It's hard on LAN, I definitely had criticisms but under the circumstances they did perfectly fine

is it harder on lan than it is online?

i can appreciate its not easy but so many plays were missed, including super obvious spy/sniper plays

didnt know we had cheat feed in tf2, my bad

No cheat feed, high stress environment, more distractions, fatigue from working multiple 14 hour days in a row

You can use cheat sheet online because of the 90sec stv delay

Im not sure how possible this would be but if they could have a second observer but delayed by like 15 seconds. I know they had the 2nd observer for picture in picture and a third observer would be hard to do. If they did have this delayed observer they could use it to catch plays that the primary one missed then show it in picture in picture so we can at least see what the casters were going crazy about.

LAN has no cheat feed, so everything is in real time.

We were definitely tired today and missed some stuff, but I felt observing at the end of day 1 and on day 2 was some of the best observing weve some, especially on lan.

The picture-in-picture stuff helps a lot to cover more bases to try catch random deaths, along with the minimap. The synergy between the main and secondary (PIP) observer was really amazing to feel

That's a lot of fucking effort

why not use +-3 seconds stv delay, it is minor difference if you're looking to players screens and on the big screen simultaneously, would make it easier for observer and would improve the stream experience
anyway camerawork was really good given the circumstances

I agree, its not something that would be possible to achieve between now and next lan but i still think brainstorming an overall quality improvement is healthy for the scene.

You'd need at least 5 seconds for cheat feed, I tried online with varying delays on the feed and anything under 10 is difficult. Also anything but live requires a lot of infrastructure that you can otherwise skip without a delay.

Don't think it would help much, 3 seconds is way too short for any major benefits and even that kind of ruins the stage and player reactions a bit

Would it not be possible to set the STV to be at 15 seconds and have the actual observer in spectate. Or do some of the casting/observing plugins require you to be in STV.

I did mention it because I saw it in CS, and it is very effective there. They have setups where they have 10 pcs with one of each player which they can use for replays.

I understand that TF2 is far far smaller and has far less resources but if we could do the delayed feed that would be amazing and I think it would fix the issues of missing unpredictable sacks and airshots. You could even use a cheat feed with the delayed secondary observer so that they get everything.

Moist_PenguinWould it not be possible to set the STV to be at 15 seconds and have the actual observer in spectate. Or do some of the casting/observing plugins require you to be in STV.

I may be wrong but the plugins they use requires -insecure in the settings what means you won't be able to connect to any game servers.

dashners response was better

Would there be any way to have two versions of the game open, one where the observer can be in freecam and one where they are in first person?

Freecam is purely to allow the observer to see what's happening better.

For lan there were two instances for the split picture. It's possible but doubles the manpower needed for something that would probably just get ignored by the main observer. As I said above, the tools there were for lan were plenty good enough to catch things, it's just we slipped up a bit later on. Cheat feed would help but it's something that's been sacrificed to make other parts of the production much better while being more efficient for setup time

Yeah the 3rd day was rough... I think most of us were running on like 4 or 5 hours if sleep from the previous night? It also doesnt help that a lot of the tech was literally being developed at the event so there was some new stuff that most of us had to work into our usual workflow so there were some stumbles. Along with that we were working in observing pairs to run the tech and i dont think any of us are used to having other people do stuff for us so things get missed, or doubled up on because of assumptions.

I also just need to not start drinking @ noon :P

I'm following this thread with interest - I think I know what I needed to do better, but others might have a better idea from watching the stream.

only use split screen for the rollouts, and end it before they shoot their first sticky. with split screen, you can't really look at both perspectives at once, so the goal of that then becomes that the viewer has a choice who to watch, but it's pretty ugly and annoying to watch a half a screen. thus, only use it to show something a viewer would watch with their peripheral vision (only thing I can think of are rollouts). the small cam in the bottom left corner was cool, I'd recommend that as a direct replacement of the times you wanted to split screen.

you should spectate singular players in stalemates for longer periods of times to grasp what a player is trying to initiate, and generally, it should be the attacking team/players to spectate. for the most part, stalemates are only boring to watch because the casters and cameraman sort of miss what the attacking team is working out, and only really notice the actual play.

if a player is behind, spec that player for the entire time unless/until he's so far away from the enemy team that the enemy team can just cap the next point, then once/if he gets closer, spec the player behind again.

I generally agree with what you say, just a few things (and I promise that even though some of them look like leading questions, they are genuine things that I'm thinking about):

• if you saw the use of split screen to show the product forward hold, what did you think about that? Would you have preferred that as PiP?
  
• some related questions: do you think that you might want different things from an observer than a less experienced player would? Since you understand what the whole team is likely to be doing, do you think that makes you want to watch individual players more? I've heard several times that tf2 is confusing to spectate, do you think that spectating/casting single players more in stalemates would make it more or less so?
  
• how do you think we should balance PiP and main camera for players behind? I've found that often when I spectate the player behind things like uber exchanges happen offscreen - do you think it's worth it?

Cheers,
Sim

When it comes to games being confusing for viewers, I think a minimap in place of the PiP (when it's not being used) would help a lot in that area. Viewers would be able to see the "flow" of the game, on top of following the individual plays captured by the observer.

