I hadn't been able to find out any info on this before I decided to make my purchase but now that it's here I'm a little taken aback about TIR's performance. Maybe it's just me not tweaking it enough but I'm starting to fear it's more a limitation of the hardware that causes the mouse lag.
Don't get me wrong, a bit of mouse lag isn't too bad for looking around out of a cockpit but I was really hoping to be able to use the TIR for FPS games as well. Yes, it is possible to play them but there is a serious loss of precision when having to compensate for cursor lag.
Is this because of the processing and filtering being done by TIR software or does the hardware not refresh or poll the cursor enough?
I have the TIR set up well enough to track just fine in case someone is inclined to reply about speed settings but is there something else I'm missing? I don't use a VIA chipset.
Regards
Charvel
Cursor lag...hardware, software or user?
Re: Cursor lag...hardware, software or user?
Hello:
Mouse lag is a very subjective thing, I would like to first just establish that the unit is working in a baseline way, ie. not performing badly. If you move the cursor around in a normal fasion, not in an FPS game, does it seem reasonable? When you adjust the Smoothing slider does this effect the lag? With the smoothing slider set to minimum, does it react faster?
If it is working fine, which it sounds like it is, then there are several things to do to help the performance. First, we have a new version of our program coming out in a week or so, it should be much better for FPS use, we will have optimized the smoothing algorithm for game play. Second, put the dot on the end of a hat brim, this makes for a larger dot movement and therefore you can lower your scaling value and then the smoothing value, which makes it more responsive. Also, increasing the size of the dot makes for a larger area to be tracked and allows for decreasing the smoothing.
Let me know if any of this helps, or ideas from you on how to make the unit work beter for FPS games. Perhaps we can add you to our beta group, with an eye toward FPS as we only have flight sim guys in there now.
Mouse lag is a very subjective thing, I would like to first just establish that the unit is working in a baseline way, ie. not performing badly. If you move the cursor around in a normal fasion, not in an FPS game, does it seem reasonable? When you adjust the Smoothing slider does this effect the lag? With the smoothing slider set to minimum, does it react faster?
If it is working fine, which it sounds like it is, then there are several things to do to help the performance. First, we have a new version of our program coming out in a week or so, it should be much better for FPS use, we will have optimized the smoothing algorithm for game play. Second, put the dot on the end of a hat brim, this makes for a larger dot movement and therefore you can lower your scaling value and then the smoothing value, which makes it more responsive. Also, increasing the size of the dot makes for a larger area to be tracked and allows for decreasing the smoothing.
Let me know if any of this helps, or ideas from you on how to make the unit work beter for FPS games. Perhaps we can add you to our beta group, with an eye toward FPS as we only have flight sim guys in there now.
Re: Cursor lag...hardware, software or user?
Hi Jim,
Thanks for your reply! I've tweaked around quite a bit more with the TIR and managed to find a reasonable setup for use with IL2.
The most important thing is to have the distance between the sensor and dots fairly close, which in my case necessitates a "boom" extension on a baseball cap with a 5-dot pattern (stuck on a guitar pick actually ). The tracking picture on the advanced screen was showing too short of a travel distance side to side before but with the reflector dots closer I can get a much better sweep allowing a lower sensitivity setting and less extrapolation of the movement allowing a better accuracy.
I have quite a bit of experience (whatever that really means ) in getting FPS games to respond well with mice. Sometimes it requires combinations of video card vsync settings with in-game or windows tweaks to get it just right.
The smoothing feature is fine to compensate for lower polling rates with flight sims but for FPS games even a very small lag from the smoothing algorithms won't be acceptable for serious gamers. I can't stress how important this is for FPS games where accuracy is paramount.
Most FPS'ers will even forego the USB mouse refresh (80 if I recall correctly) for 120 on PS/2 port with proggies like PS2rate but the difference really isn't that important. 40-60 is much too slow, however. Accelleration and speed settings will get you there quicker but sacrifices smoothness. I'm sure this isn't really news to you though.
I've read some of the features you guys are planning for version 2.0 drivers and I'm quite impressed with the options for deadzones and panning. Also, as you mention a better algorithm for smoothing looks promising but I still have some reservations about its use at all in FPS.
I'd be curious to know what the polling rate is on the TIR and whether there is the possibility of perhaps enhancing the hardware end to improve this. Don't get me wrong, this is an absolutely wonderful device and know that pretty soon I probably won't be able to live without it!
