When BBC 1 or 2 do a split screen for an end credit promo, there always seems to be a "mini frame roll" just before it - presumably at the point where they switch whatever DVE device they are using into circuit. It's particularly noticeable when the switch happens during the credit roller, as you see the lettering jump on-screen.
A few years back, when I was a lass in the BBC, this would have been deemed unacceptable, and engineers would have been summoned to provide a suitable timing modification to stop it happening. Does nobody care any more?
It is more noticeable on some occasions than others. On BBC TWO NI (Digital) last night, at the end of the Restoration programme, very clear jump just ahead of the BBC FOUR ECP graphic. To top it all off, just before the ECP disappeared, the Restoration programme, still running on the left of the screen, froze, before going completely grey (the announcer now stumbling). The ECP then disappeared giving us a full screen of grey for five secs or so.
These very slight distortions are nothing to the years of obscene picture rolls that BBC N. ireland viewers suffered (up until only a few years ago) when the local desk went in/out of circuit on the Analogue service. BBC Engineers certainly weren't tripping over themselves on that one.
Although then again, if you think that's bad, you should see ITV's ECP!
Last night on ITV2 they showed a programme that had the old style credits with the smoky background, but the ECP system chopped it down into the new style credits box; so it chopped out half the credits on the right-hand side.
It messes up more often than not. Usually, it's just a case of the credits going back a few seconds late; so the video freezes; but sometimes you get it where the wrong bit gets shrunk, so it looks all messed up.
I think they should re introduce a system like the hearts ecp on the right hand side. It gave a unified look to all the channels, was neat and stylish, could run programme information on the left or have a background to the programme that they could match the ecp to.
These very slight distortions are nothing to the years of obscene picture rolls that BBC N. ireland viewers suffered (up until only a few years ago) when the local desk went in/out of circuit on the Analogue service. BBC Engineers certainly weren't tripping over themselves on that one.
Well they should have been! All they needed to do was adjust the station timing to match Network and hey presto - no frame roll!
It wasn't until I saw sites like this that I found that all regions on all channels DIDN'T Time-base Correct (TBC) all output as standard.
I grew up in the Midlands and the North West, then moved to Kent, then London, and NEVER saw a loss of field-sync from presentation to programme.
Then, on seeing captures and tape copies of other regions, I was stunned to see how BBC NI, and Anglia, for instance, didn't TBC their optings at all, leaving the vertical/horizontal shudder which causes, upon live viewing, your TV set to make a hiccuping sound, and on video recordings, causes a picture tearing band to sweep up the image.
On BBC TWO NI (Digital) last night, at the end of the Restoration programme, very clear jump just ahead of the BBC FOUR ECP graphic. To top it all off, just before the ECP disappeared, the Restoration programme, still running on the left of the screen, froze, before going completely grey (the announcer now stumbling). The ECP then disappeared giving us a full screen of grey for five secs or so.
It happened on Network BBC Two as well - I thought those credits finished a bit too quick before they burst on screen again.
It's wrong to say that BBC NI didn't sync up timings with network - they did - twice a day, every day (and more than that if it was deemed necessary). The network and local timings both drifted, sometimes very slowly and other times quickly resulting in the frame roll Mike described. Of course once things all went digital problems were solved.
The effect seen on the ECPs is not anything like the same issue. What you describe is a slight "hiccup"... generally of around 4 frames which occurs as the DVE equipment is put in circuit. Effectively you get a repeat of 4 frames. Mixing on the DVE helps hide this but not completely. That said most people won't notice it.
There's no easy way to avoid it - if you put the DVE in at the top of the programme you lose the ability to edit or preview the ECP. In the case of a film that could be 2 hours.
It's not about network timings, but about circuitry timings - this is something that happens millions of times a second - quite a bit more than twice a day! If vision circuits aren't exactly in phase to a roughly 12.5 MHz frequency, a phase shift will occur - anything less than one scan line will cause only a horizontal shift, and anything more with cause a vertical (and sometimes horizontal aswell) shift.
The manifestation of a lack of timebase is different on analogue and digital, sure, in terms of either an analogue wave or a digital image description being delayed, but in both cases it's about vision signals becoming interrupted mid-field and therefore corrupting the overall frequency of the frame.
As has been stated, with DVE equipment you'll have a frame or two's lag, because the DVE will 'buffer' a frame or two as it processes the items to go onto the frame, or the effects to apply to it. However, this doesn't account for the timing issues originally described, which seems to suggest that dodgy routing is to blame, whereby lots more digital processing is occurring on the DVE input than that occurring on the live source (ie. the playout server), as even a partial-field lag on a digital signal will cause the receiving device to buffer it to yet another frame later - or attempt to render the data coming in on the current frame - causing a 'roll'-like effect.
It's not about network timings, but about circuitry timings - this is something that happens millions of times a second - quite a bit more than twice a day! If vision circuits aren't exactly in phase to a roughly 12.5 MHz frequency, a phase shift will occur - anything less than one scan line will cause only a horizontal shift, and anything more with cause a vertical (and sometimes horizontal aswell) shift.
Erm yes, but I was trying to simplify things for the less technically aware!
Erm yes, but equally I was pointing out that checking a station's time against network's twice a day is hardly the same thing as vision signal timing. Which it isn't!