I have been wanting to know how the MAVIC lens resolution compares to the P4. while I was at it, I also compared it to the Inspire-1 x3, the xStar and the Yuneec Hpro... click through for the video and the results.
The goal was to create a fair comparison of the MAVICpro and the DJI Phantom 4. this turned out to be far more complicated than anticipated. and I ended up spending almost 10 days working on this project. in the end I did get some good solid numbers that give me some understanding how these compare, based on actual imaging system resolution and dynamic range.
Check out the video and look at the images below to follow along.
4K resolution composite of the video screen-grab's from the resolution test.
Here is the 6 up step wedge comparisons.. note there are 3 steps to 1 f-stop
The comparison is really rather simple, the devil is in the details. this is a matter of reducing errors.
Interestingly the flat part of the MAVIC grey-step-wedge only show up in video and JPG's, the DNG (raw) files did not display the flat part.
What we have in the test is two simple metrics for the lens-sensor combination that tells us a lot about the imaging-system's performance. I would say for photographers the biggest feature is dynamic range.
Im fully aware that a lot can be done to further refine the accuracy of these two tests. however the important part is that all the drones were measured equally, so the tests are consistent and can be used to see the differences between these cameras, the biggest question is not how the cameras ranks, but there could easily be a small inaccuracy in the f.stop or resolution number, just that inaccuracy likely applies to all of them so the ranking will remain the same, just with slightly different values.
BUT there are naturally more to fully knowing all the details about a sensor, some of these I had to skip to reduce the amount of time it took to finish this test. some of the BIG ones would be
LENS FLARE / diffraction
SENSOR NOISE LEVEL a good sensor can have a lot of noice, but typically a good DR also is a pretty clean sensor.
COMPRESSION & Bitrate.. not strictly camera, however the max bitrate is a huge determining factor to how good and how "flexible" the video will be later in editing.
I did shoot the grey-wedges also in RAW. all of the cameras dynamic range improved roughly 1 f.stop in raw.
Hey Roger, Im not planning on selling those, they are fairly simply to mimic though.
Posted by: Bo Lorentzen | 05/12/2017 at 11:33 AM
Bo- I;m alos using a Nividia shield, & I;m wondering if you are going to make/sell the shield holder for the Mavic??
cheers
Roy
Posted by: Roger | 05/11/2017 at 08:14 AM
Hi Bo, I am very impressed with your attention to details when doing your videos. I am interested in buying a Mavic and as I cannot try one out before purchasing I would like to ask you if you would allow me to have a RAW file, or 2, from the Mavic to test the dynamic range. I have an old DJI FC40 to which I have fitted a 2 axis gimbal to carry a Gopro Hero 3+, but I am not impressed with the quality of the still images from it. My original intention was to shoot aerial 360° panoramas, but due to the Gopro's limitations on exposure and white balance I gave up the attempts. I have some panoramas on my website, these are shot from ground level or a dual monopod approx 3.4 mtr max height.
I am going to Iceland in March and would really love to shoot some aerial stuff, and the Mavic looks ideal for portability, and I am hoping quality stills (raw) as well. Please let me know if you could send me some raw files from the Mavic to process.
Jeff
Posted by: Jeff Starley | 01/11/2017 at 08:00 PM
Hey Bo, I watched your Mavic un-boxing vid, good job by the way. In it you mounted your Nvidia Shield to the Mavic controller with 2 'S' clips. I assume you made them, any chance you will be providing them for sale? I want to use my Shield and don't like any of the other bulkier options, the one you used is perfect.
Jesse
Posted by: Jesse | 12/08/2016 at 02:34 PM
Hey Mark, yes absolutely, micro-vibrations is a huge issue, we used to see this with big DSLR gimbals even, the other big thing is fast moves changes a lot of pixels quickly resulting in the system compressing or dropping fine details to stay in the bandwidth.
Posted by: Bo Lorentzen | 11/15/2016 at 07:05 AM
Hey Bo,
Great stuff! Thanks for doing all of this.
Curious if you think that the stabilization systems of each camera can effect the "flying resolution test" results?
Naturally, we can't judge the camera without the drone in flight (the "whole system" as it were), but there are variables at play here that maybe should be either explored, controlled, or just disclosed (video compression, windspeed, distance to target, etc.).
Posted by: MarkB | 11/15/2016 at 06:56 AM