PDA

View Full Version : Comparing Flac - CD Transport



James
2004-05-27, 12:27
Hello Everyone,

I am curious to get opinions from serious audiophiles about the quality of
steaming Flac - Wav files to their DA converter compared to using a CD
transport. I created a sample library using EAC and Flac, then stream over
the wireless network. I use the coax out to my DA.

Now, without throwing big prices and audiophile words around, I can say I
have a very nice setup.

About testing. I can easily go back and forth from Squeezebox and the CD
transport, so it is very easy to hear any differences there are. The
streaming Flac - Wav is very very good, in fact, if you haven't heard any
better you would think it was outstanding. But, when you listen via the CD
transport you hear 'more'. Again, without getting into fancy audiophile
language, the instruments have better detail. For example, with Flac, a
bongo in the background is noticeable and distinguishable. With the
transport, the bongo is closer to you, more detailed and pronounced. Overall
a better presentation. I don't have special ears either, you would hear it
too I am sure.

So, I would like to hear from those who have heard the gap in performance
and what can be done to close it. Perhaps there are better ways to create
the Flac file from CD in the first place? Also, I wonder if anyone has
created a test Flac file? This way there would be a way to hear the best it
can be and compare to the original CD.

thanks,




---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.689 / Virus Database: 450 - Release Date: 5/21/2004

Ron Thigpen
2004-05-27, 12:59
James wrote:

> So, I would like to hear from those who have heard the gap in performance
> and what can be done to close it. Perhaps there are better ways to create
> the Flac file from CD in the first place?

The FLAC file contains all of the information that was on the original
CD. The TCP/IP network makes certain that the SB is getting all of
those bits to work with. Trying to make improvements at this end of the
pipeline isn't likely to pay dividends as it's thought to be pretty
nearly perfect already.

The next two steps in the pipeline are the S/PDIF transport of the
bitstream and the digital to analog transformation.

S/PDIF seems to do the job and hasn't gotten all that much attention.
Many feel that good digital cables do yield sonic benefits.

The D/A process is a bit more of an art and has received a lot of
attention. There are some listeners who greatly prefer their external
D/A converters (DACs) to the one built into the SB. A good next step
for you might be to audition some external DACs in your system. Same
bitstream, but potentially a much improved analog signal and sound over
what you are now hearing.

--rt

Brooks Davis
2004-05-27, 13:03
On Thu, May 27, 2004 at 12:27:58PM -0700, James wrote:
> Hello Everyone,
>
> I am curious to get opinions from serious audiophiles about the quality of
> steaming Flac - Wav files to their DA converter compared to using a CD
> transport. I created a sample library using EAC and Flac, then stream over
> the wireless network. I use the coax out to my DA.
>
> Now, without throwing big prices and audiophile words around, I can say I
> have a very nice setup.
>
> About testing. I can easily go back and forth from Squeezebox and the CD
> transport, so it is very easy to hear any differences there are. The
> streaming Flac - Wav is very very good, in fact, if you haven't heard any
> better you would think it was outstanding. But, when you listen via the CD
> transport you hear 'more'. Again, without getting into fancy audiophile
> language, the instruments have better detail. For example, with Flac, a
> bongo in the background is noticeable and distinguishable. With the
> transport, the bongo is closer to you, more detailed and pronounced. Overall
> a better presentation. I don't have special ears either, you would hear it
> too I am sure.
>
> So, I would like to hear from those who have heard the gap in performance
> and what can be done to close it. Perhaps there are better ways to create
> the Flac file from CD in the first place? Also, I wonder if anyone has
> created a test Flac file? This way there would be a way to hear the best it
> can be and compare to the original CD.

Since Flac is lossless, there shouldn't be any issues there unless some
sort of extra processing was incorrectly enabled. I can think of a few
possaibilities.

- Audio extraction (ripping) wasn't clean so your source wasn't what
was on the CD. This is hard to check though using another ripper
might help unless the problem is with your CD drive.
- Bogus processing while encoding. This is fairly easy to check by
encoding and decoding the file and seeing if the streams are
identical.
- The CD transport is doing some processing before output. I've heard
there are some really nice up sampling processes out there so this is
a possiability. HDCDs might be an issue as well, but I don't have a
firm grasp of exactly how they work. This is verifiable with good
digial capture equipment.
- The digial outputs are imperfect so the DAC is making mistakes.
My audiophile uncle has reported major improvements after installing
some sort of signal conditioner between his transport and his DAC.
I'm not sure if he is using one today, but he used to use an Audio
Alchemny one a while back. I think what the device does is use a
very good input stage to insure it has a correct view of the inputs
and then outputs a very precicly clocked, low noise stream. There's
some discussion of the things that can go wrong when reading a
digital stream on RME's website: http://www.rme-audio.com/english/.

Once I've got enough of my library in Flac format, I'm hoping to take my
squeezebox down to try it out on my uncles system. That will be a
serious listening test.

-- Brooks

--
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFAtkmZXY6L6fI4GtQRAvSBAKDatAlNabir5a09b8LYBR v8J2li7QCfXY1+
pqXV+YRZceLfAz5ocYbWH+E=
=jY9R
-----END PGP SIGNATURE-----

Pat Farrell
2004-05-27, 13:26
At 03:59 PM 5/27/2004, Ron Thigpen wrote:
>James wrote:
>>So, I would like to hear from those who have heard the gap in performance
>>and what can be done to close it. Perhaps there are better ways to create
>>the Flac file from CD in the first place?
>
>The FLAC file contains all of the information that was on the original CD.

The extraction process/program can sometimes make a difference.
EAC claims to reread the data until it gets the same data.
I just use CDex, because it is lots faster, but YMMV.

RedBook audio is a pretty lame format, or is so now in hindsight.
It was close to state of the art in the late 70s when it was designed.
A more modern digital format would have more redundancy
and error correction built it, but hey, we have to live with what we have.

Flac doesn't change the bits at all, so if you extracted it right, it is right.

> The TCP/IP network makes certain that the SB is getting all of those
> bits to work with. Trying to make improvements at this end of the
> pipeline isn't likely to pay dividends as it's thought to be pretty
> nearly perfect already.

More specifically, TCP/IP guarantees bit-exact delivery (or complete failure).

>S/PDIF seems to do the job and hasn't gotten all that much attention. Many
>feel that good digital cables do yield sonic benefits.

S/PDIF is fine for any reasonable length transport. There are no such
things as digital cables, except in the dreams of marketing departments.
All cables are analog. Digital signals either are transmitted or not. Binary
and all that. :-)


>The D/A process is a bit more of an art and has received a lot of
>attention. There are some listeners who greatly prefer their external D/A
>converters (DACs) to the one built into the SB. A good next step for you
>might be to audition some external DACs in your system. Same bitstream,
>but potentially a much improved analog signal and sound over what you are
>now hearing.

DAs are where it is at, assuming that the rest of the analog chain
is the same, and properly gain staged.

The human ear is highly sensitive to tiny changes in relative loudness (gain)
and almost always thinks that the louder signal is better.
It is very hard to get the levels matched, and if they are not within 0.1 dB,
you simply can't make judgements over fine distinctions.

So if your CD transport has a little more gain in its chain, it will sound
better.
And if your amp and speakers and room acoustics are not first rate, you
won't be able to properly tell.

Getting the room acoustics right while keeping the decorating committee
happy is a real challenge.

Pat

Mark Bennett
2004-05-27, 13:42
I've done some testing using a Benchmark DAC1 external DAC on
a high-end system (Naim Pre/Power Amps, Kef Reference speakers)
and comparing the use of a Squeezebox and an Audiolab 8000
transport (was ~$2k just for a CD transport...).

The details can be seen on an earlier posting:
http://article.gmane.org/gmane.music.equipment.slimdevices.general/8640/match=external+dac

The bottom line is that there is no perceivable difference
between the squeezebox and a normal CD transport even on a
$20k+ system.

Just one thought comes to mind on this. Are you sure you're
streaming uncompressed? In recent versions of SlimServer
Slim "helpfully" introduced a feature where all music was
re-encoded to 320kb MP3 for a wireless SB. I'm not quite sure
when this was introduced, but I think it was in 5.1.5. Try
one of the more recent nightlies and make sure the bitrate
limit is set to "no limit" in the player settings.

We could clearly tell the difference between 320kb MP3 and
uncompressed audio. This might explain why you're hearing a
difference.

Others have suggested alternative causes, which are also
worth checking if it isn't being re-encoded to MP3.

James wrote:
> Hello Everyone,
>
> I am curious to get opinions from serious audiophiles about the quality of
> steaming Flac - Wav files to their DA converter compared to using a CD
> transport. I created a sample library using EAC and Flac, then stream over
> the wireless network. I use the coax out to my DA.
>
> Now, without throwing big prices and audiophile words around, I can say I
> have a very nice setup.
>
> About testing. I can easily go back and forth from Squeezebox and the CD
> transport, so it is very easy to hear any differences there are. The
> streaming Flac - Wav is very very good, in fact, if you haven't heard any
> better you would think it was outstanding. But, when you listen via the CD
> transport you hear 'more'. Again, without getting into fancy audiophile
> language, the instruments have better detail. For example, with Flac, a
> bongo in the background is noticeable and distinguishable. With the
> transport, the bongo is closer to you, more detailed and pronounced. Overall
> a better presentation. I don't have special ears either, you would hear it
> too I am sure.
>
> So, I would like to hear from those who have heard the gap in performance
> and what can be done to close it. Perhaps there are better ways to create
> the Flac file from CD in the first place? Also, I wonder if anyone has
> created a test Flac file? This way there would be a way to hear the best it
> can be and compare to the original CD.
>
> thanks,
>
>
>
>
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.689 / Virus Database: 450 - Release Date: 5/21/2004
>
>

