In article , tony sayer
wrote:
In article , Jim Lesurf
scribeth thus
Hi,
Just to let people know that I've just put up a new page of results.
These compare the BBC iPlayer with the audio on DAB and TV for the
Proms this year. No surprise that I prefer the iPlayer.
Landline by another name
..
But some of the other results were unexpected!
Interesting but do we know what the signal paths were and what else was
in them transcoding anywhere perhaps?.. Even an Optimod set to classical
protect or similar?..
The reply involves me using the word 'source' in two ways...
My source(s) at the BBC (i.e. the people there who've helped with this)
tell me as follows, IIUC...
That the R3 24bit 48k from RAH comes into the central distribution desk,
then gets split there for iPlayer, FM, DAB, and the versions of R3 on 'TV'.
So common mode to that point. BBC4 TV also uses that feed. *But BBC1/2
don't*
The FM gets optimod and then is NICAM to the TXs.
The DAB goes into the DAB mp2 coder and then multiplexed out. That has a
gain adjust set. But no optimod. The filtering is part of the encoding.
(One of the things that I found out during this was the decision to lift
the level on DAB. Whereas on iPlayer they shy away from about -4dBFS.)
The iPlayer goes direct to Coyopa (iPlayer server) and sent (and saved for
'listen again' as the input LPCM at first). No processing beyond sample
rate (and perhaps level) conversion as part of the AAC encoding setup.
So in each case the splits are when any resampling (e.g. to 32k for FM) and
encoding is done.
The R3 for 'TV' gets sent to the TV people. So far as I know that just gets
mp2 encoded with the filtering as part of the encoding.
The TV side is more complex, and varies from BBC1 to BBC2 to BBC4.
BBC1 and 2 take their own feed from the hall, and may well have their own
balancing, etc. So a change may be imposed almost anywhere along the path
that the TV side use. From mics in the hall, onwards. It could be their
optimud sic, or something else...
Given this, no surprise that the BBC1/2 sound *was* different to R3. But
the puzzle is the strange 'peaks and dips' behaviour with a period that is
log-uniform! If it were different hall acoustics caused by different
balancing or mic placement I'd still be puzzled by the log-uniform pattern
being so clear. Weird.
I was also given a copy from 'source' - i.e. the same desk. But this was
resampled there to 16/44.1 to match my main target - iPlayer.
So far as I can tell, the resampling is done well. And the iPlayer output
is very close to the 'source' version.
But I can't say why the TV version has the odd behaviour. It could be a
quirk with the way the TV side process the sound. No idea. I can't tell if
the level compression is automated or simply someone with a score twiddling
the faders.
It doesn't appear when I compare iPlayer with source. What causes it, and
where in the chain, I have no idea. My source(s) at the BBC are also
puzzled. If I/they suddenly have a brainwave and twig the cause, I'll say
so.
It may be something wrong with my analysis. But, if so, we have no idea
what! So at present it remains a puzzle.
When I get a chance I may see if I can do a similar comparison for one of
the proms broadcast on BBC4 TV to see if that is 'clean'.
But at present I'm immersed in a load of research on a more 'historic'
topic. The development of the BBC FM distribution, NICAM, stereo, etc.
Having reported on the continued use of NICAM in HFN the reaction of some
readers was surprise/disbelief. That made me decide to check, and then led
me to wondering about the history of PCM and NICAM for FM distribution as
many of the details are little-known or forgotten. At the time I didn't
realise what an extensive and interesting set of events lurked... Have
since been told a number of stories and details that I hope to relate in
due course. The more I learned, the more I've felt that the good work done
should be better known and appreciated. :-)
Can't say I've heard any singing light bulbs at all;! I've heard of some
florescent ones emitting UHF signals..
I've also caught the neons in an old Quad tuner interfering with its FM
reception!
I've not heard any bulbs for decades. But then I can't hear 15kHz any
more! I suspect that any bulbs that sing at lower frequencies are heard and
dealt with. But the ones at HF pass unnoticed. The levels are also pretty
low. So you need to integrate over a long series of FFTs to show them
clearly. In a way it is a testament to the AAC that it can reproduce them
so well, even though they are probably inaudble in normal listening.
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
iPlayer versus the rest - Proms 2011
...
Linear Mode
