IZ3LSV

G8MNY  > DIGTV    16.02.11 23:56l 108 Lines 4997 Bytes #999 (0) @ EU
BID : 7718_GB7CIP
Read: GUEST
Subj: Digital TV so far
Path: IZ3LSV<F6CDD<GB7CIP
Sent: 110216/2244Z @:GB7CIP.#32.GBR.EU #:7718 [Caterham] $:7718_GB7CIP
From: G8MNY@GB7CIP.#32.GBR.EU
To  : DIGTV@EU

By G8MNY                                      (Updated Jan 08)
(8 Bit ASCII graphics use code page 437 or 850, Terminal Font)

In Croydon, South London, I am line of sight to main Tx for London, but DVB-T
(Freeview) multiplex channels are still about 25dB weaker here than the 1MW ERP
analogue signals.

       Analogue     DIGITAL MUX
   <-----8MHz-----><----8MHz----->
    Carrier                         v
    ³         Sound
    ³          ³                   25dB
    ³   Colour ³Nicam
    ³       ³  ³ Û ÛÛÛÛÛÛÛÛÛÛÛÛÛÛ   ^
  ÄÄÁÄÄÄÄÄÄÄÁÄÄÁÄßÄßßßßßßßßßßßßßßÄÄ

To cut down on the number of boxes connected to the TV, I use a Freeview VCR
(now obsolite as the freeview Data format has changed) in front of the main
16:9 Plasma 42" TV, fed from a large 25 element loft Group A aerial which I
have broadbanded a bit by trimming elements for a more even signals display on
the STB/VCR, giving 25-27dB S/N on all muxes. Note that 17dB S/N is the no go
error threshold.

  Mux                        Mux
 Signal   Original          Signal   Modified Aerial
  30dB´                       30dB´
  25dB´ Û Û Ü Ü               25dB´ Ü Û Û Û Û Ü
  20dB´ Û Û Û Û Ü             20dB´ Û Û Û Û Û Û
  15dB´_Û_Û_Û_Û_Û_Ü_\Freq     15dB´_Û_Û_Û_Û_Û_Û_\Freq
       Ã GROUP A ´  /              Ã GROUP A ´  /

I used another identical STB/VCR a 1st floor bedroom from a small 4 element set
top aerial into an standard 4:3. This gives similar good signal results.

PICTURE QUALITY
On the big TV, picture definition is generally higher than I expected, with the
16:9 format apparently giving about 7MHz (720 pixels/line) resolution over the
RGB SCART feed.

Of course there is no visible noise (snow) with a digital picture, just other
artifacts! In fact real falling snow can't be encoded in Mpeg as there are no
similar frames to allow for compression!

However the pictures are not @ the 50Hz new frame rate of true interlaced
analogue. e.g. the net frame rate is often very low! So parts of the most
visible picture are updated 1st then a few frames later the darker bits!
A actor's face half in the dark turning, will appear s t r e t c h e d   o u t
until the darker bits get updated. A very strange effect when it happens &
sometimes repeatedly!

     Û²±° ³         Û²±°   ³         Û²±° ³
     Û0 0 ³         Û0 0   ³         Û0 0 ³
     Þ U ³          Þ U   ³          Þ U ³
      Ý= ³           Ý=   ³           Ý= ³
      ßß             ßß                ßß
      HEAD       HEAD < MOVED     1/2 Sec later

But this is a rare affect, the more usual effect will be highlighted forehead
detail not following the head movement at all well, or a shakey camera shot
with parts of the picture cut out & moving independantly.

Also there is what seems to be data bandwidth competition (bidding) on any of
the broadcaster MUXs, so that occasionally a NEEDED new "I" (full) frame for a
shot change has no bandwidth available for say 1 second, so the picture just
hangs for that time with no errors or sound faults.

As far as I know these 2 affects are all the result of the EXTREME MPEG OVER
COMPRESSION that the broadcaster has chosen to use for that programme, & are
nothing to do with weak signals & error rate that cause the [] blocks to appear
& "glitch" noises in the sound. I no longer get those after a new coax feed to
the loft aerial.

DATA RATES
From a multi Mux Tx site stats printout I saw (published on the Internet) the
typical data rates per channel vary from an extremely low 50kB/S to a peak of
6MB/S in any 5 mins giving means of 1-2MB/ch. This is still extremely low
compared to the uncompressed >200MB/S from a studio camera source. (a HDTV
source is 1200MB/S)

BBC LIPSYNC
Since using the larger TV, poor lipsync problems are far more noticeable. Other
programme makers don't seem to suffer from this, it definitely is a BBC thing.
I know the plasma TV might add a 1/50 or 1/25S picture delay in its picture re-
formatting & hence add to the problem, as might the larger screen make the lip
movements more apparent, but why should this be just a BBC thing? I have even
seen reasonable lip sync degrade during a BBC interview, why? I have even seen
the same problem with digital cable system on BBC channels! Are the other
broadcasters' running 1/25S delayed sound to compensate for large screen TVs,
or is the BBC output just inconsistent? I noticed they have improved their act
recently.

DVB-T TV channels.
Since the addition of loads more junk TV channels the STB's programme guide is
corrupted once/twice a day if it is put into standby. This seems to be centred
on some of the channels that keep swapping their allocations (on & off service
in the menu guide). I assume they actually go missing from the Tx guide menu
for that mux I am using for a few seconds & that my STB instantly sees the
change & flags up it need for a rescanned (early STB software fault?) each
time.


Y don't U send an interesting bul?

73 de John G8MNY @ GB7CIP


Read previous mail | Read next mail