Michel Fombellida
2004-05-27, 13:50
Hi,

Using a good ripping tool and a lossless compression, you should get "bit
exact" music on your PC. What could be the difference then? Answer: jitter and
possible processing (upsampling/oversampling) that a transport could handle. My
current solution, pass the digital output of the squeezebox through an dCS
purcell and an Apogee Big Ben before reaching the DAC. In this case the
Squeezebox output is really first class. Perhaps a next step would be to modify
the Squeezebox to add a word clock input to sync all the systems together. I
know a squeezebox/Big Ben user who also has a heavily tweeked Wadia transport
and he can hardly tell any difference.

Michel

Ron Thigpen
2004-05-27, 13:51
Pat Farrell wrote:

>> The FLAC file contains all of the information that was on the original
>> CD.
>
> Flac doesn't change the bits at all, so if you extracted it right, it is
> right.

yep.

>> The TCP/IP network makes certain that the SB is getting all of those
>> bits to work with. Trying to make improvements at this end of the
>> pipeline isn't likely to pay dividends as it's thought to be pretty
>> nearly perfect already.

> More specifically, TCP/IP guarantees bit-exact delivery (or complete
> failure).

yep.

>> S/PDIF seems to do the job and hasn't gotten all that much attention.
>> Many feel that good digital cables do yield sonic benefits.

> S/PDIF is fine for any reasonable length transport. There are no such
> things as digital cables, except in the dreams of marketing departments.
> All cables are analog. Digital signals either are transmitted or not.
> Binary and all that. :-)

more precisely, digital signals are modulated into an analog signal,
which is transmitted, and demodulated back into a (hopefully) faithful
representation of the original signal.

a cable design that will transmit a modulated digital signal such that
the receiving equipment is able to reproduce the original signal with
fidelity may be somewhat different in characteristics than a design
intended to carry line-level analog signals faithfully. cables modeled
on such designs are generally referred to as "digital cables".

without getting into the whole "you can hear it and i can't" routine,
there are many who report that well designed "digital cables" when used
in their systems, sound better, to their ears. please note the
qualifiers in the above statement.

yes, there are those who will gladly accept large sums of your
presumably hard earned money in exchange for their versions of these
"digital cables". as in all other things: "caveat emptor". let your
ears be your guide, when auditioning cables in your system, before
opening your wallet.

>> The D/A process is a bit more of an art and has received a lot of
>> attention. There are some listeners who greatly prefer their external
>> D/A converters (DACs) to the one built into the SB. A good next step
>> for you might be to audition some external DACs in your system. Same
>> bitstream, but potentially a much improved analog signal and sound
>> over what you are now hearing.
>
> DAs are where it is at, assuming that the rest of the analog chain
> is the same, and properly gain staged.

yep. a good next step.

--rt

James
2004-05-27, 14:35
Thanks,

I have the Slimserver set up to only stream Flac - Wav. The bitrate for the
player set for no limit.

I use an external DAC with same digital coax cable to either the Squeezebox
or transport to test.

From what you and others say it is possible to close the performance gap
between Flac and using a transport. That would mean that I have something
wrong in the creation of the Flac. I use EAC, then Flac on the fly. Is this
the best way?








----- Original Message -----
From: "Mark Bennett" <mark (AT) markandliz (DOT) co.uk>
To: "Slim Devices Discussion" <discuss (AT) lists (DOT) slimdevices.com>
Sent: Thursday, May 27, 2004 1:42 PM
Subject: [slim] Comparing Flac - CD Transport


