DJI FPV Simulator for all!

[Kilrah]

So you mean that what both Liftoff & Velocidrone says in their instructions is incorrect & what I feel & see when using Velocidrone and compare a frame rate above 200 & one down to 60 is totally wrong?

No, I'm saying that what I described had nothing to do with it. Maybe they use the same term differently.

For the record, the live view from the aircraft is also 60fps and is not floaty at all.
Analog FPV quad setups run at 60 max as well.

Also I've been flying liftoff at 60fps with no comparable issue, and when I fly FS2020 in VR even at 30fps there is no such feel either.

If the sim is floaty at 60fps then it's a result of how the simulation is coded, not the fps, or at least not to such an extent as we have here.

 

[slup]

No, I'm saying that what I described had nothing to do with it. Maybe they use the same term differently.

For the record, the live view from the aircraft is also 60fps and is not floaty at all.
Analog FPV quad setups run at 60 max as well.

Also I've been flying liftoff at 60fps with no comparable issue, and when I fly FS2020 in VR even at 30fps there is no such feel either.

If the sim is floaty at 60fps then it's a result of how the simulation is coded, not the fps, or at least not to such an extent as we have here.

You can't compare the fps from the goggles with a simulator ... the goggles records the reality, the real gravity will not be less with lower goggle frame rate, the quad will fall equal fast with or without goggles.

A software simulator on the other hand, generates a new picture as fast as the frame rate ... the simulator quad will not fall in between the frames as a quad in reality.

And for the record ... I also have the DJI FPV goggles & have used them for quite a while, 60fps there does not equal to 60fps in the simulator... if that depends on some other way of coding that you claim, is way over my head as I'm not a game creator ... but I know how to dial in a simulator so flying there feels exactly how it feels to fly my own quad in reality.

 

[Kilrah]

A software simulator on the other hand, generates a new picture as fast as the frame rate ... the simulator quad will not fall in between the frames as a quad in reality.

But it should, the whole point is that if it doesn't it's not coded right. The physics must not be tied to the frames but to time, and that's how most work.

 

[slup]

But it should, the whole point is that if it doesn't it's not coded right. The physics must not be tied to the frames but to time, and that's how most work.

Don't know what you mean by "that's how most work" ... neither of the largest most famous simulators work like that, keeping a high frame rate is crucial in order to not get a floaty feeling, the higher the better. Meaning if you have slow hardware you need to dial down the graphical load to keep the fps/refresh rate up.

 

[maelstrom]

Don't know what you mean by "that's how most work" ... neither of the largest most famous simulators work like that, keeping a high frame rate is crucial in order to not get a floaty feeling, the higher the better. Meaning if you have slow hardware you need to dial down the graphical load to keep the fps/refresh rate up.

I'm confused by this explanation. The framerate shouldn't affect the physics of the simulation, only the smoothness of the visuals. Normally, sim software sends the positional information to your system at the same rate regardless of how quickly the system subsequently sends it to your viewing device. Dropped frames (a jerky output) or a torn image are the usual results of low framerates. The smoothness of the final image is usually determined by the framerate and the physics engine should just carry on as normal. How would the physics engine know what the output framerate was? The sim is just pumping out the data to the buffer of the graphics card. How would it be a realistic simulation if the physics changed as a result of the output framerate? I agree with @Kilrah, the physics should be linked to time, not framerates. There's something seriously wrong if the force of gravity in the simulation varies as as result of output framerate.

 

[slup]

I'm confused by this explanation. The framerate shouldn't affect the physics of the simulation, only the smoothness of the visuals. Normally, sim software sends the positional information to your system at the same rate regardless of how quickly the system subsequently sends it to your viewing device. Dropped frames (a jerky output) or a torn image are the usual results of low framerates. The smoothness of the final image is usually determined by the framerate and the physics engine should just carry on as normal. How would the physics engine know what the output framerate was? The sim is just pumping out the data to the buffer of the graphics card. How would it be a realistic simulation if the physics changed as a result of the output framerate? I agree with @Kilrah, the physics should be linked to time, not framerates. There's something seriously wrong if the force of gravity in the simulation varies as as result of output framerate.

Really doesn't matter what we think ... it is what it is.

A sim used on a system too weak to keep the frame rate up, will give effects on the floatiness we are speaking of. It's not a matter of reasoning ... it's a matter of how it is. And right now the most used sims for FPV is requesting a hardware that can keep the frame rate up high in order too feel realistic. In some sims you can adjust other parameters like prop power, drag impact, quad weight, or battery power impact in order to somewhat mitigate the effects from a lower frame rate... but so far frame rate is king.

Velocidrone is well known for being the better sim for weaker hardware systems (& by weaker systems I mean a 3-5 year old top of the line game rig)... this as it's there possible to throttle down very much of the visual effects to relief the system, in order to gain frame rate so you can have a realistic feel of the game physics ... which is the most important to get a proper training... the graphic is just bling after all.

 

[Pol]

From what I've read the DJI FPV drone will use the DJI Fly app. I checked and the latest app was up yesterday. During the last ten years or so of flying DJI Phantoms and DJI build your own drones, a couple of which I flew with early Fat Sharks. The FPV experience back then was not very good. I recently got a little Cinewhoop and feel that excitement again. It seems to me one way to get us Mavic folks interested and excited about their new FPV drone is let us have access to the simulator!!! Anyone else want a chance to experience FPV without a lot of money up front? We need to convince DJI, access to their new FPV simulator (especially on an iPad), would help drive their business. The current PC based DJI simulator does not work very well.

ME TOO

 

[dsmith76]

Yep ... a floaty simulator is devastating for a proper training, you will not get a grip around how the gravity affects the quad. It's probably due to a too low frame rate in the sim. In order to improve that try to turn down the graphics if it's possible.

Anyone that have tried to hook up the new DJI FPV controller to Liftoff or Velocidrone running on a computer?

I’ve used the FPV RC with FPV FREERIDER simulator and it is much better than using my Taranis QX7. Sticks are softer. BTW, I much prefer Freerider to either Liftoff or Velocidrone. It is simpler and graphics are better, st least on my 2015 MacBook Pro. Only has 6 tracks/scenes, but they all require somewhat different skills. And, it’s free! I still do Liftoff and Velocidrone once in a while, but they seem designed more for racing. Much harder for me to fly/navigate than Freerider.

 

[maelstrom]

A sim used on a system too weak to keep the frame rate up, will give effects on the floatiness we are speaking of. It's not a matter of reasoning ... it's a matter of how it is. And right now the most used sims for FPV is requesting a hardware that can keep the frame rate up high in order too feel realistic. In some sims you can adjust other parameters like prop power, drag impact, quad weight, or battery power impact in order to somewhat mitigate the effects from a lower frame rate... but so far frame rate is king.

Then it must be something peculiar to FPV sims. If you're having to adjust the parameters of the drone to compensate for changing physics at differing framerates, there's something wrong with the sim programming. With normal flight sims, for example, if the system isn't powerful enough and framerates are low, the aircraft seems to move in a series of jumps. Each jump is to where the aircraft would have moved smoothly if the framerates had been higher. The physics (and the effect of gravity) doesn't change. That would make absolutely no sense in a simulator of any kind.

You said "a floaty simulator is devastating for a proper training, you will not get a grip around how the gravity affects the quad. It's probably due to a too low frame rate in the sim". That just doesn't make any sense and it's definitely not how real-world flight sims work (trust me, I worked in the development office of one for a number of years). If you were descending 100m in a sim at 10m/second, it would still take you 10 seconds no matter what the framerate was. It would just appear as a more noticeable series of tiny jumps as the framerate reduced. Anyone who writes a simulator program where the effect of gravity somehow changes as a result of the output framerate needs to do some recoding.

 

[slup]

Then it must be something peculiar to FPV sims. If you're having to adjust the parameters of the drone to compensate for changing physics at differing framerates, there's something wrong with the sim programming. With normal flight sims, for example, if the system isn't powerful enough and framerates are low, the aircraft seems to move in a series of jumps. Each jump is to where the aircraft would have moved smoothly if the framerates had been higher. The physics (and the effect of gravity) doesn't change. That would make absolutely no sense in a simulator of any kind.

You said "a floaty simulator is devastating for a proper training, you will not get a grip around how the gravity affects the quad. It's probably due to a too low frame rate in the sim". That just doesn't make any sense and it's definitely not how real-world flight sims work (trust me, I worked in the development office of one for a number of years). If you were descending 100m in a sim at 10m/second, it would still take you 10 seconds no matter what the framerate was. It would just appear as a more noticeable series of tiny jumps as the framerate reduced. Anyone who writes a simulator program where the effect of gravity somehow changes as a result of the output framerate needs to do some recoding.

Then I recon you need to get in touch with them & tell both Velocidrone & LiftOff that they need to redo & do it right & change their recommendations for true physics.

I'm only a user & not a game programmer ... but right now their recommendations is correct & it works as they say.

 

[maelstrom]

Then I recon you need to get in touch with them & tell both Velocidrone & LiftOff that they need to redo & do it right & change their recommendations for true physics.

No point as I won't be buying the software. Just pointing out how real sims work.

Incidentally, I have aeroflyRC7 (which includes quads) on my old and new PCs and running it at very different framerates makes no difference whatsoever to the "feel", handling or response of the drone, just how smooth the visuals look on my monitor.

 

[slup]

...Just pointing out how real sims work.

Incidentally, I have aeroflyRC7 (which includes quads)...

Well, both Velocidrone & LiftOff is "real" sims also ... very reputable within the FPV community as a matter of fact.

And if you're satisfied with Aero Fly RC7 & feels that is enough for FPV training just go for it

 

[cab124]

When using the simulator on an iPhone using the DJI FPV goggles and remote, I was noticing that the controls for navigating the app (selecting menus, etc.) from within the goggles did not work. I had to do this from the phone. Is that normal, or was I experiencing a problem?

Thanks!

 

[Kilrah]

You use the remote to navigate in the app, not the goggles' joystick.

 

[cab124]

You use the remote to navigate in the app, not the goggles' joytick.

Thanks! I wasn't sure how that was supposed to work. Totaly Newby.

 

[cab124]

You use the remote to navigate in the app, not the goggles' joystick.

So when is the goggles' joystick typically used? Maybe when the drone is in the air and you need to navigate the menus in the goggles?

 

[Tufargon]

Can someone try with the MA2 controller?

 

[Kilrah]

Thanks! I wasn't sure how that was supposed to work. Totaly Newby.

There are nice little pictograms in the app showing it...

So when is the goggles' joystick typically used? Maybe when the drone is in the air and you need to navigate the menus in the goggles?

All the system setup for operating the drone and goggles, basically everything but the simulator.

Can someone try with the MA2 controller?

Nope, only works with the FPV's one.

 

[Tufargon]

Nope, only works with the FPV's one.

Man that sucks. We've got the only RC that doesn't work with any of the DJI simulators.

 

[maelstrom]

Well, both Velocidrone & LiftOff is "real" sims also

I'd argue the accuracy of that description if the physics really do change with the framereate. But if you're happy to accept that, then who am I to try to change your mind.