PDA

View Full Version : Digtial output audio quality



Ken
2004-04-10, 08:09
I have an issue with the audio quality of the Squeezebox that has been nagging me for a bit. I use the digital output of the Squeezebox to feed an external DAC (conrad johnson), and set the digital output level setting at the server to 'fixed'. Until a month ago I was using a 5.1 nightly build (1/31 - it was very stable for me), and then finally upgraded to 5.1.1. After a few days of using the new version I had the nagging feeling that the sound quality was not quite as good as it was with the earlier release. I switched back to the older build and sure enough heard an improvement in quality. I've spent the past few weeks alternately listening to each version and find that there is indeed a small, but noticeable difference in the quality of the sound. Note that there are firmware changes between these two server versions, and I update the Squeezebox with each change.

Given that I'm simply feeding the digital source directly to the DAC, I can't account for a reason why a difference might exist. I'm guessing that it falls into one of 3 categories:

1. The timing in the digital output has somehow changed between the two firmware versions and it effects the way my DAC is able to decode the music.
2. Whatever processing is going on within the Squeezebox to transmit the bits out is somehow interfering with them to a degree.
3. There are no differences and I'm being fooled by my own expectations.

I'm curious as to whether anyone else has experienced this difference, and if anyone might be able to shed some light what might cause this change.

Thanks,
Ken

Mark Bennett
2004-04-10, 13:00
There were a number of improvements in the digital out behaviour between
5.1 and 5.1.1. All of these have so far been seen as improvements to the
compatibility of the box, with this being the first suggestion I've seen
that anything might have got worse....

The changes I can think were introduced in 5.1.1 are:

1) correction of sample frequency from 44.19kHz to 44.1kHz.
2) corrected the level of the "validity" bit in the s/pdif stream
3) correction of sonic glitches between tracks

There might be others, but I haven't gone trawling through the
checkins, and the release notes are silent on these changes.

I assume that you're using some form of lossless coding, otherwise
you probabaly wouldn't have been able to notice the difference.

I suspect that the change in sample frequency is probably the main
cause of whatever you're hearing. I suspect it just makes the
sound different from what you're used to, not necessarily worse.

Can you articulate the difference you think you're hearing? Just
"not quite as good" is difficult to discuss.


Ken wrote:
> I have an issue with the audio quality of the Squeezebox that has been
> nagging me for a bit. I use the digital output of the Squeezebox to feed
> an external DAC (conrad johnson), and set the digital output level
> setting at the server to 'fixed'. Until a month ago I was using a 5.1
> nightly build (1/31 - it was very stable for me), and then finally
> upgraded to 5.1.1. After a few days of using the new version I had the
> nagging feeling that the sound quality was not quite as good as it was
> with the earlier release. I switched back to the older build and sure
> enough heard an improvement in quality. I've spent the past few weeks
> alternately listening to each version and find that there is indeed a
> small, but noticeable difference in the quality of the sound. Note that
> there are firmware changes between these two server versions, and I
> update the Squeezebox with each change.
>
> Given that I'm simply feeding the digital source directly to the DAC, I
> can't account for a reason why a difference might exist. I'm guessing
> that it falls into one of 3 categories:
>
> 1. The timing in the digital output has somehow changed between the two
> firmware versions and it effects the way my DAC is able to decode the
> music.
> 2. Whatever processing is going on within the Squeezebox to transmit the
> bits out is somehow interfering with them to a degree.
> 3. There are no differences and I'm being fooled by my own expectations.
>
> I'm curious as to whether anyone else has experienced this difference,
> and if anyone might be able to shed some light what might cause this
> change.
>
> Thanks,
> Ken
>
>

Ken
2004-04-10, 18:09
Thanks for the response Mark. I'm glad to hear that things actually did change between the versions, and that I'm not simply imagining differences. You may be correct that the audible changes I'm hearing are simply different and not actually worse, and that the older version is my preferred sound. For reference, the changes I'm hearing between 5.1 and 5.1.1 fall into these categories:

1. Slightly slower attack time across the entire spectrum, with greatest prominence at the bottom end. However, there's an overall increase in midbass response with 5.1.1.

2. Less discrete location of instruments across the sound stage. With the earlier 5.1 version, the images are very distinct and solidly located. With 5.1.1 they tend to be more diffuse and are harder to pin down.

