Cable Cam and the Competition..

Joined
Jun 4, 2015
Messages
2,911
Reaction score
1,699
Age
67
Serious question on the P4 and other birds just introduced from Yuneec and others
:
From back to when the Solo was introduced and even today, the Cable Cam feature is one of the most desirable features and now the MPCC expands that ability greatly. For cinematographers in particular this seems to be very much wanted and appreciated and I can certainly see why. So this being the case, why was something similar not included in new introductions like the P4 and others? I haven't seen it mentioned if it is and maybe I missed it. If not, is it a technical limitation that the Solo relies on the processors in the controller and Solo that the p4 and others lack? Or is it just software that DJI has failed to implement? If it is just software, where might we see the dependence on the CPUs the Solo has that other units without them would be unable to duplicate?
@Ian [P13], @franknitty69, @EyeWingsuit, what say you?
Thanks..
Jubal
 
the implementation of the freescale imx cpu on the solo is pure genius. it makes complicated smart shots like mpcc easy to add. if you break down what mpcc is doing to basic blocks of function, technically, the same thing could be done in an ios/android app. but it would never be as smooth as a solo. its that interpolation at the beginning, between and end of movements that sets the solo apart from the rest of the field. and in my opinion, can only be performed with a companion computer that has direct connection with a flight controller. the freescale imx is also capable of computer vision btw.

as to why a DJI or Yuneec hasn't copied mpcc already, i think it's due to the lack of a true companion computer in those drones. the p4 has some additional processing power with the new obstacle avoidance system, no indication that it's a cpu. but the 3d exploded view looks like it's not. whatever it is, maybe it can handle mpcc, maybe it can't. but then again you have to remember DJI doesn't like to give things to their customers for free. so if the P4 is capable of true mpcc, expect to see it implemented in the P5.
 
Thanks Frank. My uneducated opinion was that it had to be something related to the CPUs in the controller and Solo. It would have to take a load off the app. Like when they started adding real GPUs to graphic cards and freed resources for the CPU in computers. But then I also think that the Litchi program I have heard about for the Phantoms, but never seen. It is supposed to have similar features to Tower, but I doubt it could duplicate a lot of it.

I hope to see more complicated smart shots that truly take advantage of the Solo and CPUs.
 
Thanks Frank. My uneducated opinion was that it had to be something related to the CPUs in the controller and Solo. It would have to take a load off the app. Like when they started adding real GPUs to graphic cards and freed resources for the CPU in computers. But then I also think that the Litchi program I have heard about for the Phantoms, but never seen. It is supposed to have similar features to Tower, but I doubt it could duplicate a lot of it.

I hope to see more complicated smart shots that truly take advantage of the Solo and CPUs.

yeah me too. UI support is the biggest issue right now. if you make a smart shot, there isn't a ui to trigger it, adjust settings, etc. the only way it they can be used at the moment is to hack shotmanager or trigger your smart shot from a computer connected to solo wifi. 3dr is working on an api for smart shots that will let you drop in custom smart shots and at least have some basic ui in the app for using them.
 
Processing power is one thing. Access to the processors is another. Its difficult to say what kind of processors are included in the P4, or how powerful said processors are, but I think most of us have infered that most of the hardware is dedicated to a single purpose (eg obstacle avoidance). It not unlike graphics cards in a PC. Years ago there was no way to access a GPU in a PC for generalized computing. Eventually it became obvious that there was a whole lot of processing power sitting in a PC going unused. Now a days GPUs are used for a myriad of compute tasks unrelated to rendering video on screen (for example mining bit coin).

I'd love to learn more about the system architecture in a P4, but I suspect that most hardware is dedicated to specific tasks. DJI seems like the kind of company that would be pretty hesitant to throw in a cpu running a Linux based OS, which would require compliance with open source licensing (even though it can all be fairly easily abstracted from the rest of the craft). The only other option would be to run a proprietary closed source OS on any included companion computer. That would require a whole different level of development effort in an area they don't have any real expertise.
 
Last edited:
i have been trying to find out the specs of Yuneec flight controller and processing for the ultrasonic collision avoidance. at ces, Intel CEO Brian Krzanich showed a typhoon h and said that it had intel realsense as well as an intel cpu. but on the yuneec website they make no mention of the intel cpu or intel realsense. and in this ces interview, the yuneec rep even said that intel realsense is an add-on.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

anyways, the typhoon curve cable cam sounds just like mpcc, but i have yet to see to actually see it in action. so i guess we will see.
 
i have been trying to find out the specs of Yuneec flight controller and processing for the ultrasonic collision avoidance. at ces, Intel CEO Brian Krzanich showed a typhoon h and said that it had intel realsense as well as an intel cpu. but on the yuneec website they make no mention of the intel cpu or intel realsense. and in this ces interview, the yuneec rep even said that intel realsense is an add-on.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

anyways, the typhoon curve cable cam sounds just like mpcc, but i have yet to see to actually see it in action. so i guess we will see.
Yeah, I was trying to research it more as well. But other than it being mentioned in the specs, no talk of videos of it that I can find.
 
Yeah, I was trying to research it more as well. But other than it being mentioned in the specs, no talk of videos of it that I can find.
What ever the specs end up being in the P4, I think it's far more likely that Yuneec would incorporate a general purpose cpu (a companion computer) running some variant of Linux as they already do on the controller.

I think I'll grab a Yuneec h when the full package is available.
 
What ever the specs end up being in the P4, I think it's far more likely that Yuneec would incorporate a general purpose cpu (a companion computer) running some variant of Linux as they already do on the controller.