>
> I've done some testing using a Benchmark DAC1 external DAC on
> a high-end system (Naim Pre/Power Amps, Kef Reference speakers)
> and comparing the use of a Squeezebox and an Audiolab 8000
> transport (was ~$2k just for a CD transport...).
>
> The details can be seen on an earlier posting:
>
http://article.gmane.org/gmane.music.equipment.slimdevices.general/8640/match=external+dac
>
> The bottom line is that there is no perceivable difference
> between the squeezebox and a normal CD transport even on a
> $20k+ system.
>
> Just one thought comes to mind on this. Are you sure you're
> streaming uncompressed? In recent versions of SlimServer
> Slim "helpfully" introduced a feature where all music was
> re-encoded to 320kb MP3 for a wireless SB. I'm not quite sure
> when this was introduced, but I think it was in 5.1.5. Try
> one of the more recent nightlies and make sure the bitrate
> limit is set to "no limit" in the player settings.
>
> We could clearly tell the difference between 320kb MP3 and
> uncompressed audio. This might explain why you're hearing a
> difference.
>
> Others have suggested alternative causes, which are also
> worth checking if it isn't being re-encoded to MP3.
>
> James wrote:
> > Hello Everyone,
> >
> > I am curious to get opinions from serious audiophiles about the quality
of
> > steaming Flac - Wav files to their DA converter compared to using a CD
> > transport. I created a sample library using EAC and Flac, then stream
over
> > the wireless network. I use the coax out to my DA.
> >
> > Now, without throwing big prices and audiophile words around, I can say
I
> > have a very nice setup.
> >
> > About testing. I can easily go back and forth from Squeezebox and the CD
> > transport, so it is very easy to hear any differences there are. The
> > streaming Flac - Wav is very very good, in fact, if you haven't heard
any
> > better you would think it was outstanding. But, when you listen via the
CD
> > transport you hear 'more'. Again, without getting into fancy audiophile
> > language, the instruments have better detail. For example, with Flac, a
> > bongo in the background is noticeable and distinguishable. With the
> > transport, the bongo is closer to you, more detailed and pronounced.
Overall
> > a better presentation. I don't have special ears either, you would hear
it
> > too I am sure.
> >
> > So, I would like to hear from those who have heard the gap in
performance
> > and what can be done to close it. Perhaps there are better ways to
create
> > the Flac file from CD in the first place? Also, I wonder if anyone has
> > created a test Flac file? This way there would be a way to hear the best
it
> > can be and compare to the original CD.
> >
> > thanks,
> >
> >
> >
> >
> > ---
> > Outgoing mail is certified Virus Free.
> > Checked by AVG anti-virus system (http://www.grisoft.com).
> > Version: 6.0.689 / Virus Database: 450 - Release Date: 5/21/2004
> >
> >

Pat Farrell
2004-05-27, 14:45
At 05:35 PM 5/27/2004, James wrote:
> From what you and others say it is possible to close the performance gap
>between Flac and using a transport. That would mean that I have something
>wrong in the creation of the Flac.

No, I would not assume that the difference is in the creation of the flac
file.

> I use EAC, then Flac on the fly. Is this the best way?

To be more specific, the normal way is to use EAC or CDex to
extract, then flac to compress to a .flac file for storage.
Then the SlimServer decompresses the flac file to wav on the fly.
This is normal, flac is explicitly designed to be eazy and fast to
decompress.

If you are doing this, you are not using the CD player's transport.
You may be using its DAC. It isn't clear to me from your
messages.

If you are taking the output of the SqueezeBox's digital output
and putting it to one of two DACs (one external and one inside a CD player)
then you still have to make sure you have the gain/volumes
exactly matched.

Or you may have an audiophile CD player with a better DAC than your
external one.

If the SqueezeBox has any impact on the sound of flac files to an external
DAC, then something is wrong with it. It is just a delivery system.

Pat

Brooks Davis
2004-05-27, 15:05
On Thu, May 27, 2004 at 05:45:41PM -0400, Pat Farrell wrote:
> If you are taking the output of the SqueezeBox's digital output
> and putting it to one of two DACs (one external and one inside a CD player)
> then you still have to make sure you have the gain/volumes
> exactly matched.

One thing just occurred to me. Is the Player Settings -> Digital Volume
Control value set to "Digital output level is fixed"? If not, then the
output is probably being run through a filter which will almost
certainly hurt your quality.

-- Brooks

--
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFAtmY5XY6L6fI4GtQRAnk2AJ4r3QXRMFsYMDlucNTJFR Pfd5JTYQCgkGoI
Plv7WccpuXB4rj4Q2DBBtcc=
=k8Ea
-----END PGP SIGNATURE-----

James
2004-05-27, 15:12
I will be more clear.

I connect the Squeezebox to my DAC using digital coax. To compare, I replaced the Squeezebox with my transport. In this test, the transport wins.

There are those who have replied and stated that the performance gap can be closed between Squeezebox and Transport. If this is correct, then I have to take a close look at the Flac files I created and see if improvements can be made.

I use EAC, insert a command line to use Flac automatically when copy|compress.

So, it appears the only variable is the creation of the Flac file.











----- Original Message -----
From: Pat Farrell
To: Slim Devices Discussion
Sent: Thursday, May 27, 2004 2:45 PM
Subject: [slim] Comparing Flac - CD Transport


At 05:35 PM 5/27/2004, James wrote:

From what you and others say it is possible to close the performance ga
between Flac and using a transport. That would mean that I have something
wrong in the creation of the Flac.

No, I would not assume that the difference is in the creation of the flac file.

> I use EAC, then Flac on the fly. Is this the best way?

To be more specific, the normal way is to use EAC or CDex to
extract, then flac to compress to a .flac file for storage.
Then the SlimServer decompresses the flac file to wav on the fly.
This is normal, flac is explicitly designed to be eazy and fast to
decompress.

