Does logic not apply here?

[flying_outcast]

I noticed it before, higher bitrates for apparently lower res (you'd think it would be the other way around to keep up). Never figured it out. I wanted 1080p streaming and caching, but was told someplace that the live feed is smoother at longer ranges with 720 and I decided that this reliability was more important, but never tested it.


I also remember seeing someplace where you could select 2.4 ghz or 5.8 ghz, and manually set the channel, but now I can't seem to find it (maybe firmware update, or maybe I'm just not looking in the right place anymore).

You wouldn't just think it is. It's a fact that more data needs bigger pipeline, just basic physics. (Not the only thing that has been delivered upside down from dji.

And about the channel and mode settings, they look similar as the settings in android go 4 but there you can press the buttons they have put there...

But not in ios, even tough the buttons are there, they just don't like to be touched. It's fully functional in android.

 

[geigy]

Click on HD in the main screen. Here's a thread:
Mavic Image Transmission Settings

 

[geigy]

Did some more tests. Using the Android version of DJI Go 4, I recorded 15 seconds of cached video in HD, Normal. and Smooth modes. The Mavic was handheld indoors for this test.

The HD file was 1920x1080@30fps, and the other two files were 1280x720@30fps. All three used Apple's AVC1/H264 codec. The kicker is that the bit rate for all three files was just under 12 Mb/s. Files were analyzed using mediainfo on Win 10.

So, logic is not really prevailing yet. Does the cached video file match what was received from the Mavic? Probably. Why would the HD bitrate be the same as the 720p files? Smooth is supposed to match the capability of the DJI goggles, 1080p @60fps.

Next stop: install 0700 and see if they've changed anything.

Anyone else chime in - I have not got this figured out yet at all.

 

[alexcannas]

The bitrate is not directly related to the resolution. As an example, when recording in 4k the maximum bitrate is 60Mbps. That is the same in 2.7k mode. So the file size is the same. The difference obviously is in the fact that the compression ratio used for 4k is much higher than 2.7k.

That being said, I think the OccuSync system is probably using a variable bitrate. I did some checks on mine, in HD mode it would go upto 12Mbps, but if left there the bitrate dropped to 1.7Mbps. In regular mode, the maximum bitrate reported was about 5.5 and in smooth about 7.5Mbps. Again these were the maximums and then it dropped. If you move the camera or the Mavic then the rate goes up again.

Btw: Normal is now called Regular in the settings. I just don't get it why they do not simply name them 720p30, 1080p30 and 720p60 or whatever is should be. It seems like the software team is changing every month :S

 

[geigy]

Alex,

Thanks - we are on the same wavelength here, but I'm not sure I completely get what you are saying. You are seeing a higher bitrate for HD, whereas I, and flying outcast are seeing a lower bitrate. Probably I need to get to the bottom of that before figuring this out.

In the meantime, does it make sense to you that HD would be using significantly less bandwidth than normal 720p mode? In essence, if I understand you correctly, the HD mode might have a higher compression (lower quality) setting, even though the rez is higher.

The problem I still have is that the bitrate displayed in DJI Go varies depending on the mode, but the cached video files show nearly the same bitrate for all 3 modes.

I did the experiment with the 0700 FW. Same result. All cached files had the same bitrate according to the mediainfo utility, in spite of the fact that the HD file is 1080p. Mediainfo output enclosed.

 

[Mission North]

Video links are a complex subject. There is always a compromise between bandwidth, resolution, robustness, artifacts and latency. Most compression systems look at differences between several frames of video. In order to do that there has to be a delay in the encoding system to see what is ahead. If you reduce this delay by compressing fewer frames into the data you will get worse artifacts or require more bandwidth. By including more data redundancy and error checking you can get a more robust signal, but again you would need higher bandwidth.

In cases like DVD and Blu-Ray the latency is irrelevant, so you can spend days encoding an hour's worth of video. In broadcast television a few seconds won't make much difference to the viewer. In live television production, any more than a few frames means that the linked cameras won't cut with the cabled ones. And the sound department have to delay the audio on radio linked interviews to lip sync with the pictures - usually done by sending the roving presenter's mic via the camera's radio link. Put on the DJI Goggles and any video delay is going to be a problem.

Basically, for a known bandwidth limit: low delay, high quality, robust signal - choose any two.

 

[alexcannas]

Alex,

Thanks - we are on the same wavelength here, but I'm not sure I completely get what you are saying. You are seeing a higher bitrate for HD, whereas I, and flying outcast are seeing a lower bitrate. Probably I need to get to the bottom of that before figuring this out.

