DJI challenging GoPro?

[JurijTurnsek]

11 minutes ago, Anaconda_ said:

Like this? Skip to like 3 minutes for a great preview of exactly what it's doing to stabilise. 

 

Yes, but without the added hardware (sensors are already built in).

[Anaconda_]

13 minutes ago, kye said:

You quoted me out of context  

Ah sorry, I misread and didn't want the delay to go under the radar - it's a biggy.

[Video Hummus]

40 minutes ago, kye said:

I disagree.  In the case of the gyro stabilisation, what you want is for the distinctive pixels in one frame to be as close as possible to their location in the last frame.  It's a frame matching problem.

I’m pretty positive that is not how it works with moving pictures (or, video) It’s a horizon and framing problem. You want to keep the horizon and angular velocities from changing as much as possible. You don’t have to look at the image data. That’s how stabilizer tank cannon platforms work. Anybody remember the famous leopard tank demonstration with the flask of beer on the end of the barrel?

of course doing this process off camera with a faster cpu and multiple passes is going to yield better results. I invite Olympus or whomever to standardized some file format for it and software to take advantage of it.

what your are suggesting in post can easily be done with dedicated hardware in camera and in real-time, which is still better than having to do it in post. 

40 minutes ago, kye said:

Plus, think about how well IBIS works on a 16mm lens - pretty darn well.  Think about how well it works on a 400mm lens - less well.  Stabilisation in post is just fine at 1000mm  

Yeah that’s basic physics. The farther your focal magnification the more amplified movements become. OIS and EIS can help in both situations.

The camera industry would do well to incorporate some features from the action camera market. There is a physical limit to how small a gimbal can be. 

40 minutes ago, kye said:

In image analysis based stabilisation, the computer looks for distinctive pixels in each frame, works out the movement, then works out how to change frame 2 to better match frame 1, knowing the overall direction of movement from frame 1 to frame 1000.  This method is about matching frames by analysing the frames directly.

That is interpolating acceleration and gyro data from the image. Which will always be worse than actually reading and using that data in real-time or in post from dedicated hardware.

43 minutes ago, Anaconda_ said:

Like this? Skip to like 3 minutes for a great preview of exactly what it's doing to stabilise. 

 

3 hours ago, kye said:

Yeah, this is a perfect example of what I’m talking about. Now, put the hardware in camera and have dedicated CPU correct the image in real-time and you have your gimbal footage without a gimbal. Also works in low light because you aren’t relying on image data.

It is by-far the better method. And what’s is stopping you to do more post stabilization after the fact as well?

[MeanRevert]

4 hours ago, JurijTurnsek said:

Where?

Also, I hope GoPro and DJI switch to the 48mpx sensor for even better EIS (you are not getting true 4K with EIS from the cropped 12mpx sensor). Osmo Pocket is still a more versatile film-making tool unless you need to go near or under water.

https://9to5toys.com/2019/05/14/dji-osmo-pocket-4k/

Sorry, looks like it's expired. It's a start though.

[Jermaine]

13 hours ago, MeanRevert said:

Ok, I lied. This dude has the best review out there. I found the side-by-side, the 2nd video, especially useful.

 

 

Slightly off topic but does anyone know what font is used in this thumbnail, I've tried to find a similar style but this is the exact one I want

[thephoenix]

1 hour ago, Jermaine said:

Slightly off topic but does anyone know what font is used in this thumbnail, I've tried to find a similar style but this is the exact one I want

seems maybe stupid but did you try asking the guy that did the video ?

[MeanRevert]

Now that the release has come and gone and most folks are underwhelmed, I think it's time for Sony to get going on their FDR X3000 successor before the Hero 8 comes out in September.

[sanveer]

Maybe we should Start a Thread, to Tell DJI what should be fixed via firmware. The Rocksteady is way better than anything else out there. But since it can't be mixed with HDR (the dual exposure one), the Cineline profile should be improved along with options to control (atleast) sharpening, contrast and colour.

I am sure there are many other things that can be added or improved via firmware. 

[Cinegain]

22 hours ago, Jermaine said:

Slightly off topic but does anyone know what font is used in this thumbnail, I've tried to find a similar style but this is the exact one I want

You'll find stuff like it here somewhere: https://www.dafont.com/theme.php?cat=107&text=ACTION+REVIEW

Alternatively... know your resources: https://www.myfonts.com/WhatTheFont/

[Emanuel]

No much room for many doubts now if any ever, Osmo Pocket rules over there despite the complement of each other in one or another feature:

 

 

 

[sanveer]

4 hours ago, Emanuel said:

No much room for many doubts now if any ever, Osmo Pocket rules over there despite the complement of each other in one or another feature:

 

 

 

Thanks to you, I watched the entire video, in higher res than I usually do. I am guessing the increase in Dynamic Range is mostly due to the difference between f2.0 and f2.8 on their lenses. There may be also slight issues with how the Cineline codec has been set on the Osmo Action. There is definitely a noticeable difference, and it is visible in most lighting situations (even daylight).

