure to work with an
>external DAC when the analogue outputs are working
>correctly?Also any ideas on why MP3 works perfectly, but any raw
>data stream doesn't?

MP3s are half the size of Flac/wave files, often one tenth the size.
Any constriction on the decoding, transport, delayering in the Sqbx,
etc. will have at least twice the impact.

It is easy to guess, and this is only a guess, that the Micronas MAS 3509F
has a buffer in its output so that it can easily handle short loss of
streaming.
And it would be easy to guess that the SPDIF and Toslink connections
do not have this buffering. They could assume that an External DAC would
have suitable buffering as needed in its design.

It you look a the RedBook spec, an audio player is required to handle
fairly long
dropouts and errors, and never, or nearly never go silent. A cheap CD player
will fake it, continue the prior signal if needed when it can't read the disk.
I expect that the Micronas MAS has similar (if not the same) circuits in it.

An external DAC, on the other hand, expects to be fed nice digital signals.
Error free
and all that. This is especially true of professional (production, not
audiophile) where
stoppage of signals has been a no-no since Edison wax cylinders.

There have been so many postings on this topic that I get confused.
Have you, Mark Bennett, tried all the suggested optimizations before the
signal hits the SqueezeBox? Specifically, is the music (flac files) on a
local disk
to the server that is executing SlimServer.pl (or .exe). Is the server well
tuned
without bad processes in the background or competing services?
Have you tried both wired and wireless connections to the squeezebox so
we can isolate which is the cause (or both if that is the case).
Have you changed the ServerSettings -> Maximum number of bytes of audio data t
to a larger value? did it make a difference?
Have you tried decompressing some FLAC files to wave and playing the wave files
directly, bypassing the decompression time?

There are probably more suggestions, these are all I can remember.

Thanks
Pat

Mark Bennett
2004-04-18, 03:56
I have now finally succumbed to the external DAC world and
am now a proud owner of a Benchmark Media DAC1.

This has allowed me to check these issues and experiences with
another DAC - some of the problems have gone away, but one
very minor issue seems to remain.

The new DAC is hooked up to the SB using a (cheap) optical
cable, the DAC is hooked into my Cyrus 7 Amp, and then into
some old Mission speakers. (I've also done a quick check on
co-ax connection between SB and DAC1, which seems to work,
but I haven't run it for more than a minute yet.)

(Original system was toslink connection to Audio Synthesis
DAX, then into Naim amp with huge KEF speakers.)

I've also been streaming this from the same laptop and setup
that I was using for the original trials to keep as much the
same as possible.

> So on to the issues:
>
> 1) We just could not get the DAC to lock onto the output
> of the squeezebox. Admittedly we only tried Flac and Ogg,
> with client side decoding. In hindsight we should also
> have tried WAV direct from the server, but I didn't think
> this would be different. (It might not be.)
>
> However this doesn't look like a hardware problem. When
> playing high bit-rate MP3's (lame encoded, 320kb/s) the
> digital output locked quickly and reliably.
>
> To me this sounds that the digital output frequency is
> probably still not quite right for uncompressed audio
> streams.

The good news is that all audio formats work fine with the
DAC1. I've tried WAV, MP3, FLAC & Ogg with no problems. (All
are 44.1kHz, high quality settings.)

So whatever issue we had with the other DAC is related to
some difference between the SB and Audiolab transport that
affects the original DAC, but not the DAC1......

I think that this rules out most of the suggested issues
related to network bandwidth and activity. If these were
problems I'd expect to see at least some indication of a
problem on the DAC1, and I don't.

Now I have a DAC that "works" I hope to be doing some testing
on my friends high-end system this week, so I'll get a
chance to try this again with server 5.1.3 just to see if
this makes a difference.

> 2) When the Squeezebox is actually stopped there is no signal
> output on the digital out. This means that the PLL loses
> lock and we lost the first few seconds of music when
> initially playing music from stop while the PLL locked.
>
> By contrast the Audiolab CD transport output a silence code
> all the time it was turned on, whether it was in stop or
> pause. (We didn't think to try pause on the SB - should
> have done in hindsight.)
>
> This issue should affect anyone using an external DAC, is
> anyone else seeing this?

This phenomena doesn't occur with the DAC1. As long as the SB
is connected to the server the DAC seems happy and doesn't
indicate an error. (Even when the SB is powered down.)

Again, I guess it must be some form of extreme sensitivity of
the original DAC, and we'll try again with server 5.1.3 when
we get the chance.

> 3) The digital output also seemed to stop when skipping forward
> or backward to next/previous tracks. This showed by the DAC
> briefly flashing that it had lost input signal during the
> transition.
>
> Fortunately since the frequency didn't have time to drift
> we didn't lose any music, so this is an annoyance rather than
> a real problem (especially compared to 1) & 2).
>
> Again the CD transport's behaviour was immaculate. Again, this
> should probably affect most DAC's unless this one has an
> untypically quick loss of input detector. Is anyone else
> seeing this?

This still seems to be a problem with server 5.1.1, especially
between tracks of different coding types, e.g. going from Flac
to Ogg, or WAV to MP3. I don't lose any sound, but the error light
on the DAC briefly lights up between tracks and there is a faint
click in the audio.

I've tried this with server 5.1.3, and it still seems to show an
error when switching tracks, but the click almost never appears
now, so there is some improvement.