If you are doing this, you are not using the CD player's transport.
You may be using its DAC. It isn't clear to me from your
messages.

If you are taking the output of the SqueezeBox's digital output
and putting it to one of two DACs (one external and one inside a CD player)
then you still have to make sure you have the gain/volumes
exactly matched.

Or you may have an audiophile CD player with a better DAC than your
external one.

If the SqueezeBox has any impact on the sound of flac files to an external
DAC, then something is wrong with it. It is just a delivery system.

Pat




------------------------------------------------------------------------------

dean
2004-05-27, 16:03
Is it possible that your "transport" is post-processing the data coming
off the CD before sending out the digital coax output?

If your CD rip was accurate (it likely was) and your digital output is
error free (it likely is), then the difference is probably in some
processing that your "transport" is doing to the data. Squeezebox,
assuming that you have disabled the digital volume control, should be
passing the bits through exactly as they came off the CD.

-dean


On May 27, 2004, at 3:12 PM, James wrote:

> I will be more clear.
>
> I connect the Squeezebox to my DAC using digital coax. To compare, I
> replaced the Squeezebox with my transport. In this test, the transport
> wins.
>
> There are those who have replied and stated that the performance gap
> can be closed between Squeezebox and Transport. If this is correct,
> then I have to take a close look at the Flac files I created and see
> if improvements can be made.
>
> I use EAC, insert a command line to use Flac automatically when
> copy|compress.
>
> So, it appears the only variable is the creation of the Flac file.
>
>
>
>
>
>
>
>
>
>
>
> ----- Original Message -----
> From: Pat Farrell
> To: Slim Devices Discussion
> Sent: Thursday, May 27, 2004 2:45 PM
> Subject: [slim] Comparing Flac - CD Transport
>
> At 05:35 PM 5/27/2004, James wrote:
>
> From what you and others say it is possible to close the performance ga
> between Flac and using a transport. That would mean that I have
> something
> wrong in the creation of the Flac.
>
> No, I would not assume that the difference is in the creation of the
> flac file.
>
> > I use EAC, then Flac on the fly. Is this the best way?
>
> To be more specific, the normal way is to use EAC or CDex to
> extract, then flac to compress to a .flac file for storage.
> Then the SlimServer decompresses the flac file to wav on the fly.
> This is normal, flac is explicitly designed to be eazy and fast to
> decompress.
>
> If you are doing this, you are not using the CD player's transport.
> You may be using its DAC. It isn't clear to me from your
> messages.
>
> If you are taking the output of the SqueezeBox's digital output
> and putting it to one of two DACs (one external and one inside a CD
> player)
> then you still have to make sure you have the gain/volumes
> exactly matched.
>
> Or you may have an audiophile CD player with a better DAC than your
> external one.
>
> If the SqueezeBox has any impact on the sound of flac files to an
> external
> DAC, then something is wrong with it. It is just a delivery system.
>
> Pat
>
>
>
>

Eric Brandon
2004-05-27, 16:22
Michel Fombellida <mf22433@...> writes:

>
> Hi,
>
> Using a good ripping tool and a lossless compression, you should get "bit
> exact" music on your PC. What could be the difference then? Answer: jitter
and
> possible processing (upsampling/oversampling) that a transport could handle.
My
> current solution, pass the digital output of the squeezebox through an dCS
> purcell and an Apogee Big Ben before reaching the DAC. In this case the
> Squeezebox output is really first class. Perhaps a next step would be to
modify
> the Squeezebox to add a word clock input to sync all the systems together. I
> know a squeezebox/Big Ben user who also has a heavily tweeked Wadia
transport
> and he can hardly tell any difference.
>
> Michel
>


I also send the Squeezebox digital coax output to a dCS Purcell upconverter
(which upconverts to 176khz/24 bits, re-clocks the data, and reduces jitter)
and then to a dCS Elgar DAC. This sounds great. Nevertheless, I'll run some
A/B tests of the Squeezebox vs. a Wadia CD transport later today and report to
the forum.

Some ideas for your issues:

1) I have tried to send uncompressed audio to the squeezebox via 802.11b,
802.11g with a "g" bridge, and wired ethernet.
- 802.11b didn't work in my house -- couldn't get enough bandwidth to
support uncompressed audio
- 802.11g bridge worked just fine, but (in my opinion) didn't sound nearly
as good as the wired ethernet does. I have no theory as to why -- just the
experience to report.

2) I archive my music as WAV files, and stream them without any on-the-fly
conversion. One idea for you would be rip some CDs without converting to FLAC
and then compare the WAV files to the CD transport. This would allow you to
take FLAC out of the equation in trying to figure out the differences you're
hearing.

- Eric Brandon

James
2004-05-27, 16:27
Thanks Dean!!

Did not know about the digital volume control. Found it and disabled it and
that made a big improvement.

Now, if I could find a way to get rid of the dropouts when playing Flac over
a wireless network....