3. Slight rolloff in the higher frequencies. This may be a tonal balance issue more than an actual rolloff, but to my ears there is noticeably less energy at the very top end with the 5.1.1 version.

I understand that all of these changes can just be artifacts of the DAC I'm using, and may have nothing to do with the proper transmission of the bits. However, I did want to see what might have changed between the versions (thank you for filling me in on that), and possibly bring up the issue in case one or more of the changes made did cause a degradation in the sound.

Thanks,
Ken

BTW - You're correct in assuming that I'm using lossless (FLAC) encoding of my audio files.

Mark Bennett wrote:
> There were a number of improvements in the digital out behaviour between
> 5.1 and 5.1.1. All of these have so far been seen as improvements to the
> compatibility of the box, with this being the first suggestion I've seen
> that anything might have got worse....
>
> The changes I can think were introduced in 5.1.1 are:
>
> 1) correction of sample frequency from 44.19kHz to 44.1kHz.
> 2) corrected the level of the "validity" bit in the s/pdif stream
> 3) correction of sonic glitches between tracks
>
> There might be others, but I haven't gone trawling through the
> checkins, and the release notes are silent on these changes.
>
> I assume that you're using some form of lossless coding, otherwise
> you probabaly wouldn't have been able to notice the difference.
>
> I suspect that the change in sample frequency is probably the main
> cause of whatever you're hearing. I suspect it just makes the
> sound different from what you're used to, not necessarily worse.
>
> Can you articulate the difference you think you're hearing? Just
> "not quite as good" is difficult to discuss.
>
>
> Ken wrote:
>
>> I have an issue with the audio quality of the Squeezebox that has been
>> nagging me for a bit. I use the digital output of the Squeezebox to
>> feed an external DAC (conrad johnson), and set the digital output
>> level setting at the server to 'fixed'. Until a month ago I was using
>> a 5.1 nightly build (1/31 - it was very stable for me), and then
>> finally upgraded to 5.1.1. After a few days of using the new version I
>> had the nagging feeling that the sound quality was not quite as good
>> as it was with the earlier release. I switched back to the older build
>> and sure enough heard an improvement in quality. I've spent the past
>> few weeks alternately listening to each version and find that there is
>> indeed a small, but noticeable difference in the quality of the sound.
>> Note that there are firmware changes between these two server
>> versions, and I update the Squeezebox with each change.
>>
>> Given that I'm simply feeding the digital source directly to the DAC,
>> I can't account for a reason why a difference might exist. I'm
>> guessing that it falls into one of 3 categories:
>>
>> 1. The timing in the digital output has somehow changed between the
>> two firmware versions and it effects the way my DAC is able to decode
>> the music.
>> 2. Whatever processing is going on within the Squeezebox to transmit
>> the bits out is somehow interfering with them to a degree.
>> 3. There are no differences and I'm being fooled by my own expectations.
>>
>> I'm curious as to whether anyone else has experienced this difference,
>> and if anyone might be able to shed some light what might cause this
>> change.
>>
>> Thanks,
>> Ken
>>
>>

Stuart Hickinbottom
2004-04-13, 11:17
Are you sure that the server isn't now transcoding your FLACs to MP3? I
noticed this a week or so ago in the following thread:
http://article.gmane.org/gmane.music.equipment.slimdevices.general/7005

The solution is to turn off the "FLAC/MP3/*" entry in "server
settings->file types" - that worked for me. I'm hoping for a neater
solution to this as the other posts to the thread cover.

Let me know how you get on,
Stuart

On Sat, 10 Apr 2004 18:09:28 -0700, Ken <sqz (AT) lemmingstothesea (DOT) com> wrote:

> Thanks for the response Mark. I'm glad to hear that things actually did
> change between the versions, and that I'm not simply imagining
> differences. You may be correct that the audible changes I'm hearing are
> simply different and not actually worse, and that the older version is
> my preferred sound. For reference, the changes I'm hearing between 5.1
> and 5.1.1 fall into these categories:
>
> 1. Slightly slower attack time across the entire spectrum, with greatest
> prominence at the bottom end. However, there's an overall increase in
> midbass response with 5.1.1.
>
> 2. Less discrete location of instruments across the sound stage. With
> the earlier 5.1 version, the images are very distinct and solidly
> located. With 5.1.1 they tend to be more diffuse and are harder to pin
> down.
>
> 3. Slight rolloff in the higher frequencies. This may be a tonal balance
> issue more than an actual rolloff, but to my ears there is noticeably
> less energy at the very top end with the 5.1.1 version.
>
> I understand that all of these changes can just be artifacts of the DAC
> I'm using, and may have nothing to do with the proper transmission of
> the bits. However, I did want to see what might have changed between the
> versions (thank you for filling me in on that), and possibly bring up
> the issue in case one or more of the changes made did cause a
> degradation in the sound.
>
> Thanks,
> Ken
>
> BTW - You're correct in assuming that I'm using lossless (FLAC) encoding
> of my audio files.
>
> Mark Bennett wrote:
>> There were a number of improvements in the digital out behaviour between
>> 5.1 and 5.1.1. All of these have so far been seen as improvements to the
>> compatibility of the box, with this being the first suggestion I've seen
>> that anything might have got worse....
>> The changes I can think were introduced in 5.1.1 are:
>> 1) correction of sample frequency from 44.19kHz to 44.1kHz.
>> 2) corrected the level of the "validity" bit in the s/pdif stream
>> 3) correction of sonic glitches between tracks
>> There might be others, but I haven't gone trawling through the
>> checkins, and the release notes are silent on these changes.
>> I assume that you're using some form of lossless coding, otherwise
>> you probabaly wouldn't have been able to notice the difference.
>> I suspect that the change in sample frequency is probably the main
>> cause of whatever you're hearing. I suspect it just makes the
>> sound different from what you're used to, not necessarily worse.
>> Can you articulate the difference you think you're hearing? Just
>> "not quite as good" is difficult to discuss.
>> Ken wrote:
>>
>>> I have an issue with the audio quality of the Squeezebox that has been
>>> nagging me for a bit. I use the digital output of the Squeezebox to
>>> feed an external DAC (conrad johnson), and set the digital output
>>> level setting at the server to 'fixed'. Until a month ago I was using
>>> a 5.1 nightly build (1/31 - it was very stable for me), and then
>>> finally upgraded to 5.1.1. After a few days of using the new version I
>>> had the nagging feeling that the sound quality was not quite as good
>>> as it was with the earlier release. I switched back to the older build
>>> and sure enough heard an improvement in quality. I've spent the past
>>> few weeks alternately listening to each version and find that there is
>>> indeed a small, but noticeable difference in the quality of the sound.
>>> Note that there are firmware changes between these two server
>>> versions, and I update the Squeezebox with each change.
>>>
>>> Given that I'm simply feeding the digital source directly to the DAC,
>>> I can't account for a reason why a difference might exist. I'm
>>> guessing that it falls into one of 3 categories:
>>>
>>> 1. The timing in the digital output has somehow changed between the
>>> two firmware versions and it effects the way my DAC is able to decode
>>> the music.
>>> 2. Whatever processing is going on within the Squeezebox to transmit
>>> the bits out is somehow interfering with them to a degree.
>>> 3. There are no differences and I'm being fooled by my own
>>> expectations.
>>>
>>> I'm curious as to whether anyone else has experienced this difference,
>>> and if anyone might be able to shed some light what might cause this
>>> change.
>>>
>>> Thanks,
>>> Ken
>>>
>>>

Ken
2004-04-13, 22:45
Stuart,

Thanks very much for the reference. Yes, I believe that its decoding flac to wav in both versions. Looking at active processes in the Windows task manager, I see flac.exe when the Squeezebox is playing, but never lame.exe as I'd expect if it was transcoding to mp3. I'm guessing that things are working as they should with both servers.

