Sorry, I didn't see this thread, just stumbled across the vid on you-tube.

If I was trying this, I think I'd give each racer an 'interesting' score, based upon (for example, but probably in this order of descending weighting):
> How close are they to the car in front?
> Are they currently being watched (and inversely, for how long)?
> Are they colliding?
> Are they off track?
> What is their position in the race?
> A 'Look - thats me!' lottery.
etc.
All cars are scanned each 'cycle' & the car with the highest score each cycle would be the one to watch.
It's a rather programmatical way to think of it, but should be easy to tune by weighting each category, rather than dealing with abstract things in code. I don't know whether this approach would work.
It did seem to spend quite a long time in it's 'hunt' mode sometimes.

Something else to consider, which does happen in real events, is that some parts of the track are more interesting than others. They are usually the ones where the cameras are located I suppose. It feels natural if the director switches cars between these sections of the track & not in the middle of them, or worse, just about to pass through one. Maybe there should be a track node weighting?

Quote from Becky Rose :

1) The choice of camera at the start

Quote from Becky Rose :

2) The camera focus' on 1 car, it's a classic sim racing mistake in every sim out there... But real TV does not do that, the camera typically tracks two cars and it is possible to do this in Insim as I did it for X-Cam. ( http://www.youtube.com/watch?v ... 5myGUc&feature=relmfu )

I see your point, and something I'd totally over looked. Hmmm...

Quote from legoflamb :

Seems like most of my input has already been suggested though.

Still appreciated though dude, thank you

Also thanks for your thoughts as well Kiyoshi It's appreciated - you're spot on with your scoring idea - Infact I've been working on something like that in the tv.js-priority branch on github - https://github.com/theangryang ... s-priority/plugins/tv2.js if you wanna check it out. It's far from complete

I will be honest the whole custom camera thing is something I want to work on next, which will fall into the whole track node weighting, but I've not got as much time to work on xi4n for the next few weeks Fecking work and family

Just wondering if you were still thinking about this awesome project or if you just passed to something else ? It looks very promising and a good solution to move from TV Director

Quote from the_angry_angel :

Also thanks for your thoughts as well Kiyoshi It's appreciated - you're spot on with your scoring idea

Kiyoshi's idea is very close to how I did it in X-Cam.

The heuristic I used gave each car a score based upon position, which was bell curved at the start and end of the race and straight in the middle. I then modified that score by how close they where to nearby cars (position, not on track).

A second score was calculated for throwing yellows and off track incidents, This score was multiplied if other cars in the same sector also threw yellows or travelled off speed. There wasnt collision data back then, so I didnt have that to use. This temporary score was balanced against the first to decide if a temporary "action" switch was worthwhile, and to a large extent was the result of trial and error.

I decided to make this second score totally independent of the position/battle score because I designed it for league racing so incidents where rare-ish and more important. A public pickup racing camera system likely would combine the scores.

Overides existed for the first and last lap. As its important in race coverage to show the winner crossing the line, and other cars need a strong weighting as they cross the line.

Ideally I would have liked my system to thrown in a few artistic cameras and helicopter shots when the interest scores where not that high, but I never got that far.

The other thing I really wanted to do with X-Cam was to allow it to connect to 2 copies of LFS, one being 10 seconds or so behind the first so that it could read ahead (you can do this with live races too by opening the reply file after the race has started for playback, and leave the master copy reading off the live stream). This would allow better prediction of where the camera should be - which in most sim racing broadcasts would be a huge advantage as replays cause that visual latency especially as they grow in size.

In my dreams a second computer could then have been used to bring in replays with a video A/B switch to the encoding machine, and a control interface with each incident as a button - but I didnt have anywhere near enough hardware to do that.

Quote :

I will be honest the whole custom camera thing is something I want to work on next, which will fall into the whole track node weighting, but I've not got as much time to work on xi4n for the next few weeks Fecking work and family

This is what will bring the footage alive. Once you have the custom camera's setup and tracking multiple cars the difference in final quality is immense and you can do away completely with the standard cameras.

My method involved aggregating the "position" of the car viewed with other cars it was close too. You need to give a read ahead prediction which is relatively easy to sort out and the smoothing is just a matter of visually adjusting it until the effect is right. It's pretty simple once it clicks, and the results are stunning.

EDIT: Oh yes, you also need a mechanism to make sure the camera doesnt switch too fast. I can't remember what I did for that, I think I did a lock for 2 seconds or something, and artificially inflated the car's interest score when the camera switched, then degraded it back to 0 over time.

idk if this has been mentioned, similiar to phils idea. If its used in a league that may have pace laps, maybe have a way to select that so it switches to the "race start" cameras that watch the entire field?