----- Original Message -----
From: "dean blackketter" <dean (AT) slimdevices (DOT) com>
To: "Slim Devices Discussion" <discuss (AT) lists (DOT) slimdevices.com>
Sent: Thursday, May 27, 2004 4:03 PM
Subject: [slim] Comparing Flac - CD Transport


Is it possible that your "transport" is post-processing the data coming
off the CD before sending out the digital coax output?

If your CD rip was accurate (it likely was) and your digital output is
error free (it likely is), then the difference is probably in some
processing that your "transport" is doing to the data. Squeezebox,
assuming that you have disabled the digital volume control, should be
passing the bits through exactly as they came off the CD.

-dean


On May 27, 2004, at 3:12 PM, James wrote:

> I will be more clear.
>
> I connect the Squeezebox to my DAC using digital coax. To compare, I
> replaced the Squeezebox with my transport. In this test, the transport
> wins.
>
> There are those who have replied and stated that the performance gap
> can be closed between Squeezebox and Transport. If this is correct,
> then I have to take a close look at the Flac files I created and see
> if improvements can be made.
>
> I use EAC, insert a command line to use Flac automatically when
> copy|compress.
>
> So, it appears the only variable is the creation of the Flac file.
>
>
>
>
>
>
>
>
>
>
>
> ----- Original Message -----
> From: Pat Farrell
> To: Slim Devices Discussion
> Sent: Thursday, May 27, 2004 2:45 PM
> Subject: [slim] Comparing Flac - CD Transport
>
> At 05:35 PM 5/27/2004, James wrote:
>
> From what you and others say it is possible to close the performance ga
> between Flac and using a transport. That would mean that I have
> something
> wrong in the creation of the Flac.
>
> No, I would not assume that the difference is in the creation of the
> flac file.
>
> > I use EAC, then Flac on the fly. Is this the best way?
>
> To be more specific, the normal way is to use EAC or CDex to
> extract, then flac to compress to a .flac file for storage.
> Then the SlimServer decompresses the flac file to wav on the fly.
> This is normal, flac is explicitly designed to be eazy and fast to
> decompress.
>
> If you are doing this, you are not using the CD player's transport.
> You may be using its DAC. It isn't clear to me from your
> messages.
>
> If you are taking the output of the SqueezeBox's digital output
> and putting it to one of two DACs (one external and one inside a CD
> player)
> then you still have to make sure you have the gain/volumes
> exactly matched.
>
> Or you may have an audiophile CD player with a better DAC than your
> external one.
>
> If the SqueezeBox has any impact on the sound of flac files to an
> external
> DAC, then something is wrong with it. It is just a delivery system.
>
> Pat
>
>
>
>

Pat Farrell
2004-05-27, 16:41
At 07:27 PM 5/27/2004, James wrote:
>Thanks Dean!!
>Did not know about the digital volume control. Found it and disabled it and
>that made a big improvement.

Sure, gain differences are huge.

>Now, if I could find a way to get rid of the dropouts when playing Flac over
>a wireless network....

Don't use wireless if you are serious about sound quality.
I get occasional dropouts even wired directly to a 100baseT switch.

A future version of the SqueezeBox firmware is promised to
support native FLAC over the wire (or wireless) which will
greatly help the dropout problem.

In the meantime, WAV or FLAC really wants wired connections.

Pat

James
2004-05-27, 17:22
Thanks Pat,

I will look into a dedicated wire just for the one hifi system.

I also looked into getting the best from MP3 by installing Lame and changing the conf file to -q 1 -b 320 . This yields surprisingly good results for MP3 with no dropouts.






----- Original Message -----
From: Pat Farrell
To: Slim Devices Discussion
Sent: Thursday, May 27, 2004 4:41 PM
Subject: [slim] Comparing Flac - CD Transport


At 07:27 PM 5/27/2004, James wrote:

Thanks Dean!!
Did not know about the digital volume control. Found it and disabled it and
that made a big improvement.

Sure, gain differences are huge.


Now, if I could find a way to get rid of the dropouts when playing Flac over
a wireless network....

Don't use wireless if you are serious about sound quality.
I get occasional dropouts even wired directly to a 100baseT switch.

A future version of the SqueezeBox firmware is promised to
support native FLAC over the wire (or wireless) which will
greatly help the dropout problem.

In the meantime, WAV or FLAC really wants wired connections.

Pat



------------------------------------------------------------------------------

Eric Brandon
2004-05-27, 17:26
Eric Brandon <ebran2@...> writes:

<< I'll run some A/B tests of the Squeezebox vs. a Wadia CD transport later
today and report to the forum >>

OK -- I ripped Mark Knoffler's "Sailing to Philadelphia", as WAV files, using
iTunes with "error correction" ON.

The two signal paths were
A) Slimserver -> "Buffalo" Router -> 100' ethernet cable -> Squeezebox -> Tara
labs digital coax cable -> dCS Purcell upconverting to 176.4khz/24 bits ->
taralabs dual AES/EBU digital cables -> dCS elgar plus DAC -> etc.

B) Wadia 860 CD player/transport -> 3' tara labs AES/EBU digital cable -> dCS
purcell upconverting ... (the same as above)

You should know that I am the pickiest audionut you'll ever meet who has spent
hundreds of hours tweaking his digital cables, analog cables, AC power cables,
room reflections, etc.

I can hear NO difference between the squeezebox and the CD transport.

It is entirely possible that the dCS Purcell is eliminating differences by
reclocking and upconverting the signals, and that if you were to plug both
sources directly into a DAC, differences would be heard. If so -- the next
step for someone trying to eliminate differences might be to put some kind of
digital signal conditioner between his Squeezebox and his DAC.

- Eric

James
2004-05-27, 17:42
Thanks Eric for taking the time to test. I tried ripping a CD to WAV using
EAC to do a similar test, but the slimserver does not show the files when it
rescans.

Eric - did you ever to a similar test with Flac?




----- Original Message -----
From: "Eric Brandon" <ebran2 (AT) pacbell (DOT) net>
To: <discuss (AT) lists (DOT) slimdevices.com>
Sent: Thursday, May 27, 2004 5:26 PM
Subject: [slim] Comparing Flac - CD Transport


> Eric Brandon <ebran2@...> writes:
>
> << I'll run some A/B tests of the Squeezebox vs. a Wadia CD transport
later
> today and report to the forum >>
>
> OK -- I ripped Mark Knoffler's "Sailing to Philadelphia", as WAV files,
using
> iTunes with "error correction" ON.
>
> The two signal paths were
> A) Slimserver -> "Buffalo" Router -> 100' ethernet cable -> Squeezebox ->
Tara
> labs digital coax cable -> dCS Purcell upconverting to 176.4khz/24 bits ->
> taralabs dual AES/EBU digital cables -> dCS elgar plus DAC -> etc.
>
> B) Wadia 860 CD player/transport -> 3' tara labs AES/EBU digital cable ->
dCS
> purcell upconverting ... (the same as above)
>
> You should know that I am the pickiest audionut you'll ever meet who has
spent
> hundreds of hours tweaking his digital cables, analog cables, AC power
cables,
> room reflections, etc.
>
> I can hear NO difference between the squeezebox and the CD transport.
>
> It is entirely possible that the dCS Purcell is eliminating differences by
> reclocking and upconverting the signals, and that if you were to plug both
> sources directly into a DAC, differences would be heard. If so -- the
next
> step for someone trying to eliminate differences might be to put some kind
of
> digital signal conditioner between his Squeezebox and his DAC.
>
> - Eric
>
>
>

Pat Farrell
2004-05-27, 17:52
At 08:22 PM 5/27/2004, James wrote:
>I will look into a dedicated wire just for the one hifi system.

It works well if you can drag the cable. Of course, if you bought the
Sqbx for wireless, that isn't too good a solution.

>I also looked into getting the best from MP3 by installing Lame and
>changing the conf file to -q 1 -b 320 . This yields surprisingly good
>results for MP3 with no dropouts.

Since the SqueezeBox does native MP3, you can
get very good quality (not audiophile, but very good) with
VBR and 320 or higher. And no dropouts.

I can't tell any clear difference between -q 0 or -q 1 with lame.
But I have admit that I haven't played that much with quality MP3, since
Flac does what I need without the need for subjective analysis

I am a firm believer in DAC qualities. At least between a
$80 DAC and a $800 one. I don't know if there are
"wow what a difference" distinctions between a $1000 DAC
and a $1200 one. Of course, the Benchmark DAC-1 used to
be a $850 unit, and now is $975.

I can't figure out what the audiophile transports
are doing. There isn't supposed to be
stuff like wow or flutter with RedBook. If they did something like
include an EAC style re-read on error, with a big buffer for
streaming, that would make engineering sense, but then
it would have a delay between when you push "play" and
when sounds appear. Which is probably bad user experience.
And while Dean's suggestion to check the level helped your case,
I don't expect that they really include gain functions,
or EQ or that kind of signal processing. Most "purist" audiophile
stuff doesn't have any "extra" stuff. So I don't know
what they do over a PC with EAC.


Pat

Eric Brandon
2004-05-27, 19:48
James <idata@...> writes:

>
> Thanks Eric for taking the time to test. I tried ripping a CD to WAV using
> EAC to do a similar test, but the slimserver does not show the files when it
> rescans.
>
> Eric - did you ever to a similar test with Flac?

I have not done a similar test with FLAC.

If you can't see the files, it probably because WAV files have no tags.
In "additional server settings/interface" turn on something like "browse
folders" or "browse files". Then you'll be able to use that option on the
Squeezebox to navigate to & play the WAV files even though they have no tags.