I think I'll grab a Yuneec h when the full package is available.
I was watching some videos of the H yesterday when I was looking for that smart shot. On paper the machine looks great, but when you look at the build quality it looked very cheap build. I think it is supposed to be out next month, so we'll see what the reviews look like.
 
I was watching some videos of the H yesterday when I was looking for that smart shot. On paper the machine looks great, but when you look at the build quality it looked very cheap build. I think it is supposed to be out next month, so we'll see what the reviews look like.
I have to do some more research on the H. It's available to buy on their site, but I get the impression that there is an enhanced intelisense version coming. If I do get it, I'll be sure to post back my impressions.
 
Does MPCC still have to be flown ahead of time or can it be selected from GS as waypoints?
Litchi has functioanlity that seems to be pretty close, not yet sure that the pitch transitions are as smooth.

The litchi implementation has the CC as a subset of their waypoints and they do not need to be flown in advance. But again not sure they can achieve the same dynamic smoothness on all transitions, although the yaw is pretty good.
 
Does MPCC still have to be flown ahead of time or can it be selected from GS as waypoints?
Litchi has functioanlity that seems to be pretty close, not yet sure that the pitch transitions are as smooth.

The litchi implementation has the CC as a subset of their waypoints and they do not need to be flown in advance. But again not sure they can achieve the same dynamic smoothness on all transitions, although the yaw is pretty good.

You still need to fly first, but you would almost have to. Remember in MPCC you are framing the shot at each point with the camera. Can't do that from a map. But WPs and tilt are the easy part. The part that they haven't been able to demonstrate with Litchi and some of the other 3rd party apps the ability to vary the speed of the tilt. In MPCC, the tilt between 2 points is timed to start very slowly and then tilt at a speed that is synced to arrive at the next WP as the tilt is completed. The same thing with differences in height between points. This is what give the Solo MPCC videos a cinematic feel. Here is a CC I did this morning just messing around.

I believe there were 5 or 6 points in this MPCC. Note that as the camera tilts along the route, you will see various tilt speeds depending on the length between points. The first one is the longest, and the points get closer to each other at the end. You almost don't realize the camera is tilting until you realize the horizon is now in view. Yet between each set of points, the camera is done with the tilt exactly at the next wp. It is these non linear/variable tilt speeds that cinematographers look for. In addition, each point is at a different altitude and this is timed between each point as well. Note at the very end of the video, the tilt is moving extremely slowly to the final camera view.
Start to finish was completely hands off the sticks.
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
I believe they are using Bezier spline curves/math on the camera transitions, otherwise known as "ease in/out". The MPCC points are most likely using what are called Conic Splines. Here is a graph showing various spline types typically used in 3d/cad/etc.

img_53bd73c3d0544.png

There's a lot of computations going on, which is hard if not impossible to do without a dedicated computer to process them. Take a simple 3 point MPCC in a triangle shape. It's actually splitting it up into thousands of individual gps coordinates to sit on those calculated spline curves. It's not just going straight from A to B in a mechanical fashion.
 
I believe they are using Bezier spline curves/math on the camera transitions, otherwise known as "ease in/out". The MPCC points are most likely using what are called Conic Splines. Here is a graph showing various spline types typically used in 3d/cad/etc.

img_53bd73c3d0544.png

There's a lot of computations going on, which is hard if not impossible to do without a dedicated computer to process them. Take a simple 3 point MPCC in a triangle shape. It's actually splitting it up into thousands of individual gps coordinates to sit on those calculated spline curves. It's not just going straight from A to B in a mechanical fashion.
Yes, that is what I was trying to explain to some DJI owners and why it may be difficult for an app to accomplish on the Phantoms. But like I told them, it takes your input and creates 100s or more coordinates along with the gimbal timing. But I like the way you explained it and I have posted a link to your post elsewhere. Thanks!
 
Last edited:
You still need to fly first, but you would almost have to. Remember in MPCC you are framing the shot at each point with the camera. Can't do that from a map. But WPs and tilt are the easy part. The part that they haven't been able to demonstrate with Litchi and some of the other 3rd party apps the ability to vary the speed of the tilt. In MPCC, the tilt between 2 points is timed to start very slowly and then tilt at a speed that is synced to arrive at the next WP as the tilt is completed. The same thing with differences in height between points. This is what give the Solo MPCC videos a cinematic feel. Here is a CC I did this morning just messing around.

Jubalr,

Just for point of clarification a Litchi CC can be flown ahead of time with camera framing. I posted an example in the other forum. The question still remains whether it can come close to the smooth yaw and pitch transitions possible with MPCC. It has not yet been demonstrated in any video I have seen, but that doesn't mean it can't come close, but maybe not close enough for discriminating cinematographers.
 
Jubalr,

Just for point of clarification a Litchi CC can be flown ahead of time with camera framing. I posted an example in the other forum. The question still remains whether it can come close to the smooth yaw and pitch transitions possible with MPCC. It has not yet been demonstrated in any video I have seen, but that doesn't mean it can't come close, but maybe not close enough for discriminating cinematographers.
Hi Greg, Yes I was aware that it could be flown in advance, as that was in their demo video. I was answering the OPs question about MPCC and if it had to be flown in advance. And it does, how could you frame a shot with the camera from a sat image?
 
Great improvement over flying waypoints in Tower to frame the shots. I am still experimenting with different terrain, to frame the shots. This is an incredible tool for video taping.
 
Great improvement over flying waypoints in Tower to frame the shots. I am still experimenting with different terrain, to frame the shots. This is an incredible tool for video taping.
Ahh, 'Video Taping'. You must be an old guy like me..:)
 
  • Like
Reactions: Adimuzio513

Members online

No members online now.

Forum statistics

Threads
13,100
Messages
147,774
Members
16,073
Latest member
andre felipe colorado