proms 320kbps
 

On Fri, 10 Sep 2010 09:30:31 +0100, Jim Lesurf
wrote:

In article , Don Pearce
wrote:
On Thu, 09 Sep 2010 09:33:22 +0100, Jim Lesurf
wrote:


TBH at present I've not decided if for me 320k gives a satisfactory
trade-off between reliability of stream connection and quality. Dunno
about you, but so far, each time I've tried the current experimental
stream I get drop outs on average about once or twice an hour (about 1
sec lost each time) which rather disrupts listening to a long piece of
music. Wheras at 128k or 192k I almost never get such dropouts.

Slainte,

Jim

How is your broadband connection though?

Fast enough that I can normally watch the *TV* iPlayer without problems
if wish.

For reference I just checked using the BBC iplayer diagnostic and for both
my laptop and shuttle this gave 2400 kbps for download speed and 2200 for
streaming. I'd agree that isn't very fast. But it should be ample for
320k *provided* the connection streams reliably.

I also use 256/320 from elsewhere at times. The problem can be with
occasional brief delays. My guess is that this is simply a machine
somewhere along the way being temporarily otherwise occupied or some
packets going AWOL. Hence I suspect this is a matter of the levels of
buffering/retries/requests not being quite adequate.

Slainte,

Jim

Do the Beeb even allow buffering on these services? I never see the
progress bar moving beyond the point in the programme that has been
reached. I think that if there is even a brief glitch in the stream,
you will hear it.

d

proms 320kbps
 

In article , Jim Lesurf
scribeth thus
In article , froggy
wrote:
Le 04/09/2010 18:46, David Pitt a écrit :
Jim Lesurf wrote:

Surprised no-one else seems to have mentioned this here yet. :-)

http://www.bbc.co.uk/proms/2010/audioexperiment/

There is a thread on alt.radio.digital.



Jim,

Perhaps you could use your vast influence at the Beeb [:-)]

sic :-)

so that this "experiment" continues after the The Last Night!

TBH at present I've not decided if for me 320k gives a satisfactory
trade-off between reliability of stream connection and quality. Dunno about
you, but so far, each time I've tried the current experimental stream I get
drop outs on average about once or twice an hour (about 1 sec lost each
time) which rather disrupts listening to a long piece of music. Wheras at
128k or 192k I almost never get such dropouts.

Yes thats happening here. I haven't bothered to check the other streams.
Mind you theres a couple of TV progs of iplayer being used elsewhere
here;!..

--
Tony Sayer

proms 320kbps
 

In article , Don Pearce
scribeth thus
On Thu, 09 Sep 2010 09:33:22 +0100, Jim Lesurf
wrote:

In article , froggy
wrote:
Le 04/09/2010 18:46, David Pitt a écrit :
Jim Lesurf wrote:

Surprised no-one else seems to have mentioned this here yet. :-)

http://www.bbc.co.uk/proms/2010/audioexperiment/

There is a thread on alt.radio.digital.



Jim,

Perhaps you could use your vast influence at the Beeb [:-)]

sic :-)

so that this "experiment" continues after the The Last Night!

TBH at present I've not decided if for me 320k gives a satisfactory
trade-off between reliability of stream connection and quality. Dunno about
you, but so far, each time I've tried the current experimental stream I get
drop outs on average about once or twice an hour (about 1 sec lost each
time) which rather disrupts listening to a long piece of music. Wheras at
128k or 192k I almost never get such dropouts.

Slainte,

Jim

How is your broadband connection though? I pay a little extra for the
professional option from Plusnet. It guarantees maximum line bandwidth
on all ports and services. I was downloading the latest TomTom Western
Europe map the other day, and had the network monitor running to see
how it was going. Here is what the monitor made of most of the
transfer.

http://www.soundthoughts.co.uk/look/adsl.png

Rock solid 16Mb/sec and no dropouts. That is about 1GByte on that
display. That makes light work of 320kb/sec.

d

Lucky you!. Here we're on VM and very good that is too, their 10 meg
service, does all we need they can do a 20 and a 50 IIRC, but there are
people less than three miles from Cambridge who can barely muster 1 Meg
whilst paying for 8 on god ole BT ally cable;(...

If the BBC wanted to improve their digital broadcasting then UP the
rates on Freeview and Freesat, as I might have muttered before the
Germans can manage 334 K for their Klassik service why can't olde auntie
BBC.