1) In this situation, I feel like split screen would be beneficial, but it would still be quite annoying to watch. Personally, I think a really smooth bird's eye person view as a team is setting up that forward hold would look a lot better. The moment the teams are able to exchange damage, then switch to 1st person (probably the attacking soldier or scout)

2) I think my suggestions apply to all observers, so yeah for example let's say it's a stalemate froyo vs ascent on snakewater (this applies to every single team; however, if you have experience you might know who is going to go for plays more than other players). froyo has control of mid on even ubers. It's possible that other players can make a play than blaze or with blaze (there are more thing to do than just suicide blaze), but for this particular map/point, you should probably just keep your eyes on blaze, seeing how he sets up before he actually bombs, not just switching to him the moment he's in the air. say it's same map, even ubers again, but froyo has control of our 2nd and trying to get last. you're going to want to keep your eyes on arekk and/or blaze since it's likely they'll both suicide. knowing whether or not they'll suicide 1 or 2 on last generally depends on the map, the score, how much time is left, the particular position of the teams, etc. but generally on last points, you'll see 2 people suicide. on other points, you'll USUALLY see only 1 (really depends, so it just comes down to experience, knowing what to expect).

3) is PiP the small screen in the bottom left corner? unfortunately i feel like this just comes down to experience, knowing when the action will happen first. however, when a player is behind and teams exchange ubers, you're going to see some interesting stuff from the player behind, so you should probably put him on the full screen and the same team's pocket on the small screen in the corner. so again, if he's too far away to really do any damage, you should leave the player on the small screen in the corner, but then switch to somebody involved in that uber exchange / resulting team fight. if the player (or any player) is backcapping, you should switch to his cam.

i feel like locked 3rd person camera should be used sometimes. it lets you see more of the surroundings, and looks nicer i think because TF2 does not show mouse movements of players very well when spectating them

honestly listen to corsa for most of this. Players behind almost always get ignored by the camera and if the layer behind is good, they usually change an entire push in a major way.

This applies to casters too. I feel like tf2 casing isn't SUPER developed when it comes to the roles that can apply. This results in some weird stuff where there is major action that doesn't get talked about. If you're co-caster is following and talking about the uber exchange, you as a caster don't need to care about that uber exchange. You can hear what's happening with it. Keep your eyes out for the soldier pushing the flank instead and see if their play develops.

A fixed 3rd person view is really jarring in tf2 because soldiers and medics especially do a lot of quick 180s and 90 degree flicks to rocket jump and flash players. I think it would be interesting to see what a 3rd person cam of a scout leading an uber would show, could give a good perspective to the viewers as to how the uber is going. Generally scouts being the leads in ubers is pretty heavily telegraphed (to the team pushing) and that may make it easier for a cameraman to switch to that view

there was actually a moment exactly like corsa's #2 where we got to watch drackk set up and then skip out of window on snake and it was definitely cooler than usual. as in, i thought to myself "great camerawork to take us through the entire sac play."

Props to Corsa for the thoughtful feedback

Players behind is by far the biggest challenge I had when doing camera work. Especially with how much first person is used it's really hard to tell when you should be on their camera, especially if the team with the player behind decides to start pushing due to the other team having people looking for that player. (jesus that sentance reads like word vomit i hope it make sense)

There's a reason I said this when free quit

I'm 100% out of my depth in regards to eSports coverage but here are some things/questions I have:

- What was the i55 set up? Cause I thought that coverage was great (so many headshots, airshots, spy plays, med picks etc all caught on the stream) and could that system be used again?
- I liked the split screen for rollouts but agree that it had sort of confusing impact in other situations
- I liked the mini camera in the bottom left to keep an eye on off-classes while the 'main' action was being watched BUT for example when a sniper was being watched on the minicam it was very difficult to see what was being aimed at since the sniper POV is so zoomed out. In these situations I think it would have been better for the minicam and maincam to be switched, otherwise it was difficult to see anything from the sniper POV and it was therefore almost redundant
- All of the rest of the production was great! I loved the pick/ban screens and the transitions were always on point as far as I saw and Nuze's analytics were a pleasure to watch
- Lastly I remember at Rewind there was at least one (maybe two) cameras with a live feed of the venue so that you could see the crowd in relation to the players on the stage and see the whole set up of the place. Really made it feel a lot more like you were there rather than just seein the crowd. For example I didn't even see the stage until I saw it in the i61 pictures thread. Maybe I just missed it though bc I didn't manage to catch 100% of the stream.

Hopefully this is helpful or at least nice to hear. Its also all just my personal opinion and I'm aware there is so much work involved with production and I have the utmost respect for everyone involved and thank you for your effort. Also if anything I said is out of your hands due to different productions companies(?) e.g. the tf.tv style vs the essentials.tf style then that's fair enough, I just wanted to suggest anything I could based on any past LANs I've watched. Lastly, sorry for rambling and cu@Rewind2018 :D

The main thing for us observers to massively improve any future event: being able to talk to each other while still hearing the game & casters. This LAN we did not get talkback between the observers until the final maps of the final day.

Having main & secondary observer being able to communicate while listening to the casters will help a lot! Also being able to interpret the map overview we now have better, besides sticking to the pure PoV camming should help with sac plays and players behind.