In the meantime, does it make sense to you that HD would be using significantly less bandwidth than normal 720p mode? In essence, if I understand you correctly, the HD mode might have a higher compression (lower quality) setting, even though the rez is higher.

The problem I still have is that the bitrate displayed in DJI Go varies depending on the mode, but the cached video files show nearly the same bitrate for all 3 modes.

I did the experiment with the 0700 FW. Same result. All cached files had the same bitrate according to the mediainfo utility, in spite of the fact that the HD file is 1080p. Mediainfo output enclosed.

To answer your question, I am not saying that 1080p should be lower bitrate. But it can be depending on the scene. If the scene is simple, then the bitrate drops. Also the bitrate (as you see in the app) is variable. What you see reported on Windows is an average bitrate. What I noticed (which another member noticed as well) is that if the scene is static the bitrate drops.

Also the available transmission bandwidth changes constantly. The encoder changes the bitrate to match what is available or you will get dropouts. At 1080p it will compress more to fit within the same budget with 720p.

The point is that there are more important factors affecting the bitrate, other than the resolution. I think this is part of what OccuSync does, manage a variable bitrate encoder to transmit over a constantly monitored transmission channel. So you should not read too much into it, in terms of resolution.

 

[Strafe1]

The bitrate is not directly related to the resolution. As an example, when recording in 4k the maximum bitrate is 60Mbps. That is the same in 2.7k mode. So the file size is the same.

Am I misunderstanding this.....I thought bitrate for 2.7k was 45 Mbps. Didn't know if it made a difference in the thought process.

 

[alexcannas]

Am I misunderstanding this.....I thought bitrate for 2.7k was 45 Mbps. Didn't know if it made a difference in the thought process.

You are right my bad... 60Mbps is 4k and 45Mbps is 2.7k. So my example was wrong

 

[TheLightSpeed!]

I had this exact issue or concern too. In the old Go app we all would select the transmission rate of 4mb to maximize distance. In this version of the app, I noticed that HD also lowers the transmission over "normal".

I had that same issue a few times where the video lagged like crazy rendering it unusable. Of course it hasn't always been like this, just recently.

 

[Robbyg]

Using version .0400 under IOS I get the following.

Normal Mode - 20Mhz = VBR(11.21Mbps - 12Mbps)
HD Mode - 20Mhz = VBR (1.06Mbps - 2.80Mbps)

Normal Mode - 10Mhz = VBR(11.21Mbps - 12Mbps)
HD Mode - 10Mhz = VBR (1.06Mbps - 2.80Mbps)

In the past I have tried to figure this out and came to the conclusion that either DJI is displaying the Bit Rates for each mode in reverse or I am missing something completely.

Why does the Transmission bandwidth have no effect, I have no clue! I wish to God I had a spectrum analyzer that worked up to 3Ghz so I could see if the BW really changes when set to 20Mhz.

My conclusion based on very little though or effort is that it is possible that DJI is reserving bandwidth slices for the three possible devices that could be connected at the same time.
So maybe the system is not dynamic in nature. Makes sense since you could be in mid flight and switch on two sets of goggles. One would think that the video would all be on the same carrier frequency but maybe they have dedicated slots within the bandwidth for each device if 1080 is menu selected. Maybe its because each device can independently change to 720 dynamical.

Maybe when you select 720 as the default it uses a common slice of bandwidth for all devices and therefore it can increase the bit rate.

Can the goggles receive 1080 while the phone or tablet is receiving 720? That would be a possible indicator.

As to why the recorded video is affected and cannot be recorded in 4K while in HD mode is a mystery to me. I can only assume its an internal video processing limitation.

This is a nice thread.

Rob

 

[geigy]

I think the displayed bitrates are wrong. From my tests the displayed bandwidth does not reflect the bitrate of the video stream in the cached files, which I think represent reality. The bitrate for I used a static scene in my 0700 tests. The most surprising result is that the bitrate does not change in HD mode.

The HD image is 1040p, and has more detail in the midtones, and the shadows seem to more aggressively smoothed, which could explain the constant bitrate. It may be that the CPU requirements for the extra processing in HD limit the Mavic's ability to also handle a 4K image and higher framerates.

Lots of guessing here, but my take-home is that the Mbps numbers displayed in the app are wrong, and Regular and Smooth is the same resolution and fps, at least under the conditions I tested.