Incidently, I have given it a quick go in UT2003 and see some promise besides the accuracy issue. It may be some time before TIR is ready for competetive play but it's pretty close to being good for single player FPS use especially if developers can build in support to separate head movement from forward movement. Hmmm, I should re-install Mechwarrior 4 as the TIR seems ideal for this game too...
Anyway, I don't mind beta testing if it will help open the doors to the FPS community. I am not so fortunate to have as much time as I used to but I'm sure I could give you some valuable feedback at least.
Again, thank you for your response. It's nice to know there are developers out there that take the time you guys do to interface with their customers to improve their products.
Thanks for your reply! I've tweaked around quite a bit more with the TIR and managed to find a reasonable setup for use with IL2.
The most important thing is to have the distance between the sensor and dots fairly close, which in my case necessitates a "boom" extension on a baseball cap with a 5-dot pattern (stuck on a guitar pick actually ). The tracking picture on the advanced screen was showing too short of a travel distance side to side before but with the reflector dots closer I can get a much better sweep allowing a lower sensitivity setting and less extrapolation of the movement allowing a better accuracy.
I have quite a bit of experience (whatever that really means ) in getting FPS games to respond well with mice. Sometimes it requires combinations of video card vsync settings with in-game or windows tweaks to get it just right.
The smoothing feature is fine to compensate for lower polling rates with flight sims but for FPS games even a very small lag from the smoothing algorithms won't be acceptable for serious gamers. I can't stress how important this is for FPS games where accuracy is paramount.
Most FPS'ers will even forego the USB mouse refresh (80 if I recall correctly) for 120 on PS/2 port with proggies like PS2rate but the difference really isn't that important. 40-60 is much too slow, however. Accelleration and speed settings will get you there quicker but sacrifices smoothness. I'm sure this isn't really news to you though.
I've read some of the features you guys are planning for version 2.0 drivers and I'm quite impressed with the options for deadzones and panning. Also, as you mention a better algorithm for smoothing looks promising but I still have some reservations about its use at all in FPS.
I'd be curious to know what the polling rate is on the TIR and whether there is the possibility of perhaps enhancing the hardware end to improve this. Don't get me wrong, this is an absolutely wonderful device and know that pretty soon I probably won't be able to live without it!
Incidently, I have given it a quick go in UT2003 and see some promise besides the accuracy issue. It may be some time before TIR is ready for competetive play but it's pretty close to being good for single player FPS use especially if developers can build in support to separate head movement from forward movement. Hmmm, I should re-install Mechwarrior 4 as the TIR seems ideal for this game too...
Anyway, I don't mind beta testing if it will help open the doors to the FPS community. I am not so fortunate to have as much time as I used to but I'm sure I could give you some valuable feedback at least.
Again, thank you for your response. It's nice to know there are developers out there that take the time you guys do to interface with their customers to improve their products.
Re: Cursor lag...hardware, software or user?
Hello:
Thanks for the great post. Really a lot of good information in there. It represents exactly how we feel about FPS; it is the future and we are okay now, but have a ways to go for the ideal feature set. I think that 2.0 will really help out, but nothing would be better than built in support.
You are dead on in your assessment of the accuracy and lag issues with regards to FPS. We currently poll at 60Hz, which would be fine except that the resolution is too low, so we smooth the data. Your hat solution is very good for increasing the accuracy, as you move the dot over a wider range and therefore have more data. We can increase the framerate, but are also working on getting more resolution out of the system. More of both is better!
The idea of how to use our system in FPS is very interesting to me, and perhaps the biggest question. In flight sims it seems to present itself obviously, you are in a cockpit with an obvious forward view and need to look around, we are the perfect solution. In a FPS the world has no "forward view" and therefore is totally relative to your current position. What it does have, that is different from sims, is the ability to move separate from aiming. This is the case today, aim with the mouse and move with the keyboard. Are we good for aiming? Maybe. With enough resolution then we can be like a laser pointer coming out of your head, but do you want that? I would prefer eye tracking for aiming (don't get me started on that please). Obviously we can't change users current interface dynamic, keyboard and mouse, so we have to augment it and make it better. Is this where we separate the view of the user from his aiming and moving? I can do this in the real world, shooting from behind, but is this better in a computer world or too complex, is it a functional improvement. One positive point is that I would then have a "forward view", meaning my view is lined up with my body, and our viewing experience becomes more meaningful. Also, the precision factor goes down slightly, because you are still using the mouse to aim. Of course, how this would really play out in a game I don't know, but I do think it would be more relevant to slower paced games like counterstrike and doom III. Perhaps if you could move your head side to side as well, to look around corners without exposing your body.
I almost hesitate to promote or hint that our product can be used at all in FPS games, as I only want users to have the best experience possible, and tell all their friends about it. In this case I believe that all or nothing is the way to go, unless there really is a compelling function that we can provide in existing mouse view FPS. Hopefully feedback from 2.0 will give us some more direction on this, without trying to sell it as the ideal FPS aiming solution. We all know that is eye tracking anyway, right?
Regardless, as from the lessons in the flight sim community, we have to provide a better, enhanced, experience over the existing interface method. I know it is out there and welcome all ideas on how to make it work.
What we really need is some developer who is slightly crazy to work with us on the ideal implementation, so we know the answers up front!
Just realized I got a little carried away on this post!
If you would like to be part of the beta group, please email Halstead York at halstead@naturalpoint.com
We will have a new beta out around Friday that will have new autopanning code, would love to have you try it out.
Thanks for the great post. Really a lot of good information in there. It represents exactly how we feel about FPS; it is the future and we are okay now, but have a ways to go for the ideal feature set. I think that 2.0 will really help out, but nothing would be better than built in support.
You are dead on in your assessment of the accuracy and lag issues with regards to FPS. We currently poll at 60Hz, which would be fine except that the resolution is too low, so we smooth the data. Your hat solution is very good for increasing the accuracy, as you move the dot over a wider range and therefore have more data. We can increase the framerate, but are also working on getting more resolution out of the system. More of both is better!
The idea of how to use our system in FPS is very interesting to me, and perhaps the biggest question. In flight sims it seems to present itself obviously, you are in a cockpit with an obvious forward view and need to look around, we are the perfect solution. In a FPS the world has no "forward view" and therefore is totally relative to your current position. What it does have, that is different from sims, is the ability to move separate from aiming. This is the case today, aim with the mouse and move with the keyboard. Are we good for aiming? Maybe. With enough resolution then we can be like a laser pointer coming out of your head, but do you want that? I would prefer eye tracking for aiming (don't get me started on that please). Obviously we can't change users current interface dynamic, keyboard and mouse, so we have to augment it and make it better. Is this where we separate the view of the user from his aiming and moving? I can do this in the real world, shooting from behind, but is this better in a computer world or too complex, is it a functional improvement. One positive point is that I would then have a "forward view", meaning my view is lined up with my body, and our viewing experience becomes more meaningful. Also, the precision factor goes down slightly, because you are still using the mouse to aim. Of course, how this would really play out in a game I don't know, but I do think it would be more relevant to slower paced games like counterstrike and doom III. Perhaps if you could move your head side to side as well, to look around corners without exposing your body.
I almost hesitate to promote or hint that our product can be used at all in FPS games, as I only want users to have the best experience possible, and tell all their friends about it. In this case I believe that all or nothing is the way to go, unless there really is a compelling function that we can provide in existing mouse view FPS. Hopefully feedback from 2.0 will give us some more direction on this, without trying to sell it as the ideal FPS aiming solution. We all know that is eye tracking anyway, right?
Regardless, as from the lessons in the flight sim community, we have to provide a better, enhanced, experience over the existing interface method. I know it is out there and welcome all ideas on how to make it work.
What we really need is some developer who is slightly crazy to work with us on the ideal implementation, so we know the answers up front!
Just realized I got a little carried away on this post!
If you would like to be part of the beta group, please email Halstead York at halstead@naturalpoint.com
We will have a new beta out around Friday that will have new autopanning code, would love to have you try it out.
Re: Cursor lag...hardware, software or user?
Hi Jim,
I sent of an email to Halstead and we'll take it from there...
I've put a little bit of thought into implementing TIR into FPS games (although haven't totally wrapped my head around all the possibilities yet) and have to agree with you that it will be difficult if not impossible to replace the mouse with a TIR unit as it stands now irregardless of accuracy issues which are technical and ultimately solveable. However, it could be very useful in its current implementation as an "assistive" device if support were added by the games developer of course.
Play would use the keyboard and mouse as normal but the TIR could be activated with the pause/unpause command (like most people seem to use anyway in flight sims) to "unlock" the players view from their movement direction. In fact using strafe keys do this linearly already, which enables circle strafing but when the player decides to evade or run forward it could be very useful to disjoin the neck so to speak and have a look around (situational awareness) while still moving in the desired direction, with the press of a key.
TIR can't replace the mouse itself anyway (unless there is something I'm not thinking of ) because the mouse needs to be physically picked up and moved on the mouse-pad to circle strafe.
On another note, I have to say I'm looking forward to acceleration features in the new drivers for flight sims. After some tweaking I've managed to get a decent enough setting that lets me check "six" but to do so looses much of the forward view precision which necessitates pausing TIR. Also, because of the lack of precision it makes it twice as difficult to get an idea of the speed and direction of a relatively far off plane with it on. It's almost akin to halving one's framerate which isn't preferable by any means.
60hz polling is only part of the problem since it is also necessary to extrapolate movement from the smaller input data range of side-to-side head movement as you noted in your post. I wonder if its an inherent limitation of IR or if there is enough room for improvement to make up for it, say 80-120 hz effectively? Maybe wireless radio wave or laser is more up to the job? Anyway, I'm digressing here, sorry...
I won't get into eye-tracking too much but I personally have difficulty seeing it (no pun intended ) as being practical for FPS targeting in the long run for some of the same reasons like lag and scaling but more importantly if it would be simply too accurate if it could be made to work properly! I know that sounds crazy but if you take out manual coordination (skill) from a game by making it too easy (shooting at something just by looking at it) then it can take away much of what is attractive to dedicated players. This, I know is debateable and depends on the structure of the game itself to some extent but it really does depend on how it's implemented in future games.
Ok, I'll stop running on now. Thanks again for your responses and look forward to seeing the upcoming improvements!
Regards,
Charvel
I sent of an email to Halstead and we'll take it from there...
I've put a little bit of thought into implementing TIR into FPS games (although haven't totally wrapped my head around all the possibilities yet) and have to agree with you that it will be difficult if not impossible to replace the mouse with a TIR unit as it stands now irregardless of accuracy issues which are technical and ultimately solveable. However, it could be very useful in its current implementation as an "assistive" device if support were added by the games developer of course.
Play would use the keyboard and mouse as normal but the TIR could be activated with the pause/unpause command (like most people seem to use anyway in flight sims) to "unlock" the players view from their movement direction. In fact using strafe keys do this linearly already, which enables circle strafing but when the player decides to evade or run forward it could be very useful to disjoin the neck so to speak and have a look around (situational awareness) while still moving in the desired direction, with the press of a key.
TIR can't replace the mouse itself anyway (unless there is something I'm not thinking of ) because the mouse needs to be physically picked up and moved on the mouse-pad to circle strafe.
On another note, I have to say I'm looking forward to acceleration features in the new drivers for flight sims. After some tweaking I've managed to get a decent enough setting that lets me check "six" but to do so looses much of the forward view precision which necessitates pausing TIR. Also, because of the lack of precision it makes it twice as difficult to get an idea of the speed and direction of a relatively far off plane with it on. It's almost akin to halving one's framerate which isn't preferable by any means.
60hz polling is only part of the problem since it is also necessary to extrapolate movement from the smaller input data range of side-to-side head movement as you noted in your post. I wonder if its an inherent limitation of IR or if there is enough room for improvement to make up for it, say 80-120 hz effectively? Maybe wireless radio wave or laser is more up to the job? Anyway, I'm digressing here, sorry...
I won't get into eye-tracking too much but I personally have difficulty seeing it (no pun intended ) as being practical for FPS targeting in the long run for some of the same reasons like lag and scaling but more importantly if it would be simply too accurate if it could be made to work properly! I know that sounds crazy but if you take out manual coordination (skill) from a game by making it too easy (shooting at something just by looking at it) then it can take away much of what is attractive to dedicated players. This, I know is debateable and depends on the structure of the game itself to some extent but it really does depend on how it's implemented in future games.
Ok, I'll stop running on now. Thanks again for your responses and look forward to seeing the upcoming improvements!
Regards,
Charvel
-
Charlie251
- Posts: 955
- Joined: Fri Jan 03, 2003 5:00 am
- Location: Melbourne, Florida
Re: Cursor lag...hardware, software or user?
Hmmm... IR reflective contacts for eye tracking? Wonder if I could get Acuview to make my perscription contacts with a little "reflextivity" added on the top.