That's why DJI needs to start getting out their firmware updates to sort things out. Also I don't know why nobody gets an action out that has an f2.0 lens. 

[kye]

22 hours ago, MeanRevert said:

Now that the release has come and gone and most folks are underwhelmed, I think it's time for Sony to get going on their FDR X3000 successor before the Hero 8 comes out in September.

It would be nice, and it would be a long time coming too.  Unfortunately I can't wait for that because I have a trip in mid-September that I need it for, so in a way I hope I don't buy an X3000 just before they release the updated model, but anyway.

In case anyone isn't familiar with EIS vs OIS in low-light, here's why EIS is fundamentally flawed when the sun goes down...  (link goes to the specific time in the video)

Also, the X3000 seems to have more DR than the others and has the higher bitrate 100Mbps so the files are less brittle in post.... (link to specific time in video)

I'd love to see a direct comparison video between the Hero 7 (with current firmware) vs Osmo Action vs Sony X3000, but because the Sony is old it seems to no longer get included.

My logic is that if the Osmo Action isn't clearly better IQ than the GoPro, and the X3000 has some serious advantages over the GoPro then the X3000 is still a very strong contender.

If you shoot low-light or want to use it more like a normal camera then it's still in first place, especially with things like a tripod mount, 3.5mm mic port, etc.

[MeanRevert]

2 hours ago, kye said:

I'd love to see a direct comparison video between the Hero 7 (with current firmware) vs Osmo Action vs Sony X3000, but because the Sony is old it seems to no longer get included.

 

[kye]

Yeah, I watched that one a few times, I guess I'm more referencing that at some points in the video he talks about the three of them, but then at other points it's just the two.  The Hero 7 video was great because the coverage was much more even between cameras.

There are some aspects that I can't compare the X3000 and the Osmo directly, but have to compare the Osmo to the GoPro, then in the GoPro review compare the GoPro to the X3000.

But thanks  

[Anaconda_]

 

 

[sanveer]

There seem to opinions on both sides. Like this one:

 

 

[MeanRevert]

Dude in the video says Osmo has more DR but if you look at the side-by-sides and the clouds, he's clearly wrong. Haha.

[sanveer]

18 minutes ago, MeanRevert said:

Dude in the video says Osmo has more DR but if you look at the side-by-sides and the clouds, he's clearly wrong. Haha.

I agree about the clouds, but in the grass the GoPro is more darker (which makes me wonder why DJI didn't try snd make an f2.0 lens action camera, since the brighter lens would have added advantages in super high speed as well as low light).

I am suddenly looking forward to the Philip Bloom promised testing. Hopefully he compares them to the absolute limit, in all departments, that matter to filmmakers (shooting and post). 

[kye]

 

5 hours ago, sanveer said:

I agree about the clouds, but in the grass the GoPro is more darker (which makes me wonder why DJI didn't try snd make an f2.0 lens action camera, since the brighter lens would have added advantages in super high speed as well as low light).

I am suddenly looking forward to the Philip Bloom promised testing. Hopefully he compares them to the absolute limit, in all departments, that matter to filmmakers (shooting and post). 

If you look at the tests from the iPhonedo Hero 7 review showing the waveforms you'll notice that different cameras expose slightly differently (where they put the darkest part of the histogram) and also have different DR (how stretched the histogram is) and also have different clipping points (where the highest values in the histogram are).  

How fast the lens is will mean that a camera can get a certain exposure level with a lower ISO, or faster shutter speed at base ISO, but that will only matter in low-light.  The exposure levels aren't a factor of the lens, they are a factor of the cameras auto-exposure function and where it tries to expose.  Action cameras are typically used outdoors in bright light and considering they have a very wide angle the auto-exposure algorithm will probably be tuned for when the sun is in frame and to just let it clip.  If you use auto-exposure on a normal camera and put the sun in frame they often darken the frame radically such that your actors are silhouettes so that they're not blowing out the majority of the sky, which is a different approach because they're cameras for different things.

Ultimately what you want is something with the shadows raised up a little so that there's a bit more information there, you want the DR to be kind of squashed to get more in, and you want the clipping point to be as high as possible so you clip the least.  You also then want the highest bit-depth (although they're all 8-bit because they're action cameras) and the highest bit-rate so you can push and pull the image around more without breaking it.

In terms of PBs video, it will be very interesting but it's a pity he didn't get more cameras - the X3000 of course but also Yi 4K and RX0ii as well.  I get it when an extreme sports person only looks at action cameras, but PB is a pro cinematographer who should cast the net a bit wider.

[webrunner5]

Yeah, but I would imagine the RXO II would slaughter the others overall on imagine quality, low light, DOF, on and on.