Genlock circuits

Some more info would be useful. Such as tell me more about the "network feeds". We used to genlock a bunch of stuff at a satellite news gathering company I worked for.

Bob

You want this sort of thing ?

Modify the search terms to find more instances.

Graham

the

Some more info would be useful. Such as tell me more about the "network feeds". We used to genlock a bunch of stuff at a satellite news gathering company I worked for.

This app is for a company that provides systems for CATV. They want to put a scrolling alert message out from one character generator onto multiple network feeds. Having designed commercial inserters, I understand cleanly inserting a video feed (like a commercial) using the vertical interval, or a digital buffer, but I am not familair with the details of overlaying video characters or pictures onto an existing video stream. That is why I am searching out information on genlocking circuits, ICs, and app notes.

The character generators are built for single feeds, and do use a genlock circuit, but that does not help much when trying to feed the same video output to more than one network channel at the same time.

Work is hard to come by, so even when I don't know it all, I still like to take it on, and work it out, whenever I can. :-)

Thanks,

John

the phrase you want is "frame synchronizer" or time base corrector"

but using one character generator and a bunch of TBCs may be more expenise then using a bunch of character generators each with a gen lock to it's network, one for each feed.

Mark

Thanks Mark,

Not at all in an argumentative way, but I don't think that is a solution. I am not going to claim lots of knowledge, and I am happy to defer to the more wise among us, such as yourself, but a TBC seems way overkill, and why would not some type of genlock setup do the job? Is it just something that you figure is not doable?

Is there just no way to use a "genlock" circuit on each network to sync up the single feed? I never meant to say I was sure this was the right thing to do, that is why I am here solicting input. If it cannot be done with one character generator into multiple streams, I would like to understand why, so I can try to explain it to the customer's satisfaction, if possible.

We have frame sync and TBC hardware and of course throwing money at it would make it all better, but like you say, way too costly. :-)

Much obliged,

John

Sorry.

" Networks" are TV networks, CNN, FOX, etc. The goal is to send crawling text from one character generator, out to multiple instances of TV network feeds. The timing for each network is going to be unique to each network feed.

The CG is capable of genlocking to one feed, so I am soliticing methods of an interface to allow its output to lock on to each network independently. Not saying it can be done, but just trying to figure it out in order to reply to the request.

It may well end up being a reason to use many cheap CGs with genlock, rather than trying to get by with one, or at least fewer, of the rather costly ones that the client already offers. Of course, considering the number of networks CATV delivers, its easy to see that a "one CG" solution would be preferred from many points of view. Hoping to find out if what they want is even reasonable.

Thanks,

John

you either need to use FS and TBC for each network

or

multiple CGs, one for each network..

Mark

is this for "emergency alert" applications

Mark

Primarily, yes.

John

destinations?

The source is the same. The number of destinations is actually limitless. So far, not every network needs to provide the alert information, but power granted is power abused, so sooner or later the FCC will force everyone to get on board with it.

and if the

Watch the weather channel. Much of that is their CG stuff. Its quality video.

when switching

That is why it needs to be genlocked, or TBC'd.

simple relay to "bypass" the entire feed if a fault is detected, but having a fault like this could drive a video tech bonkers. Sounds like a fun project. :)

All broadcast CATV hardware has to be outfitted with a failsafe bypass to return everything to network if there are problems. Of course, the best laid plans.......

logos, ads, channel numbers, etc. A surprising number of older folk can't read it, so if the text is important, consider making it big and high-contrast.

You will have to take that one up with the FCC, as I think they mandete the text size and other details when it comes to their alert system. FWIW, I agree I don't own a TV with the resolution to make out the scores during the basketball games, even with my glasses on

Thanks,

John

multiple CGs, one for each network..

:-) I am not ignoring your advice, just waiting for others to either agree or disagree. I am not sure either of those is going to be a good solution for my client, even tho they may well be the only ones available. Being cheap, they want to get away with using one of the high end CGs that they already build, rather than creating a new product line. Its their call, I just want to give them what they need to make it the right one, and make me some $$$ if I can

I appreciate the points of view and experience.

