[opendtv] Re: New Philips HDD/DVD recorder

I've always wondered how a single receiver deals with multiple time 'values'
for the current time.  I don't think that ignoring PSIP time is a reasonable
way to do that, but it's easier than having a different time base for each
channel.

By the way, A/65 requires time to be accurate within +/- 1 second.  So,
stations outside that are liable for FCC fines (and, this isn't theoretical;
the FCC has imposed such fines, but usually the delta is greater than a few
minutes.)

Since some of us are using this thread to discuss mini-failures in the DTV
transition, here's another one: never changing the times of sporting events
to deal with program overruns.

This directly relates to PSIP time, since the time propounded by the station
is supposed to relate to the start time of programming events.  Without
updated start times to reflect actual start times, PVRs don't know when to
start or stop the recording of a sporting or other such event that can
over-run.

Sinclair seems to be the only station group to have addressed this issue.
Not the network owned groups, not the networks.  

The real key, Bert, is to see if the programming events start at the clock
time point, or at the PSIP time point.  If the latter, PVRs that use the
station's time should still work okay, at least recording the scheduled
event.

I've just spent the last month or so solving all the PSIP time issues with
my PSIP generator at the first customer's station.  We never had time more
than +12 seconds from what it should be, but I fixed even that.  Indeed, I
'fixed' it a few times, but there hasn't been an issue there for more than a
few weeks.

Here's how you do it.  You have the PSIP generator based on windows 2000 or
later, and you set the time server in windows to one or more authoritative
network sources via the internet.  Windows does the updating automatically.
Then, you have the PSIP generator application derive system_time from the
computer's time base, and use that system_time value for all internal
time-sensitive matters (like changing event information tables and extended
event text tables every three hours.)

As a check, EtherGuide Emissary on the front panel displays UTC time and
local time each second, and the former is derived directly from the
transmitted system time table, while the latter is derived from the
computer's internal clock independent of the application.  So, at a glance,
a user can see if the times are offset by the appropriate value.

As a competitor, I'm glad to see that there are plenty of PSIP generators
out there that haven't solved this relatively simple task, one that can
cause FCC fines.  My goal is to have time accurate to within a few
milliseconds, but the latter will require customers to adjust the "time
transmission delta" to deal with encode and decode latency -- which is
usually around 4 or 5 seconds.

I think my PSIP generator
(http://www.EtherGuideSystems.com/Systems/Emissary/Default.aspx)
 is the only unit that permits fine-tuning of PSIP time, in 1 ms increments.

John Willkie
www.EtherGuideSystems.com




-----Mensaje original-----
De: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx] En
nombre de Albert Manfredi
Enviado el: Thursday, December 27, 2007 5:33 PM
Para: opendtv@xxxxxxxxxxxxx
Asunto: [opendtv] Re: New Philips HDD/DVD recorder


John Shutt wrote:

> I don't think that is as big an issue since the NTIA coupons
> are intended for STBs to be used with NTSC 4:3 displays.
>
> The amount of letterboxed 4:3 material far outweighs
> pillarboxed 16:9 material. It is only the latter that would
> have the "postage stamp" effect on a 4:3 display.

Hmmm. From what I can determine, as an end user with one 16:9 and one 4:3
set connected to Accurian and Digital Stream STBs, when told that the
monitor has a 16:9 aspect ratio, it's the STB that creates the pillarboxed
effect. At least, that seems to be by far the most frequent reason for
pillarboxing. On the 16:9 set, all non-distorted SDTV programming comes in
that way. So it's very common.

What these STBs allow, that the Philips PVR does not do, is to set each
channel and have that setting stored in memory. The STBs give these options:

1. Normal, unstretched video.
2. Stretch out the video horizontally, to fill the 16:9 screen.
3. Zoom in, to expand a letterboxed+pillarboxed image to fill 16:9.

Mode 2, stretching out the picture, is usually required to undistort a
widescreen image. Most of the subchannels that transmit HD seem to need that
set all the time, even if they transmit SD. On the other hand, the SD-only
subchannels do not have that set, or you get the short and fat effect.

Mode 3 is not always allowed. But it is perfect for shows that are
transmitted as SD letterboxed to 4:3 sets, and come in as
letterboxed+pillarboxed on my 16:9 set. PBS uses this a lot. The problem
seems to be that zooming is not allowed if the station is requiring the mode
2 to be set. You can't zoom in on a horizontally-stretched image. So you're
out of luck if a station transmits letterboxed on a subchannel that it
requires to be stretched.

Just having these three settings in the receiver, remembered for each
station, is actually enough. I'd be totally satisfied if all receivers
allowed this.

Here are some other observations.

Evidently, Philips is so unimpressed with PSIP time that they don't even use
it in the PVR. You can either use the analog PBS time service, or you can
set the clock manually (I chose the latter option). In fact, the Philips
won't even display time of day from PSIP, as far as I've been able to
determine.

So, having set my STB back to full auto time (since I don't need the STB
timer for my time-shift recordings anymore), I have now been able to try all
of my DT stations to see how well they're doing. I have one of those clocks
next to the TV that's tuned to Boulder Colorado on 60 MHz. To the nearest
minute, here are the results, where only the good guys are identified. By
the way, this is much improved from last time I checked.

Ch 2:    accurate  (WMAR ABC Baltimore)
Ch 4:   +2 min
Ch 5:   +1 min
Ch 7:   +4 min
Ch 9:   accurate  (WUSA CBS DC)
Ch 11: +3 min
Ch 13: +3 min
Ch 14: +2 min
Ch 20: +2 min
Ch 22: accurate (WMPT PBS Annapolis)
Ch 26: accurate (WETA PBS DC)
Ch 30: no reception with Accurian today, no time shown on Philips
Ch 45: -3 min
Ch 50: +17 hours 20 min
Ch 54: +2 min

It's no doubt for this reason that timed recording can't depend on PSIP.
Still, last time, many stations were more than 5 minutes off, so I consider
this a step ahead. Not too say acceptable, of course. This time thing just
isn't that hard to get right, sorry, guys. Imagine how many ships would run
aground if they navigated this way?

And, I really do enjoy the HDD in the Philips.

Bert

_________________________________________________________________
Don't get caught with egg on your face. Play Chicktionary!
http://club.live.com/chicktionary.aspx?icid=chick_wlhmtextlink1_dec 
 
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.




 
 
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at 
FreeLists.org 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: