A few words about RockSteady and real resolution.

In an attempt to select the optimal recording mode, made simple measurements to assess the resolution of the visible area of the matrix after turning on the DEWARP and RockSteady modes.
Measurements were taken relative to the mode without DEWARP and RockSteady.

As a result, I received the following approximate results:

DEWARP ON                ROCKSTEADY OFF        3014x1695
DEWARP OFF        ROCKSTEADY ON                2828x1591
DEWARP ON                ROCKSTEADY ON                2519x1417

If you believe the results, the ROCKSTEADY ON mode in the current implementation makes sense only for 2.7K mode. Given the same bitrate, 2.7K mode should give a better picture than a 2519x1417 crop scaled to 4K.

In order for the camera to be called 4K, it is necessary to implement stabilization at which the real resolution would be more than 2.7K.

In addition to RockSteady, I would like to see the following stabilization modes:

1. The minimum mode due to excess pixels relative to the 4000x3000 mode (4K mode without crop), let's call it “RockHands”.
This mode will be sufficient for static handheld shooting (the operator is stationary).

2. The average mode with crop at the level of the DEWARP ON mode (~ 3000x1680), let's call it “RockClear”.
This mode is well suited for accurate moving with maximum image quality.

3. RockSteady - the only wish is to try to slightly reduce the crop with DEWARP ON to at least correspond to the 2.7K mode.

Alexey

It would be great if there would be an option to use as a Dash Cam  so it would start loop recording when USB power is pressent ! i think is should be a great easy to implement feature !

id4alexey Posted at 1-20 08:14
A few words about RockSteady and real resolution.

In an attempt to select the optimal recording mode, made simple measurements to assess the resolution of the visible area of the matrix after turning on the DEWARP and RockSteady modes.

My question is did you test it on 2.7K to see if it cut the resolution from there?

1. Fix rock steady and take it back to what it was before they ruined it.

2. Remove wifi and location requirements for using a camera on a phone.

Interceptor_Alpha Posted at 1-22 17:01
My question is did you test it on 2.7K to see if it cut the resolution from there?

I made quick test. The difference is minimal. There are the crops: 2.7K and 4K . 2.7K scaled to 4K. After scaling I see minimal difference in the angle of view - 2.7K some wider.

Yup rocksteady is now just terrible since last update. Fix this please.

id4alexey Posted at 1-20 08:14
A few words about RockSteady and real resolution.

In an attempt to select the optimal recording mode, made simple measurements to assess the resolution of the visible area of the matrix after turning on the DEWARP and RockSteady modes.

Interesting comparison of resolution, I’m intrigued - how did you obtain measurements of the cropped resolution ?

MKosmo Posted at 1-26 07:54
Interesting comparison of resolution, I’m intrigued - how did you obtain measurements of the cropped resolution ?

I used simply method:
1. Supose the camera without RS & DW has full resolution 3,840 x 2,160.
2. Fix the camera against any wall.      
2. Look through the camera on the wall, and mark on the wall the left and the right boder of the image.
3. Change RS & DW mode to ON and mark the boders on the wall again.
4. Measure distance between the marks and calculate the resolution:
    RES=3,840 * DISTrsdw=on / DISTrsdw=off
5. So we define the resolution of part of the video sensor, which is using for video coding. Afte that, as I think, the result frame is coding to  4K or 2.7K depending of the choosing video mode

id4alexey Posted at 1-28 06:11
I used simply method:
1. Supose the camera without RS & DW has full resolution 3,840 x 2,160.
2. Fix the camera against any wall.      

Ahh I see, using scaling - thanks for answer.

It is a bit scary how much resolution is being lost though.

It does explain when comparing OA to OP how you can see the loss of detail. ( considering it’s using the same sensor )

so guess  makinga timer for video is not happening. It's Feb 2020 the CES show came and went, May is around the corner, hoping this is no the End of a OA1 and seeing a OA2 come with the features we asked. some other company did that and well uh. guess my camera still records and that will be my FATE!.

Could we expect some software updates soon?

I am happy with the Osmo Action. Purchased it in November, and returned the Hero 7. Way better usability.