OK, this is a step in the right direction and as time goes on the net
may become the platform of choice, well until theres a lot more Fibre to
the home or nearer.. like what VM do.

And then theres the net connection and PC to the living room or place
when the good audio system is. I have got a satellite receiver that has
got LAN facilities on but its not too easy to get that to decode the
stream required and fortunately we have a house wired for CAT 5 and a 24
port switch most domestic premises aren't so well equipped and Radio on
2.4 or 5.8 is OK but in some premises the interference experienced
nowadays is limiting..

And for most a noisy PC added to the living room complement, whereas a
digital sat receiver is quite commonplace and a much easier way of
receiving the higher quality "Radio" Broadcast signals;)...
--
Tony Sayer

proms 320kbps
 

"tony sayer" wrote

OK, this is a step in the right direction and as time goes on the net
may become the platform of choice, well until theres a lot more Fibre to
the home or nearer.. like what VM do.


Yes, but fibre to the home cost money, and who is going to pay? The public
have been lead to believe that access to the internet should be really cheap
so resent paying the money that would allow adequate investment in the
infrastructure. So unless the government can be persuaded to stump up the
cash (unlikely in the present circumstances!) this country is likely to
continue to suffer from a patchy quality of internet service.

David.

proms 320kbps
 

In article , David Looser
scribeth thus
"tony sayer" wrote

OK, this is a step in the right direction and as time goes on the net
may become the platform of choice, well until theres a lot more Fibre to
the home or nearer.. like what VM do.


Yes, but fibre to the home cost money, and who is going to pay? The public
have been lead to believe that access to the internet should be really cheap
so resent paying the money that would allow adequate investment in the
infrastructure. So unless the government can be persuaded to stump up the
cash (unlikely in the present circumstances!) this country is likely to
continue to suffer from a patchy quality of internet service.


I quite agree David. But the net was a sort of data info place like a
large file server for not that big docs. Its grown since and perhaps we
need or someone somewhere needs to grow. I don't suppose whatsisname who
dreamt it up would have foreseen iplayer etc..


However give her her due, old auntie BBC has scored on this one as most
all of the comments on the "experiment" have shown...

Course they could up the rates on satellite a much more practical method
of digital radio "broadcasting" and simpler for most all users too, just
a skyboxen or digital sat receiver;)).. No TV side PC required;!..


http://www.bbc.co.uk/blogs/bbcintern...xtra_high_qual
ity_audio.html
--
Tony Sayer

proms 320kbps
 

In article , Don Pearce
wrote:
On Fri, 10 Sep 2010 09:30:31 +0100, Jim Lesurf
wrote:



How is your broadband connection though?

Fast enough that I can normally watch the *TV* iPlayer without problems
if wish.

For reference I just checked using the BBC iplayer diagnostic and for
both my laptop and shuttle this gave 2400 kbps for download speed and
2200 for streaming. I'd agree that isn't very fast. But it should be
ample for 320k *provided* the connection streams reliably.

I also use 256/320 from elsewhere at times. The problem can be with
occasional brief delays. My guess is that this is simply a machine
somewhere along the way being temporarily otherwise occupied or some
packets going AWOL. Hence I suspect this is a matter of the levels of
buffering/retries/requests not being quite adequate.

Slainte,

Jim

Do the Beeb even allow buffering on these services?

Depends on what you mean by "allow".

Yes, the stream is certainly bufferred. If I monitor the data transfers
with gnome-system-monitor I can see a graph of how the transfer rate varies
with time. And with the 128/192k 'normal' services it is easy to see the
flickering of the lights on my router varying in sympathy with this.

For the 128/192k streams I tend to get a low steady level, and every few
tens of seconds a 'burst' or higher rate transfer. So I assume the steady
rate is just slightly too low to keep the rx buffer up to target and an
occasional burst is called for to help fill it to a higher level.

The difficulty is that I suspect this all takes place interactively under
control of the Flash code. So not easily changable by the user. The problem
then becomes that any delays or underfilling may mean that the 320k rate
exhausts the RX buffer and it then takes time to refill.

If you think about it, aac is a spectral block code. So the data *must* be
buffered to some extent - albeit only for a few blocks if the transfer was
utterly reliable and ultra-fast.