Many thanks,
Ken

Stuart Hickinbottom wrote:
> Are you sure that the server isn't now transcoding your FLACs to MP3? I
> noticed this a week or so ago in the following thread:
> http://article.gmane.org/gmane.music.equipment.slimdevices.general/7005
>
> The solution is to turn off the "FLAC/MP3/*" entry in "server
> settings->file types" - that worked for me. I'm hoping for a neater
> solution to this as the other posts to the thread cover.
>
> Let me know how you get on,
> Stuart
>
> On Sat, 10 Apr 2004 18:09:28 -0700, Ken
> <sqz (AT) lemmingstothesea (DOT) com> wrote:
>
>> Thanks for the response Mark. I'm glad to hear that things actually
>> did change between the versions, and that I'm not simply imagining
>> differences. You may be correct that the audible changes I'm hearing
>> are simply different and not actually worse, and that the older
>> version is my preferred sound. For reference, the changes I'm hearing
>> between 5.1 and 5.1.1 fall into these categories:
>>
>> 1. Slightly slower attack time across the entire spectrum, with
>> greatest prominence at the bottom end. However, there's an overall
>> increase in midbass response with 5.1.1.
>>
>> 2. Less discrete location of instruments across the sound stage. With
>> the earlier 5.1 version, the images are very distinct and solidly
>> located. With 5.1.1 they tend to be more diffuse and are harder to
>> pin down.
>>
>> 3. Slight rolloff in the higher frequencies. This may be a tonal
>> balance issue more than an actual rolloff, but to my ears there is
>> noticeably less energy at the very top end with the 5.1.1 version.
>>
>> I understand that all of these changes can just be artifacts of the
>> DAC I'm using, and may have nothing to do with the proper
>> transmission of the bits. However, I did want to see what might have
>> changed between the versions (thank you for filling me in on that),
>> and possibly bring up the issue in case one or more of the changes
>> made did cause a degradation in the sound.
>>
>> Thanks,
>> Ken
>>
>> BTW - You're correct in assuming that I'm using lossless (FLAC)
>> encoding of my audio files.
>>
>> Mark Bennett wrote:
>>
>>> There were a number of improvements in the digital out behaviour between
>>> 5.1 and 5.1.1. All of these have so far been seen as improvements to the
>>> compatibility of the box, with this being the first suggestion I've seen
>>> that anything might have got worse....
>>> The changes I can think were introduced in 5.1.1 are:
>>> 1) correction of sample frequency from 44.19kHz to 44.1kHz.
>>> 2) corrected the level of the "validity" bit in the s/pdif stream
>>> 3) correction of sonic glitches between tracks
>>> There might be others, but I haven't gone trawling through the
>>> checkins, and the release notes are silent on these changes.
>>> I assume that you're using some form of lossless coding, otherwise
>>> you probabaly wouldn't have been able to notice the difference.
>>> I suspect that the change in sample frequency is probably the main
>>> cause of whatever you're hearing. I suspect it just makes the
>>> sound different from what you're used to, not necessarily worse.
>>> Can you articulate the difference you think you're hearing? Just
>>> "not quite as good" is difficult to discuss.
>>> Ken wrote:
>>>
>>>> I have an issue with the audio quality of the Squeezebox that has
>>>> been nagging me for a bit. I use the digital output of the
>>>> Squeezebox to feed an external DAC (conrad johnson), and set the
>>>> digital output level setting at the server to 'fixed'. Until a
>>>> month ago I was using a 5.1 nightly build (1/31 - it was very
>>>> stable for me), and then finally upgraded to 5.1.1. After a few
>>>> days of using the new version I had the nagging feeling that the
>>>> sound quality was not quite as good as it was with the earlier
>>>> release. I switched back to the older build and sure enough heard
>>>> an improvement in quality. I've spent the past few weeks
>>>> alternately listening to each version and find that there is indeed
>>>> a small, but noticeable difference in the quality of the sound.
>>>> Note that there are firmware changes between these two server
>>>> versions, and I update the Squeezebox with each change.
>>>>
>>>> Given that I'm simply feeding the digital source directly to the
>>>> DAC, I can't account for a reason why a difference might exist.
>>>> I'm guessing that it falls into one of 3 categories:
>>>>
>>>> 1. The timing in the digital output has somehow changed between the
>>>> two firmware versions and it effects the way my DAC is able to
>>>> decode the music.
>>>> 2. Whatever processing is going on within the Squeezebox to
>>>> transmit the bits out is somehow interfering with them to a degree.
>>>> 3. There are no differences and I'm being fooled by my own
>>>> expectations.
>>>>
>>>> I'm curious as to whether anyone else has experienced this
>>>> difference, and if anyone might be able to shed some light what
>>>> might cause this change.
>>>>
>>>> Thanks,
>>>> Ken
>>>>
>>>>