- Joined
- Aug 24, 2014
- Messages
- 219
- Reaction score
- 75
Don't have the android version - sorry..
I didn't notice zoom on this beta software but there is a very good chance it will be be added (IMHO) in either the first production SW or later.
Has anyone flown the P4P
- Thread starter Alf
- Start date
Don't have the android version - sorry..
I didn't notice zoom on this beta software but there is a very good chance it will be be added (IMHO) in either the first production SW or later.
Meta4
Premium Pilot
- Joined
- Jan 8, 2014
- Messages
- 15,354
- Reaction score
- 9,296
Good question, but I'd like to hear verification that with OA enabled that it will actually do 30mph. My P4 will only do about 15 to 16mph with OA enabled, nowhere close to the advertised 22mph.
- Joined
- Aug 16, 2016
- Messages
- 86
- Reaction score
- 16
- Age
- 57
The main downside to H265 and the reason adoption is going slow is that it is very processor intensive. Generally it is decoded on the GPU but depends on the OS and driver support. Currently Apple does not have drivers for quicktime but Premier Pro installs it's own decoder. I tried CPU and GPU (CUDA and OPEN-CL) options and all were about the same which means on my Mac at least it is just using CPU (which is why it is slow). But I am not on Sierra so not sure if that has better support for H265
- Joined
- Mar 10, 2016
- Messages
- 322
- Reaction score
- 83
I downloaded both, the short video (original from P4P) and your first movie. Here are some observations:
1) DJI_0015.MP4 is a high bitrate (100 Mb/s) encoded as MPEG-H Part 2 HEVC/H.265 Video Elementary Stream. The quality of the frames where the P4P is not moving are absolutely awesome. No significant mpeg artifacts visible at 400%. However, you must have had a very slow shutter speed because the moving portion of the video has a lot of motion blur taking away from the quality.
2) Your produced video unfortunately uses a relatively low bitrate encoding using MPEG-4 AVC/H.264 Video Elementary Stream with a 29 Mb/s bitrate which is not enough for UHD footage using AVC/H.264.
I am very impressed by the original recording quality. It is an order of magnitude better than similar footage taken with a P3P.
Totally amazing....
- Joined
- Aug 24, 2014
- Messages
- 219
- Reaction score
- 75
Unless I personally test stuff I don't like to guess! So all we have to go by is the factory specs for now....
I actually turn all OA off on all of my craft 95% of the time. Also, since I don't do much following of high speed objects the speed of my machines tends to suit what is best for landscape and landmark videos - usually quite slow.
Sorry - can't be accurate about this part...but users are starting to receive them today so I'm sure some more daring pilots will be flying them soon.
I'm a boring pilot - don't fly them except to use then for what they are made for - taking pics and vids in LOS.
Meta4
Premium Pilot
- Joined
- Jan 8, 2014
- Messages
- 15,354
- Reaction score
- 9,296
Unfortunately the specs don't say.
All they offer is this: Max Speed 45 mph (72 kph) (S-mode); 36 mph (58 kph) (A-mode); 31 mph (50 kph) (P-mode)
I need to be able to fly offshore where headwinds can be a big issue and object avoidance isn't.
I hope that the pro version will go faster without OA like the P3 and P4 do but it would be good to get confirmation.
Video with movement will be harder to compress with comparable quality and drone video is particularly a problem in that regard as every pixel changes every frame when the drone is in motion. This is magnified when the camera is tilted down and lessened when tilted up. Also, turning or panning aggravate the problem.
I can't say I know the detailed specifics of how h.265 differs from h.264 or if it is better at some thing and worse at others. Is it able to retain IQ better with static images but not so well with movement?
As an owner of an Inspire 1 Pro with M43 X5 camera the single thing that pisses me off most of all is the flickering/pulsing/breathing thing is does in the shadow detail when you stretch the contrast with curves or similar. I've had to adjust my grading to reduce the amount of contrast stretching that I'd prefer to do to minimize the flickering thing. It's been suggested that this is a function of the IPB thing where the PB frames drift too far between the I frames. I wonder if DJI/Zenmuse are using the correct coding to handle the IPB frames and further wonder just how they'll implement that work with h.265. That is, I'm not all that confident that DJI/Zenmuse are doing the job properly with the Inspire 1 Pro and potentially the I2 etc.
Brian
- Joined
- Mar 10, 2016
- Messages
- 322
- Reaction score
- 83
The stream analysis tool I have is freeware and does not give me more detailed H265 encoding parameters settings. Frankly given the extreme computational cost of encoding H265 has me very surprised to see it in such a small package with power consumption constraints.
