PDA

View Full Version : syncronization drift?



Gregory P. Smith
2003-11-22, 16:42
I have two slimp3s. When playing synced, the sync tends to drift apart
so that after more than 15 minutes of playing the drift is noticable
in audio coming from the other room (and makes for a very odd sensation
walking between the rooms). After an hour or more the drift is easily
a quarter to half a second.

Is this due to each slimp3's 44.1khz source being independant with its
own inaccuracy? regardless, shouldn't the server correct for this at
every song gap by re-syncing all players at that time to avoid drift
accumulation?

Audio files are VBR mp3s (lame --alt-preset=standard). One slimp3 is on
a 10/100 switch with the server. The other slimp3 is on a much higher
latency connection due to an ethernet over powerline adaptor and two
10/100 switches.

Server software is 4.2.6.

side note: why does the server restart the current track at the beginning
when synchronize is selected from the menu? this is annoying given
that i often want to turn on sync when i'm moving between rooms having
started a playlist in one room or another.

greg

dean
2003-11-22, 21:07
The SLIMP3s do re-sync at every new song. Are you playing streams that
are longer than a few minutes?

-dean

On Nov 22, 2003, at 3:42 PM, Gregory P. Smith wrote:

> I have two slimp3s. When playing synced, the sync tends to drift apart
> so that after more than 15 minutes of playing the drift is noticable
> in audio coming from the other room (and makes for a very odd sensation
> walking between the rooms). After an hour or more the drift is easily
> a quarter to half a second.
>
> Is this due to each slimp3's 44.1khz source being independant with its
> own inaccuracy? regardless, shouldn't the server correct for this at
> every song gap by re-syncing all players at that time to avoid drift
> accumulation?
>
> Audio files are VBR mp3s (lame --alt-preset=standard). One slimp3 is
> on
> a 10/100 switch with the server. The other slimp3 is on a much higher
> latency connection due to an ethernet over powerline adaptor and two
> 10/100 switches.
>
> Server software is 4.2.6.
>
> side note: why does the server restart the current track at the
> beginning
> when synchronize is selected from the menu? this is annoying given
> that i often want to turn on sync when i'm moving between rooms having
> started a playlist in one room or another.
>
> greg
>
>

Gregory P. Smith
2003-11-23, 15:31
On Sat, Nov 22, 2003 at 08:07:46PM -0800, dean blackketter wrote:
> The SLIMP3s do re-sync at every new song. Are you playing streams that
> are longer than a few minutes?
>
> -dean

hmm. that means that mine are drifting apart noticably within a single
track (~5 mins). assuming its the differences in the 44.1k sources on
both causing the issue there's not much to be done.

I do have some 1h+ tracks but i expect noticable drift on something that
long. A 5 minute track drifting 1/4sec will happen if the frequency
bases are a mere 0.1% apart. any idea what the expected accuracy of
the oscillators on the slimp3s is?

this isn't a big deal for me, just something i've noticed.

> >I have two slimp3s. When playing synced, the sync tends to drift apart
> >so that after more than 15 minutes of playing the drift is noticable
> >in audio coming from the other room (and makes for a very odd sensation
> >walking between the rooms). After an hour or more the drift is easily
> >a quarter to half a second.
> >
> >Is this due to each slimp3's 44.1khz source being independant with its
> >own inaccuracy? regardless, shouldn't the server correct for this at
> >every song gap by re-syncing all players at that time to avoid drift
> >accumulation?

Roy M. Silvernail
2003-11-23, 15:57
On Sunday 23 November 2003 17:31, Gregory P. Smith wrote:
> On Sat, Nov 22, 2003 at 08:07:46PM -0800, dean blackketter wrote:
> > The SLIMP3s do re-sync at every new song. Are you playing streams that
> > are longer than a few minutes?
> >
> > -dean
>
> hmm. that means that mine are drifting apart noticably within a single
> track (~5 mins). assuming its the differences in the 44.1k sources on
> both causing the issue there's not much to be done.

I noticed a ~3 second lag between two synchronized slimpies on a particularly
low bit-rate stream right from the start. That was the 11/20 nightly. I
can't test again just now, though. The second slimpy is on the way back to
trade for a SqueezeBox. :)