If you use this method, I suggest that you take several photos. I expect that the latency will be different each time. The videos I posted earlier show a range of latencies from 200 to over 300 msecs. And of course, there's the inherent variation due to the lack of synch between the display frame rate and the still camera shutter. This would be 16.7 msecs if the Lightbridge frame rate was 60 fps - but I think it's probably lower so the variation will be proportionately greater.
Video latency (lag) what is yours and what's your set up?
- Thread starter CactusJackSlade
- Start date
Thanks for posting this. I looked at several places and saw lags ranging from as low as 200 msecs up to 290 or 300 msecs. The most typical lag was around 240 msecs.
I wonder if video with more fine detail and dynamically changing content would produce greater lag. Our stopwatch test differs from a real flight in having a mostly out-of-focus image with little frame to frame changes.
CactusJS, everyone with Lightbridge 1 has lag, there's no getting around it, even if you bypass the Go app with a monitor. The H.264 video compression at the craft (encoding) and the decompression at the display device (decoding) takes time, even if it's done in hardware. H.264 has an inherent delay because of all the compress/decompress work.
Here's my iPad Mini delay, measured by using the technique mentioned earlier in post #22.
I don't think anyone has experimented with this. On the P3P you can dial down the video link bandwidth so it should be possible to compare latency at 2mbps vs 10mbps. But other things might also affect latency, including other processor-intensive functions like recording video. It might even depend on the complexity or rate of change in the scene being transmitted.
I'm not a FPV pilot, but my impression is that the Phantoms were primarily designed as great video platforms, and support for usable FPV was a low priority.
So I've got an Ipad Mini 1. running IOS 8.4. video latency using the video and stopwatch ranges between 0.55 sec and 1.5 sec. average of 0.82 sec with 7 measurements. hardware decoding was set to off. I could not find out where to reduce the downlink quality... My other problem is that the DJI Go app crashes during flight the last two times I've used it. that's got me a bit nervous. sounds like everyone likes their ipad air2. Just spent a grand on a drone. will the spending ever stop? 
- Joined
- Jan 3, 2016
- Messages
- 741
- Reaction score
- 302
- Age
- 72
So I've got an Ipad Mini 1. running IOS 8.4. video latency using the video and stopwatch ranges between 0.55 sec and 1.5 sec. average of 0.82 sec with 7 measurements. hardware decoding was set to off. I could not find out where to reduce the downlink quality... My other problem is that the DJI Go app crashes during flight the last two times I've used it. that's got me a bit nervous. sounds like everyone likes their ipad air2. Just spent a grand on a drone. will the spending ever stop?
As you probably know, the iPad Mini 1 isn't an approved device since it's processor isn't up to par. I've got the next gen, an iPad Mini 2 with the A7 processor and it works great. I picked it up on eBay for $160 and looks new.
S
I wasn't aware that the mini 1 isn't approved - Thanks. My iphone 6 works great comparatively - just not big enough! shopping around...
ianwood, the logic of what you say seems feasible, but when I tested HDMI vs the GO app running on my iPad Mini2, the latency is always identical. Have you done testing and found other results?
I think we all know that when measuring latency, each time you snap a photo to capture the latency, you will get different results. With my P3P and iPad mini2 I have measured between 180ms and 250ms (1/4 second). But when I have the HDMI goggles connected to the RC, as well as the iPad Mini2 connected to the RC USB port running GO, I always see the exact same latency number on both displays, down to the thousandth of a second. It's uncanny, almost unbelievable. I expected some difference, but it's always identical. I can't explain it. Here's a photo of my test last night. I ran about 6 tests, all of them show identical latency.
I use the Nvidia Shield with great results and chose this tablet specifically for using with my P3A. I considered ipad but preferred android and have been totally happywith the shield's performance. Two gripes shield users have are that it does not come with its own micro USB charging cord. Second, the micro SD card slot can be a pain to insert/eject because it is fairly recessed in the side of the tablet. Overall though, I still give the tablet solid marks for its display performance during drone flights.
I can. the cause of the latency is primarily in the encoding. H264 (and 265) use schemes that minimize data by comparing frames. You can not begin to encode without first buffering a few frames and then calculating the differences between them. Thats the cause of the lag, and there is no practical way around it if dont know what is coming in the future.
Well, there is, you can use a different encoding scheme that compresses each frame independently, like MJPEG does. With such a scheme, you could get close to zero latency, but that comes at a huge price, as its way less efficient and thus requires much more Kb per second (ie bandwidth) or has far worse quality for any given bandwidth. Since bandwidth is very limited in our application, that either means it will look terrible or you will have terrible range.
Here is an alternative to lightbridge that picks other priorities:
CONNEX™ ProSight HD Vision Pack for FPV US Version
It has virtually no latency, but you only get HD at ~100m or so range. If you go further than that, image quality looks horrible, worse than any analog system. But you dont have a latency problem.
Range, image quality, latency.
pick 2.
I would miss the 2X zoom on the iPad while recording in 1080p. It's nice to see up close while flying, great resolution on the iPad since it's cropping the 4k sensor.
Another tablet discussion thread,one just got shutdown today already by mod,for what I dont know.
Lots of differing opinions and different results with the same tablets on this thread
Tablets and phones that work well with DJI GO
as far as ipad mini 1 goes,i use one strictly for flying with just dji go app installed and so far it has worked fine for me even in 95 degree texas heat but it doesnt work for others,why I dont know.
Sorry, I wasn't clear. What is explainable is my testing of HDMI out, and the GO app latency shows identical latency. I mean exactly the same, all the time. Logically, I would think the iPad would be slower than HDMI, given the USB transport of the compressed video to the iPad, which is and extra step. But when the Ipad has 243ms latency, so did the goggles connected to the HDMI port.
So I have one video path, a cable from the HDMI port directly to HDMI googles. Apparently the video is decoded in the DJI module on the back of the RC, via hardware, then sent to the google via HDMI cable. The second path is through the USB cable from the RC into the iPad. I presume the decoding is done in hardware of the iPad (hardware decoding is enabled in the app). So I have two separate video paths that use different communications technology (USB and HDMI) sending the same video stream (a video stream of a stop watch), one being decoded by the RC HDMI module (I presume) and the other being decoded by the iPad in hardware. How can both independent video paths delivery exactly the same latency, down to 1 thousandth of a second, exactly ? See my test photo above. I ran about 6 tests, and every time both displays have the exact same latency delay. Each test resulted in different delay times, varying from 180 to 250ms, but the delay between both displays is identical !. Perplexing.
Last edited:
Similar threads