Shame the test is so short. One thing I wanted to try was altering the
workspace assigned to the Flash plugin to see if that helped. Alas, I often
can't use a 'live' service as 'real world' (TM) interrupts take priority.
So occasions for me to participate in the public test are limited. Since
the 1sec gaps are rare I need some time to get any meaningful stats.

And I've now found that up here in the frozen north we are *not* going to
be allowed to see the first half of the 'Last Night' from the RAH! Instead
we are being given an 'opt out' for what looks like it may be a 'shortbread
and tartan' alternative from Dundee. I won't say my reaction as it isn't
for delicate ears! :-) Oh well, at least we have Radio 3 where the
pictures are better, anyway. But it means I won't have a 'BBCTV Freeview'
version for the first half as a comparison.

Or would someone would be willing to make either a (English) Freeview or
FreeSat LPCM wav recording of the first half on TV for me to analyse?

I never see the progress bar moving beyond the point in the programme
that has been reached. I think that if there is even a brief glitch in
the stream, you will hear it.

The progress bar tends to be too crude a display for this and tends to show
what has been played or 'elapsed'. And brief gaps tend to be 'faded' by the
system, so not as abrupt as you might assume. 1 sec is obvious. But smaller
ones may easily pass unnoticed.

All being well, I'll be able to say more when I've finished collecting data
and do some analysis. Currently working on some more analysis routines to
go with the ones I already have.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

proms 320kbps
 

On Sat, 11 Sep 2010 09:54:59 +0100, Jim Lesurf
wrote:

In article , Don Pearce
wrote:
On Fri, 10 Sep 2010 09:30:31 +0100, Jim Lesurf
wrote:



How is your broadband connection though?

Fast enough that I can normally watch the *TV* iPlayer without problems
if wish.

For reference I just checked using the BBC iplayer diagnostic and for
both my laptop and shuttle this gave 2400 kbps for download speed and
2200 for streaming. I'd agree that isn't very fast. But it should be
ample for 320k *provided* the connection streams reliably.

I also use 256/320 from elsewhere at times. The problem can be with
occasional brief delays. My guess is that this is simply a machine
somewhere along the way being temporarily otherwise occupied or some
packets going AWOL. Hence I suspect this is a matter of the levels of
buffering/retries/requests not being quite adequate.

Slainte,

Jim

Do the Beeb even allow buffering on these services?

Depends on what you mean by "allow".


Well, consider a comparison with something like Youtube. When I use
that, I see the "downloaded" portion of the progress bar move swiftly
across to completion in not much more than a few seconds. the video
then effectively plays off my hard drive. I don't get that with BBC
streams. There is clearly a small amount of buffering, but only a tiny
amount gets downloaded ahead of playing and certainly not the entire
programme. I don't know what management is available to the Beeb for
this, but I suspect they know lots of people don't watch whole
programmes, so they want to limit the amount of wasted data transfer
by preventing the buildup of a huge buffer. A result of not quite
getting the balance right here would be occasional dropouts.

Yes, the stream is certainly bufferred. If I monitor the data transfers
with gnome-system-monitor I can see a graph of how the transfer rate varies
with time. And with the 128/192k 'normal' services it is easy to see the
flickering of the lights on my router varying in sympathy with this.

For the 128/192k streams I tend to get a low steady level, and every few
tens of seconds a 'burst' or higher rate transfer. So I assume the steady
rate is just slightly too low to keep the rx buffer up to target and an
occasional burst is called for to help fill it to a higher level.

The difficulty is that I suspect this all takes place interactively under
control of the Flash code. So not easily changable by the user. The problem
then becomes that any delays or underfilling may mean that the 320k rate
exhausts the RX buffer and it then takes time to refill.

If you think about it, aac is a spectral block code. So the data *must* be
buffered to some extent - albeit only for a few blocks if the transfer was
utterly reliable and ultra-fast.


Yes, but that is, as you say, a necessary buffer to allow extraction
of the error-corrected signal.

Shame the test is so short. One thing I wanted to try was altering the
workspace assigned to the Flash plugin to see if that helped. Alas, I often
can't use a 'live' service as 'real world' (TM) interrupts take priority.
So occasions for me to participate in the public test are limited. Since
the 1sec gaps are rare I need some time to get any meaningful stats.