The detailed frame grabs taken into photoshop from the moving parts of the clip clearly demonstrate motion blur - not encoding deficiencies. Unlike the general consensus of P3P users attempting to slow down the shutter speed to minimize temporal distortions, I attempt to go for the fastest possible speed. I much prefer clear frames over motion blur (the latter referred to as cinematic). Modern TV such as the LG 65" 4K OLED which I use compensate for motion flicker incredibly well.
I use RESOLVE for grading only (create a clip specific LUT) then edit the video in Adobe Premiere PRO (latest version). Adobe offers very good encoder capabilities including H265. I don't use that encoder because it takes so much compute time. H265 has been developed to allow streaming high quality video including UHD over internet and satellite as it emphasizes on much higher compression than H264. Memory is cheap and as long as I don't publish online, high bit rate H264 does a fine job. My UHD material is H264 encoded at variable rate 60-90 Mb/s. That looks as good as the original from the P3P and your 100Mb/s H265 video loses only little.
Sent from my iPad using PhantomPilots
The detailed frame grabs taken into photoshop from the moving parts of the clip clearly demonstrate motion blur - not encoding deficiencies. Unlike the general consensus of P3P users attempting to slow down the shutter speed to minimize temporal distortions, I attempt to go for the fastest possible speed. I much prefer clear frames over motion blur (the latter referred to as cinematic). Modern TV such as the LG 65" 4K OLED which I use compensate for motion flicker incredibly well.
I use RESOLVE for grading only (create a clip specific LUT) then edit the video in Adobe Premiere PRO (latest version). Adobe offers very good encoder capabilities including H265. I don't use that encoder because it takes so much compute time. H265 has been developed to allow streaming high quality video including UHD over internet and satellite as it emphasizes on much higher compression than H264. Memory is cheap and as long as I don't publish online, high bit rate H264 does a fine job. My UHD material is H264 encoded at variable rate 60-90 Mb/s. That looks as good as the original from the P3P and your 100Mb/s H265 video loses only little.
Sent from my iPad using PhantomPilots
- Joined
- Jan 31, 2016
- Messages
- 112
- Reaction score
- 53
- Age
- 80
Will the android monitor that comes with the controller support installing applications such as Litchi? This would be very important for those who want more advanced flight features.
Official DJI answer is no, but we will have to wait until some P4P+ models get out into user's hands to know for sure. I would be surprised if someone doesn't come up with a hack to do so.
- Joined
- Mar 10, 2016
- Messages
- 322
- Reaction score
- 83
I downloaded the Straight-from-the-Phantom-4-Pro file DJI published. If DJI did this to demonstrate the picture quality of the P4P camera - it failed! It has 4 pictures taken by a P4, a Canon 5D, an Inspire (Zenmuse Camera) and the P4P. Unfortunately, the elapsed day-time between the pictures is about 3 hours and the subject is a skyline near sea-level with some fog rolling through clearly dissipating over the 3 hours. The 5D Mark III picture was taken at 9:16am with lots of fog. The P4P was taken at 11:32am. The Inspire/Zenmuse was taken at 12:06pm and the P4 picture was taken at 12:21pm. The Canon picture is a useless comparison due to the extensive fog at the time of the shot. The other three are reasonably comparable. After some picture analysis and "massaging" in Photoshop it is clear that the OPTICS on the Zenmuse FC550 are quite a bit better than the optics on the P4P. Even with a resolution advantage (20MP) over the Zenmuse FC550 (16MP) the Zenmuse produced a noticeable better, much clearer picture with more mid-range detail. In fact even the P4 (12MP) picture has more clairty and definition in the midrange than the P4P. This makes me suspicious about the impact of the fog (slightly more significant while the P4P was taking the picture).
I hope DJI is publishing a better set of 4 pictures with little time and condition differential. This set of pictures makes the P4P very disappointing.
I hope DJI is publishing a better set of 4 pictures with little time and condition differential. This set of pictures makes the P4P very disappointing.
I don't think you can tell much from the set of photos you described since the conditions were so uncontrolled. When testing cameras it takes multiple exposures under controlled/consistent conditions before one can really make any judgments.
- Joined
- Mar 10, 2016
- Messages
- 322
- Reaction score
- 83
Yep - that was the point of my note. Shame on DJI - why publishing something that useless!
I think the only somewhat valid observation relates to apparent superior quality of the optics of the Zenmuse camera.
Last edited:
