If Mini 2 SDK support isn't released soon, it may be too late to matter

[CLAYB.7]

"Today at 03:24
Dear Client,

The release date is out and we plan to release MSDK 4.16 at the end of December. We greatly appreciate your patience.

Hopefully our solution can help you.

Kindly Regards,
DJI Developer Support"

Yeah Brother I pretty much got the same thing from DJI a while back and so I am going to sit back and chill and wait until Mid December and hopefully DJI will release it and they will not put it off no more so here's to them really releasing it this time!!! God Bless Sincerely Clay B.

 

[TT3023]

So, how about getting waypoints and more on the Mini 2 ... TODAY! Yeah, we're now hearing that Mini 2 support in the MSDK has been pushed back (yet again) to the end of December. And frankly, I wouldn't want to bet the farm on that date, given the history up to now.

But what if there were a way to do waypoints and more with the Mini 2 right now. It turns out that this is apparently possible!

Months ago I asked if there might be a way to spoof the existing MSDK (which supports the Mini) to also work with the Mini 2. Unbeknownst to me until today, that work has been taking place and apparently has been accomplished.

The project apparently creates a wrapper around the existing MSDK to support Mini 2 via MAVLink and QGroundControl.

I've just started looking this all over but it looks extremely promising.

The key software that makes this possible is here:

GitHub - m4xw/rosettadrone_mini2: The Rosetta Drone 2 Project. Enable Mavlink and H.264 on DJI drones. Supports Mini 2 with Virt Stick Litchi CSV and DJI Waypoint class interpreters

The Rosetta Drone 2 Project. Enable Mavlink and H.264 on DJI drones. Supports Mini 2 with Virt Stick Litchi CSV and DJI Waypoint class interpreters - m4xw/rosettadrone_mini2

github.com

Pretty cool, eh?

Here is a YT video that summarizes the project. It's in Italian but auto-translate captions from Italian to English works well:

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

Accept third party cookies

 

[TT3023]

Here are some English YT videos about the Rosetta Drone project. Apparently it has been in development for several years, but Mini 2 support was seemingly only added a few months ago.

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

Accept third party cookies

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

Accept third party cookies

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

Accept third party cookies

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

Accept third party cookies

 

[TT3023]

I now have this built, but not tested yet. We shall see ...

 

[SkittleBob]

I now have this built, but not tested yet. We shall see ...

I'm watching from a distance, to be honest a long one.

Is this going to spoof (say) Litchi into thinking it is a Mini, or are we limited in software?

 

[CLAYB.7]

I'm watching from a distance, to be honest a long one.

Is this going to spoof (say) Litchi into thinking it is a Mini, or are we limited in software?

I am starting to think that DJI is waiting until it is right at the time that they are going to release the DJI MAVIC MINI 3 before they release the SDK update for the MINI 2 and that way they can sell as many as they can of the DJI MAVIC MINI 2 because they are going to have to go all out on the DJI MAVIC MINI 3 because all you have to do is look at what the other guys are doing and it looks like they are doing it well I mean just look at what AUTEL just released there EVO NANO AND NANO+ and that thing is loaded with all kinds of top notch features and they still stayed under the 250g. limit to do it to!!! and let's not forget about the HUBSAN ZINO MINI PRO so we know that it can and has been done given top notch features and staying under the 250g. limit and I think that is why DJI is waiting or it's pushing back the release of the SDK UPDATE but I guess that we'll have to wait and see but I hope that this DECEMBER is going to be it and they'll finally release it!!! I WOULD LOVE TO FINALLY GET TO USE WAYPOINTS AND THE OTHER FEATURES ON OUR DJI MAVIC MINI 2 BUT WE'LL SEE WHAT HAPPENS. God Bless Sincerely Clay B.

 

[TT3023]

Every one of these additional features stuffed in to a sub 250 drone is compromising something else to stay under that weight. As far as I'm concerned at this device level obstacle avoidance is a crutch at best and potential problems at worst. Poor obstacle avoidance systems are worse than having none at all. They add unpredictability and they're something else to fall out of calibration or react in undesirable ways at the worst possible time. Yes, when you get to very high end devices the implementations can get much better, but that's not what we're talking about here. I get so tired of MINI 3 this and MINI 3 that and oh man this one from the company is gonna be the DJI KILLER! OK, I know a lot of this is about engagement but c'mon. The MINI 2 is a solid product and utterly reliable. The controller doesn't look and feel like a toy. The camera is more than good enough for the vast majority of users. It is unclear if most MINI 2 owners would have any *real* need to upgrade to a 3 when such a device eventually appears under whatever name. And I'll address the Rosetta Drone question next ...

 

[TT3023]

I'm watching from a distance, to be honest a long one.

Is this going to spoof (say) Litchi into thinking it is a Mini, or are we limited in software?

You're misunderstanding I think. Rosetta Drone permits virtually ANY standard MAVlink-based ground station to be used. That means QGroundControl, for just one example, and DroneKit for programmable and AI-based flight ops. You don't NEED software like Litchi, you can use the same ground control software used commonly with PX4-based drones, etc.

So I now have it compiled and it runs. However, I'm having an issue getting the DJI SDK to authenticate, which appears to be a common issue with the SDK if the key is generated wrong, etc. Hopefully there isn't a more complex issue in the build but we'll see.

 

[SkittleBob]

You're misunderstanding I think. Rosetta Drone permits virtually ANY standard MAVlink-based ground station to be used. That means QGroundControl, for just one example, and DroneKit for programmable and AI-based flight ops. You don't NEED software like Litchi, you can use the same ground control software used commonly with PX4-based drones, etc.

So I now have it compiled and it runs. However, I'm having an issue getting the DJI SDK to authenticate, which appears to be a common issue with the SDK if the key is generated wrong, etc. Hopefully there isn't a more complex issue in the build but we'll see.

TBH I have no idea what you are doing, what a 'standard MAVlink-based ground station' is, or what a 'PX4-based drone' is.

As you were ?

 

[TT3023]

QGC Guide (master)

How to use and develop QGroundControl for PX4 or ArduPilot powered vehicles.

docs.qgroundcontrol.com

Building an Open Source Drone with PX4 using Pixhawk Open Standards thanks to Upverter

The Open Source and Open Standard efforts under the Dronecode Foundation umbrella make it easy to design and build a drone. Upverter takes advantage of the existing standards to build out your own Pixhawk FMU compatible board and I/O board.

www.dronecode.org

Introduction · MAVLink Developer Guide

mavlink.io

 

[SkittleBob]

So it's an open sourced version of Litchi?

 

[TT3023]

So it's an open sourced version of Litchi?

Well, yeah -- though since the QGroundControl project started around 2010 I think, as part of the Pixhawk/PX4 project (which began a couple of years earlier), it would probably be more accurate to call Litchi a closed-source version of QGroundControl. Here's a page with the history:

The history of Pixhawk | Auterion

auterion.com

 

[TT3023]

However, a key difference is that apps like Litchi have to be written to communicate specifically with the DJI MSDK. But Rosetta Drone provides a bridge so that essentially any of the many programs, apps, and tools that have been written to talk via MAVLink to PX4-style controllers can be used via Rosetta Drone to operate via the DJI MSDK -- they do not have to be modified to use with (for example) the Mini 2.

 

[photo2be]

The continuing delay in seeing support for the Mini 2 in the SDK is increasingly problematic. At this rate, there will be a new Mini model before the SDK support sees the light of day ... if ever. And I'm seriously wondering if that support will ever arrive. Given that competing small drones in the same class are appearing, some with features included integrally that could only be provided for the Mini 2 using apps via the SDK, the situation seems even more ominous for the future of this platform.

Hi TT3023, face fact...

Like all other manufacturers DJI doesn't even want to support or improve a released product too much. It will compete with the upcomming new prodcuct like Mini3 maybe. Enjoy and use what you already invested in.

Best from Sweden

 

[TT3023]

My view is that one of the factors that distinguishes DJI from the others is that they do offer SDKs and API interfaces. That's a lot of work, and potentially a lot of headaches for them. But it's great for a diversity of apps and use cases.

 

[SkittleBob]

Hi TT3023, face fact...

Like all other manufacturers DJI doesn't even want to support or improve a released product too much. It will compete with the upcomming new prodcuct like Mini3 maybe. Enjoy and use what you already invested in.

Best from Sweden

There's 'too much' and 'too little'. I for one won't buy another DJI drone until the SDK was out.

 

[photo2be]

Maybe a good idea but then you will have to wait fot the next drone forever..

There's 'too much' and 'too little'. I for one won't buy another DJI drone until the SDK was out.

 

[SkittleBob]

Maybe a good idea but then you will have to wait fot the next drone forever..

It'll be cheaper then.

 

[TT3023]

So I now have it compiled and it runs. However, I'm having an issue getting the DJI SDK to authenticate, which appears to be a common issue with the SDK if the key is generated wrong, etc. Hopefully there isn't a more complex issue in the build but we'll see.

And I've now solved the authentication problem (as I suspected, I needed to regenerate the key correctly) so will be able to start testing when I have some time.

 

[Storyline]

FWIW, as a new Mini 2 owner I've been trying to climb the learning curve and have watched loads of youtube videos on various aspects from flying techniques to how to get the best out of the camera and out of the hundreds of comments I read I don't think I can remember a single person who was clamouring for collision avoidance in a future model. It certainly has no interest for me. One thing that did seem remarkably popular was wanting some kind of follow me mode.
I am quite a keen photographer but know nothing about video photography so have a mountain to climb but one thing I've learnt over the years is that in any imaging system it is quite common that the weakest link is the photographer and getting the absolute best out of the hardware takes a lot of skill and practice.
Of course it would be great to get a bigger sensor and enhanced controls in the next model in the Mini series but one thing that I think I learnt today is that DJI could enable people to get better video images by firmware changes to the existing camera.