And I've now found that up here in the frozen north we are *not* going to
be allowed to see the first half of the 'Last Night' from the RAH! Instead
we are being given an 'opt out' for what looks like it may be a 'shortbread
and tartan' alternative from Dundee. I won't say my reaction as it isn't
for delicate ears! :-) Oh well, at least we have Radio 3 where the
pictures are better, anyway. But it means I won't have a 'BBCTV Freeview'
version for the first half as a comparison.

Or would someone would be willing to make either a (English) Freeview or
FreeSat LPCM wav recording of the first half on TV for me to analyse?

I never see the progress bar moving beyond the point in the programme
that has been reached. I think that if there is even a brief glitch in
the stream, you will hear it.

The progress bar tends to be too crude a display for this and tends to show
what has been played or 'elapsed'. And brief gaps tend to be 'faded' by the
system, so not as abrupt as you might assume. 1 sec is obvious. But smaller
ones may easily pass unnoticed.

All being well, I'll be able to say more when I've finished collecting data
and do some analysis. Currently working on some more analysis routines to
go with the ones I already have.

Slainte,

Jim

I'll be interested.

d

proms 320kbps
 

In article , Don Pearce
wrote:
On Sat, 11 Sep 2010 09:54:59 +0100, Jim Lesurf
wrote:


Do the Beeb even allow buffering on these services?

Depends on what you mean by "allow".


Well, consider a comparison with something like Youtube. When I use
that, I see the "downloaded" portion of the progress bar move swiftly
across to completion in not much more than a few seconds. the video then
effectively plays off my hard drive.

From what I've been told that is because the BBC iPlayer works in its own
way. The process is effectively dominated by the BBC wanting to 'wrap' it
all into a Flash process So the iPlayer isn't necessarily like Youtube,
etc.

FWIW I've just checked and my Flash setup was limiting the BBC to just
100kB of space. So I've changed that to 1MB in case I get a chance to see
if that makes any difference.

All being well, I'll be able to say more when I've finished collecting
data and do some analysis. Currently working on some more analysis
routines to go with the ones I already have.


I'll be interested.

One preliminary result may interest you. This is that some 'continuity' and
'interval talk' portions have a large number of very short sequences of
zero-value samples.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

proms 320kbps
 

On Sat, 11 Sep 2010 12:11:52 +0100, Jim Lesurf
wrote:

In article , Don Pearce
wrote:
On Sat, 11 Sep 2010 09:54:59 +0100, Jim Lesurf
wrote:


Do the Beeb even allow buffering on these services?

Depends on what you mean by "allow".


Well, consider a comparison with something like Youtube. When I use
that, I see the "downloaded" portion of the progress bar move swiftly
across to completion in not much more than a few seconds. the video then
effectively plays off my hard drive.

From what I've been told that is because the BBC iPlayer works in its own
way. The process is effectively dominated by the BBC wanting to 'wrap' it
all into a Flash process So the iPlayer isn't necessarily like Youtube,
etc.

I thought Youtube videos were Flash - they certainly download with a
..flv postscript.

FWIW I've just checked and my Flash setup was limiting the BBC to just
100kB of space. So I've changed that to 1MB in case I get a chance to see
if that makes any difference.

All being well, I'll be able to say more when I've finished collecting
data and do some analysis. Currently working on some more analysis
routines to go with the ones I already have.


I'll be interested.

One preliminary result may interest you. This is that some 'continuity' and
'interval talk' portions have a large number of very short sequences of
zero-value samples.


Huh?

d

proms 320kbps
 

In article , Don Pearce
wrote:
On Sat, 11 Sep 2010 12:11:52 +0100, Jim Lesurf
wrote:

In article , Don Pearce
wrote:
On Sat, 11 Sep 2010 09:54:59 +0100, Jim Lesurf
wrote:


Do the Beeb even allow buffering on these services?

Depends on what you mean by "allow".


Well, consider a comparison with something like Youtube. When I use
that, I see the "downloaded" portion of the progress bar move swiftly
across to completion in not much more than a few seconds. the video
then effectively plays off my hard drive.

From what I've been told that is because the BBC iPlayer works in its
own way. The process is effectively dominated by the BBC wanting to
'wrap' it all into a Flash process So the iPlayer isn't necessarily
like Youtube, etc.

I thought Youtube videos were Flash - they certainly download with a
.flv postscript.

That does not mean that the Flash code you download and use is exactly the
same in both such cases. Flash is essentially a programming / scripting
language aimed at specific sorts of tasks. So the two organisations each
use it in their own ways.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html