My sincear apologise if this has been posted before.
I know, or atleast I think, it is a completely impossible thing to do, but maybe it isnt.
Can we have video streaming of servers on websites, so when big time leagues, IE STCC are holding races, they can stream them onto their website and "Spectators" can watch it from their website.
Please dont slate me if it is impossible. If you dont ask, you dont find out. Because I am not a scripter, so I'm not sure if it possible.
You can stream webcams on websites...It might work for LFS.
what do you mean? AFAIK STCC is either a download or a youtube video (well if it was live no-one told me, and im a driver!), live streaming could be possible with some WMEncoder setup, but im not sure how it would "edit" it (ie switch between cars for crashes/overtaking etc)
Then, you can have live race action, live race commentary etc.
You can have a race controler (or a better name :schwitz that controls where the camaras point, can show replays of crashes if the camara was looking at someone else?
So, it is possible?
Be nice to see what a programer or a big league admin says
I'm no expert on this stuff but I'll spam some more anyway
All it (the MoE broadcast) was was a session of the race being broadcast as it is in LFS, with no extra trickery (if I recall right). I suppose theoretically a "director's" version of something like noobTV and LFS could be more flexible but replays would be very difficult simply because of their nature (you can't rewind them).
Bandwidth considerations dictate that streaming mpeg style content is extremely heavy on bandwidth. If anyone were to broadcast a race's-worth of video to a "mass" audience, in a live event, the bandwidth required would be colossal. It might work if there were a relay system, where a server had say 6 concurrent streams, and one of those was given over to relaying to the next server in the chain, streaming 6 concurrent streams, relaying one of them to the next server in the....
The superior concept is LFSTV, which has various interpretations, but the one I like most is a dedicated replaying/spectating client which receives InSim data from the race, and also camera position/focus information from the LFSTV "director". An audio/commentary stream, which is much less bandwidth-intense than video, completes the service. Because it's not transmitting video, but instead just positional/directional/angular/etc data, it's much less burden on bandwidth. The LFSTV resolution is of course whatever you make your client window size. If needed, a relay system/service can also be incorporated, greatly increasing the number of possible viewers.
There's many different permutations.. those are just the aspects I like the most, from the proposals/suggestions I've picked up on. The above thread is well worth a good read
IPv4 and IPv6 have such a relaying system (well a better actually), it's called multicast (unlike the wrong definition of unicast in the thread you mentioned). Every 'viewer' is given a special IP address from a range of addresses that are grouped together into another single IP address. Each router in between the two endpoints (viewer, and broadcaster) creates a multicast group and combines the IP addresses into a single one at the other end of its connection. The result is that the broadcaster needs to send his stream to only 1 IP address. This single IP address is converted into more than one IP address at the first point of convergence (ie, where multiple routes to 'viewers' occur) and the stream is sent to each of those new multicast IPs only once until it reaches the next point of convergence, where they get split up again, etc, etc, until a packet reaches a 'viewer'.
The difference between multicast and regular (unicast) connections is that a unicast connection is 1:1, ie. it sends each packet to one single destination, and multicast is 1:M, ie. it sends each packet to an infinite number of destinations. Broadcast would be overkill since it would send each packet to all possible destinations (and for very obvious reasons is not routed at all (try sending a packet to IP 255.255.255.255 every 0.000000001 seconds and you'll kill the internet).
Multicast was disabled for security reasons in the early years of the internet, but with the arrival of IPv6 it's getting turned back on on all newer equipment by default.
Not superior... Completely different... It's a good concept, but it is limited to supplying people that use special software (this viewer), which is basically my only point of criticism on this concept.
Yeah, and possibly worth reviving (or we could just continue here). If we could only get some word of some dev (not naming names, Scawen, honest ) on his views on the matter.
Let's carry it on here.. people who want to back-track can follow the link to the other conversation. If that causes the other thread to get re-ignited, I might use my "special magical powers" to merge the threads into one later
Multicast sounds yummy. If I understand your explanation right, it solves the problem with the internet that makes video transmission so difficult in comparison with a TV transmitter. That'll break the back of that problem for sure! Yeeharr!
I don't think Scawen's ever given any positive or negative indications with regard to LFSTV directly. I remember seeing a post made by someone I've always considered to be "in the know", "finger on the pulse" etc, which suggested that it was a future consideration.. but I've no idea if it ever came to him from Scawen, or it was his own best guess. I also don't know if it's an S2 or S3 "thang", if it's a "thang" at all.
Yep.. it's a fair criticism. But consider these points...
The graphics would be superior if all the textures, tracks and car models were local to the player. You could run LFSTV at 1600x1200 resolution, and it would be scrumptious quality.. better than HDTV progressive scan. If your monitor could do it, you could beat interlaced scan resolution.
An hour-long race would be a pretty big download, regardless of the codec used to compress the image.. to download a player, and then stream the race data would likely be a much smaller download, even for a one-off view. The disadvantage is the preparatory download, to get the viewer. There's something really rather kick-ass about clicking and streaming instantly. Once you've got the viewer though, bandwidth usage is extremely economical.
Even if it does cost...Maybe you can sell it to leagues....£10 a year.
20 Leagues @ £10 pa = £200.
This can also be the little seed of the oak tree that will REALLY put LFS on the map, and I mean on the map...as famous at BTCC, WTC etc.
You have it watched over the internet, and say...we get 32,000 viewers per race (random number that is reasable) one driver then e-mail a company, lets say www.blahdvd.com says, we get 32,000 viewers a race, if I plaster my car in BlahDVD.com you pay me say..£100 a race?
They agree, a few other drivers do it, earn cash.