- Eric

Michel Fombellida
2004-05-28, 00:37
Hi Eric,

I did compare FLAC to WAV to Apple Lossless and I cannot detect any difference
either (except that at the moment Apple Lossless is a bit less stable and
sometimes mov123.exe is crashing).

I am running wireless but will test wired soon.

Michel

Ho Yin Ng
2004-05-28, 02:13
All this talk of Flac and DACs is making me very interested.

I also want to enquire whether the type of CD-rom you use in your PC to
read the audio disk could have some effect?

Ho Yin

At 00:41 28/05/2004, you wrote:
>At 07:27 PM 5/27/2004, James wrote:
>>Thanks Dean!!
>>Did not know about the digital volume control. Found it and disabled it and
>>that made a big improvement.
>
>Sure, gain differences are huge.
>
>>Now, if I could find a way to get rid of the dropouts when playing Flac over
>>a wireless network....
>
>Don't use wireless if you are serious about sound quality.
>I get occasional dropouts even wired directly to a 100baseT switch.
>
>A future version of the SqueezeBox firmware is promised to
>support native FLAC over the wire (or wireless) which will
>greatly help the dropout problem.
>
>In the meantime, WAV or FLAC really wants wired connections.
>
>Pat
>

Mike Kozlowski
2004-05-28, 06:36
On Thu, 27 May 2004, Pat Farrell wrote:

> Don't use wireless if you are serious about sound quality.
> I get occasional dropouts even wired directly to a 100baseT switch.

The Squeezebox is a 10Mbps device. 802.11g is (up to) 54Mbps. If you
connect wirelessly via a g bridge, it's entirely possible that the wired
Ethernet connection between the bridge and the Squeezebox is the
bottleneck.

--
Mike Kozlowski
http://www.klio.org/mlk/

Ron Thigpen
2004-05-28, 07:46
Eric Brandon wrote:
> OK -- I ripped Mark Knoffler's "Sailing to Philadelphia", as WAV files, using
> iTunes with "error correction" ON.
>[...]
> It is entirely possible that the dCS Purcell is eliminating differences by
> reclocking and upconverting the signals, and that if you were to plug both
> sources directly into a DAC, differences would be heard.

Eric,

Would you be willing to do a comparison without the dCS Purcell in the
chain and report your impressions?

--rt

Pat Farrell
2004-05-28, 20:37
At 05:13 AM 5/28/2004, Ho Yin Ng wrote:
>All this talk of Flac and DACs is making me very interested.
>I also want to enquire whether the type of CD-rom you use in your PC to
>read the audio disk could have some effect?

If an extractor package, such as EAC can read the audio on the CD,
then it reads it "exactly" as it was encoded. So proper extraction will
get you the exact bits from the audio CD. The quality of the PC CD-reader
makes no difference in the quality of the result.

However, not all computer CD-ROM readers (or writers) are equal.
The quality of the CD-ROM reader will have direct impact on the speed
that EAC can extract the audio. You will not get "50x" extraction with
complete accuracy. And with some disks, and some readers, you may
never get accurate reading of the bits.

So the quality of the CD-ROM reader will have impact on how
quickly you can do the processing.

Once it is extracted to a wave (.wav) file, then flac will compress it
without loss, since that is what flac does. I see compression
rations ranging from 25% (nearly four to one) to as bad as 70%
(only a 30% saving) depending on type of music, etc.

The EAC website has some recent experiences with brands of CDs.
http://www.exactaudiocopy.org/
Their "tips & specs" link has specific models that they have tested.

A few years ago, my company extracted audio from 40,000 CDs. At
that time, the TDK Velo CDs were the fastest and most repeatedly
accurate drives, but all the drives are better and faster since then.

Pat

Mark Bennett
2004-05-29, 01:01
I would be very interested in hearing the answer to this one as well.

Cheers,
Mark.

Ron Thigpen wrote:
> Eric Brandon wrote:
>
>> OK -- I ripped Mark Knoffler's "Sailing to Philadelphia", as WAV
>> files, using iTunes with "error correction" ON.
>
> >[...]
>
>> It is entirely possible that the dCS Purcell is eliminating
>> differences by reclocking and upconverting the signals, and that if
>> you were to plug both sources directly into a DAC, differences would
>> be heard.
>
>
> Eric,
>
> Would you be willing to do a comparison without the dCS Purcell in the
> chain and report your impressions?
>
> --rt
>

Eric Brandon
2004-05-29, 18:37
Mark Bennett <mark@...> writes:

> > Eric,
> >
> > Would you be willing to do a comparison without the dCS Purcell in the
> > chain and report your impressions?
> >

A good deal of cable switching required, and no time this weekend. But will
try it out early next week & report to the group.

Cheers,
- Eric