To comment on a couple of comments made by Gavin about broadcasters using funny terms to confuse people, I thought as a good broadcast engineer I might enlighten those who are interested.
The term 'constellation' refers to the pattern created by the phase diagram of a phase modulated carrier. Obviously as engineers we don't actually spend lots of time drawing, and its a shame I don't have a scanner here to copy in my BBC study sheets but I just found this set of slides from Cambridge university: http://www-lce.eng.cam.ac.uk/~mrdr3/MPC.pdf Being able to see the constellation is an option that can be used as a great diagnostic tool. When demodulating the signal you are left with two signals called I and Q, and by putting these into an oscilloscope and applying one as the X and the other as Y you actually get the phase constellation displayed and you can see some kinds of distortion. n-PSK transmissions have no amplitude component and are less affected by non-linear distortion of carriers caused if an amplifier is over driven. But QAM carriers are particularly susceptable to non-linear amplification, this is because they vary amplitude as well as phase and you need to have your entire transmission chain well balanced. Distortion of QAM can cause 'decision' problems where the decoder can't decide of a symbol belongs in one state or another. Thanks for saying my job's interesting Gavin, its not, but thanks for saying it sounds that way. :P Do you feel like coming in and doing a few shifts for me? Its one of those jobs where for the past few days I've done virtually no work to speak of but I know that at any moment I could have to deal with a major incident. Combination of an overpaid TV repairman and security guard. :) I've been rather inspired this weekend, plus a dose of boredom, and I've setup a website for discussing broadcasting related issues. It has very little, but I hope soon to populate it with interesting articles, QAs and discussions. So feel free to make use of it, be aware that its a little under-construction at the moment: http://www.orbit.me.uk/ Bob_ -- Is realities bitrate is too low? ----- Message from [EMAIL PROTECTED] --------- Date: Sat, 15 May 2004 12:15:29 +0100 From: Gavin Hamill <[EMAIL PROTECTED]> Reply-To: Gavin Hamill <[EMAIL PROTECTED]> Subject: [linux-dvb] Re: [OT] Reality To: [EMAIL PROTECTED] On Saturday 15 May 2004 05:24, Jerico Webmail wrote: >I am seeing > more newsgathering people using MPEG-4 and interestingly we will be testing > MPEG-4 multicasted over IP encapsulated in DVB very soon. If demand for > HDTV grows the money people may demand MPEG-4 to save bandwidth but more > likely just higher modulation schemes (8PSK). Apart from the sitting-in-a-dark-room-for-half-a-day part, your job sounds interesting :) All the technical aspect of television without having to worry about the quality of the programming =) > I'm also interested to see how many broadcasters make the transition to > 8PSK over satellite for DTH. I didn't even think there were any other options than QPSK - I stand corrected :) > I must say, we've tried using 16QAM over > satellite (for DVB-IP) and even with a 2.4m antenna and stressing the > satellite transponder the reliability is quite poor. I wouldn't hope for > 16QAM anytime soon over satellite, Yep, and some cablecos are really packing it in at 128QAM... I just love the phrase 'constellation density' - it sounds like it was designed by geeks, for geeks to the exclusion of all else =) > but 8PSK is more than ready for > deployment when the receivers can be deployed. Alot of outside broadcast > sporting events are using 8PSK and where a decent setup can be deployed and > demand is high the occasional 16QAM is used. Out of curiousity, what kind of additional bandwidth is 8PSK giving? > For a channel 15Mbps is very nice bandwidth, but if thats being sat muxed I'm quite sure the '15Mbps' is purely for show, since the AV7110 is unable to decode streams of more than 10Mbps, and this chip is used in various STBs not to mention our old friends, the 'full feature' PCI cards from Siemens/Fujitsu et al. (if I'm wrong here, please correct me) > then that could take a big hit. The biggest problem with DVB broadcasting > is that no one will pay for pre-processing of the media, <nod> I was involved with a VHS -> XviD project for a while, and the amount of VirtualDub filters we had to send the raw analogue capture through just to get a decent quality encode was quite mad. Even on a highly specced P4, processing 480x360 (I think..) data was running at one or two frames per second max. The output *was* worth it though - MPEG4 with very little background noise / artefects / movement, so all the bitrate was truly going to describing the action. > Snell & Wilcox provide some lovely processing hardware that claim to be > able to save significant bandwidth by preprocessing analogue noise out. I bet there's some insane FPGA and other custom DSP stuff going on in there, and a price tag to boot to be able to do that in realtime and at broadcast quality.. Cheers, Gavin. -- Scanned by The Sheriff - http://www.isheriff.com/ -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject. ----- End message from [EMAIL PROTECTED] ----- -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