John

It seems to me that each feed could be at any point in the video frame unless the master transmitter/muxer was one unit with a TBC. So each "channel" could be

The problem with any design which "sits there waiting to inject text" is the device is active, inline with the signal path, 24/7. Any device (unless very high quality (and complex and costly)) will degrade video quality, and if the device fails, the signal could vanish or become corrupted - not a good thing. Active bypassing could be implemented, but you might get glitches when switching the CG Mux in and out. And some fault-tolerance is in order - perhaps even a simple relay to "bypass" the entire feed if a fault is detected, but having a fault like this could drive a video tech bonkers. Sounds like a fun project. :)

I'm by no means an expert, so take my ramblings with a grain of salt. :)

p.s. Many people dislike scrolling junk on the screen. Same with logos, ads, channel numbers, etc. A surprising number of older folk can't read it, so if the text is important, consider making it big and high-contrast.

It depends what you mean by "networks" Is it a computer network, or a whole pile of video destinations? What is the video reference for the multiple networks. If it is the same for all, you might be able to get away with analogue video delays. The CG, ir it is a reasonable one might be able to be timed into the system.

A bit more clarification would help

martin

Opinions are like assholes -- everyone has one

Ok, so all the I/C feeds are non sync. A Classic way would be to put a TBC/synchroniser on each feed, but it might be easier/cheaper to get a PC per video network that has a decent video codec and get appropriate software. I'm not in the video biz anymore, so I'd look at computer video grafix people, for an alternative approach.

PC's are so cheap, but they will Never (IMHO) be as reliable as a hardware TBC / synchroniser. A PC that crashes wont go to a bypass mode and you would loose your network feed, but a TBC probably has that facility, so you would loose the caption, but not the network

martin

Opinions are like assholes -- everyone has one

Interesting. The local news channel here has started putting "bulletins" for any [miniscule] "breaking news" on our TV. The text is TINY - about half the height of normal ticker font (and barely readable.) We've learned to tune it out, like commercials.

John, You are not going to be using genlocks for this. Let me see if I can explain it.

First, a genlock basically says that my new source is going to be locked to this other signal. So, your CG takes the incoming signal, syncs itself to that signal, adds your warning text to it, then outputs that combination. Fairly simple to do, and doesn't take that many resources from your CGs computers.

Now, you try to add a second source. You can't just genlock, because the two inputs are not sync'd to each other! Now, it is possible that they could be. In a broadcast studio all the sources are locked to a single master clock source, so that fades and transitions can be done easily and smoothly. If your CATV head-end went to the trouble to sync all your channels, then you could theoretically genlock to more than one channel.

Even if they were, you would still have to have two separate circuits in your CG, because the signal addition happens in the box, you are not just added your signal on top of the other. In other words, the CG generates a single video signal with both the original and the text together, it doesn't just output a text signal that is added to the original.

So, anyway you look at it, you need one CG for each channel. You can probably find multi-channel CGs on the market now for exactly the task you are now undertaking...

Charlie

Hello John,

I wonder how they want to do that with one CG. After all, a CG can nowadays be made cheaply, even with a plain old micro processor like an MSP430. Then you'd be able to have multiple "CG pods", each synchronizing to its respective feed but being fed the same character sequence via, say, RS232.

If you do the uC thing be aware of how they generate the clock. Some (like the MSP430) will dither and that won't be good for video applications. You'd have to make your own clock oscillator and that must be locked to H-sync of the feed. But there are chips to do that.

With PCs, even the rugged PC-104 architecture, I'd be careful unless the clock generation on the video portion is fully documented and understood. I have seen a lot of grief with PC-based video hardware. Actually, I had to redesign some of these "solutions" so they would meet their own specs. Surprisingly, the engineers at the manufacturer were often not even embarrassed about that. Seemed kind of "normal" to them.

The quality of a CG that adds text or other stuff to an exisiting feed is, in my experience, proportional to the quality of the sync separator and PLL that restores the timing info of the source and makes this timing info available to the CG. Creating a nice source is easy, extracting sync info from a feed is going to be the art here. Even more so in cases where the feed signal isn't quite 100%.

Regards, Joerg

Hello Charlie, Hello John,

Yes, and this would be one of the companies to talk to if you don't want to design your own CG chain:

Regards, Joerg

You'll at a minimum need effectively a separate frame buffer for each network you need to feed.

Graham