Then a team e-mails a company, say FedEx, says, we enter 4 cars into a championship, we get 32,000 viewers a race, over 10 races etc etc. you pay us £500 a race, and we put FedEx over all our cars. Then teams have enough money to buy drivers from other teams and before you know it you ahve your own mini-F1 series. Obviously, this is a long shot...but iy could happen
Yes, that was the whole intention of multicast back in the day they feared their big 2 inch thick copper cables would get stuffed with (2 inch thick) bits and bytes that got sent to multiple PCs using unicast.
The one problem with multicast now is that it will not work on local endpoints at customers... ie. your ISP still won't route multicast packets to the internet. Most of the rest is set to go when needed, but you'd still need an internet endpoint capable of broadcasting multicast streams (aka a streamserver). It was decided that, with the introduction of IPv6 and ever increasing bandwidth demands, multicast be phased back into general use on the internet. It may be a few years, but then, so will LFS be a few years until it's finished.
Yeah, I think that was way back on RSC that he posted something about a proxy system. I remember Scawen posting something about that subject as well.
Oh yes, I've considered those points, and I'm all for this application. Having one doesn't mean we can't have the other though. In fact... LFS-TV could be used to create a multicast videostream, so they can operate at the same time too.
Download sizes, however, are becoming less of a factor with the increasing bandwidth regular internet usage requires. Certainly the amount of time it takes to download a movie versus downloading a replay is hands down a win for the replay. That doesn't keep people from creating videos right this moment and hosting them somewhere for us to download. Ranging in size from 10 second examples of 1MB to full league races edited to have graphics (TV graphics) in them that are up to CD size, and we download them. Unless LFSTV becomes not only a proxy for the actual LFS stream of in-game data but also a fully working video-editing suite which allows you to create all of the fancy stuff Adobe Premiere and After Effects is able to create, regular video will always be in some kind of demand.
Which brings us to the other use... Live coverage. Needed bandwidth wise LFSTV will win that one easy. It has better graphics, uses the actual in-game data, etc. There is still one slight downside to it. You need the LFSTV player to watch, and you need to download all the textures used in the game (and possibly add-on textures, extra tracks if ever) to use it. Auto download will work but it will increase bandwidth slightly. Using videostreaming, one only has to input the url into Media Player and it starts playing. Using WMV streams reaches people not currently familliar with LFS, while LFSTV I fear will always be reserved for people who already know LFS. The amount of time it takes to watch the live coverage doesn't change between LFSTV and videostreams. An hour's worth of racing will take an hour to download in LFSTV and in Mediaplayer it will take just as long. The only difference will be in the amount of data you transferred in that time. For most people this will no longer be a major concern by the time either tool will be finished.
I just realised we agree... Could've saved me a lot of typing by just saying yes
At the moment (and at the state LFS currently is in) that scenario is very wishful thinking. However, it is a scenario we ultimately should be looking for. It won't ever be BTCC or WTC size, but big official leagues with a lot of money have been done before, and internet broadcast coverage has been huge (well over 50000 viewers per race for a $300000 NR2003S league). In fact, when you check on the iRacing business plans, that's exactly the kind of scenario they're building their simulator for. If LFS is ready when iRacing launches a less than perfect product then LFS could take over that niche of the market and become the ultimate simulator used for really big things...
I've always liked the idea of one of the clients (or the server) to have a relay. Connect to it is all you have to do and you get the streaming interpolations there as anyone else. In my rather limited understanding of this stuff I'd even hazard a guess that it's possible, and also "possible".
Yes, it would need a client with the textures, cars etc etc and it could end up being a whole LFS download all over again (so one could really include it in the LFS client) or it could be a small addon that uses the existing LFS installation's files. I can't see how it could be based on the streamed LFS data, be standalone and a small download because some things simply are required for it.
Looking at it from network traffic's POV it would still be cheaper ITO BW to download a 100 Mb LFSTV client (or the couple of megs for the viewer that uses existing LFS files) and then multicast/broadcast the LFS data instead of the actual video.
Just as LFS, the viewer would not need "installation" and if standalone it can work from a USB key or something similar, hence it can still be mobile while weighing a few Mb.
Like I said before, network demand (bandwidth) for live coverage is not an issue... Just look at the amount of data you are able to download per hour right now, and compare it to what you could download 3 to 5 years ago. Now step forward another 3 to 5 years (S3 should be finished sometime in the next 5 years, right?). Add to that the fact that internet TV is popping up allowing you to watch digital TV in full frames using the current available bandwidth and you'll understand why streaming an LFS video will not be a problem at all for most, if not all, potential viewers. Remember, you're mostly watching the race live, and as such won't even notice the amount of data you're sucking in. It will even allow you to do other stuff on your PC (while slowing down the internet connection a bit more), while LFSTV will seriously tax your PC demanding it be dedicated to viewing to a large degree, while leaving your internet connection running at near max speed.
As for downloading. Downloads have always been and will always be done using videostreams rather than in-game data, simply because video editing software gives you more artistic freedom than any LFS based application will ever be able to provide.
With regards the STCC I considered live streaming and ruled it out, I watched some of the GPV races and took notes. OK I couldn't understand very much because my Spanish isn't very good, just enough to go on holiday with, but the footage itself suffered from being live because it couldn't have enough spectator cameras logged in.
Really good live coverage could only happen if there where enough cameras to see all the action with a race director then just flipping between cameras.
To the point where even the first STCC edit suffers a little and for the next round I have written a race control application that will allow me to view up to 12 cameras simultaneously and flip between them on the output stream. I then remake the stream neatly afterwards.
The alternative that might work is to develop a piece of software that heuristically calculates what action to follow. I know there's a few video camera tools out there already i'm going to look at soon, and indeed I may look at doing my own too in the future.