2 options would be great for me:

- an option for smaller file size, better compression or more compression. Sometimes I want to save space for situations where max quality is not important. The files from the DJI Osmo Action can be 8 times larger than my Sony RX100 M7 files. at 1080 60fps. This is one of my main gripes. at least have 2 quality options. Even if storage is cheaper these days, sometimes smaller size is important.

- while freediving in rough water, with the Osmo in the housing it is hard to read the custom modes FPS. I wish the text in the C modes was LARGER, easier to read in an instant, underwater or in the waves, with a lot of sun. Legibility has to be improved for custom modes, as i can't access the touchscreen when the Osmo is in the housing.

These 2 are my main issues. I hope they are addressed.

Having less sharpening would be good too.

id4alexey Posted at 1-20 08:14
A few words about RockSteady and real resolution.

In an attempt to select the optimal recording mode, made simple measurements to assess the resolution of the visible area of the matrix after turning on the DEWARP and RockSteady modes.

That is exactly what I have been saying from the beginning. I don't think Gopro is any better either. For me, it is hard to be ok with a seriously upscaled, "pseudo" 4K resolution.

And it gets worse. Per my rough calculations (feel free to correct me if I am wrong), to get the level of digital stabilization people want and real 4K resolution, you need at least a 16MP sensor. And with a similar lens setup, it has to be 1/1.7" or bigger so that it does not hit the diffraction limit. So basically you need the next-gen action camera to be able to actually do what this gen promises.

I totally agree with your stabilization level suggestions. I believe there is enough playroom on the sensor for some basic stabilization (without dipping below 4K), and that should be the first level. I also would rather see the highest level labelled correctly, as 2.7K maybe, instead of upscaling it and listing it as 4K.

MKosmo Posted at 1-28 09:10
Ahh I see, using scaling - thanks for answer.

It is a bit scary how much resolution is being lost though.

Well, the quality difference is huge, but there are other factors at work there as well. OP is brighter (by two stops if my memory serves me right) which is a factor, it's fov is much smaller (translates into better definition), it has a narrower depth of field (slightly better compression) etc. etc.

But if I have to chose, having mechanically stabilized "real" 4K resolution would be at the top of my list   

I can understand the ACTION guys wanting RS to be as god as it can get, personally i am planning to get a gimbal and never use RS

stephjant Posted at 2-7 12:27
The processor has limited capacity.  I am wondering if the improvement to lag was done at the expense of something else, and RS in particular which I think is worse than in the previous firmware version.  I personally dont care about the lag but I do need excellent RS performance for my action shots taken at speed and often over rough terrain

Exactly what I think/thougt.
I was pretty p1ssed off to realize today how the new Firmware destroyed the RS...

Bring back the older better version of rocksteady.  The screen lag was not really that bad an issue, but now that that is "fixed" rocksteady is not as good as it was when I first got the camera.  The "gimbal-like" stabilization that I bought the camera for does not exist anymore.
Better yet, make it so the user can choose.  Have levels of rocksteady.  one being this current version of rocksteady that eliminates screen lag for things that don't need that much stabilization, and another setting that makes it as steady as it was originally with the only minor issue is it bringing back the screen lag.  Best of both worlds, make everyone happy.

So its like DJI  hired one Soul to take care of it.  way to go  dejing

I have had the Osmo Action for a few months. I recently got a GP Hero 8 as well.
The file size of 60fps videos on the Osmo Action are huge.
I would like an option for lower bitrate/smaller files. Below the comparisons on how much footage I can put on a Hero 8 vs the Osmo Action at different res and fps.

Most of my content is output at 1080p, I am interested in 2.7k and 1080p (by the way does the battery at 2.7K lasts longer than when shooting at 4K? and 1080 vs 2.7K?)

I noticed the file size at 1080p 60 fps and 2.7K 60fps is huge.
I would be happy with a lower bitrate option to save drive and card space.

I put the same newly formatted 128 GB memory card and I checked the recording time on both the GPH8 and the DJI OA.

Any chance that DJI may provide more compressed/smaller files at 1080p and 2.7K 60 fps?
GoPro Hero 8 files at 4K 60fps are smaller than Osmo Action at 1080p 60fps.
Osmo Action at 2.7K 60fps does make sense, 167 min, same as 4K.


That would be amazing.

GoPro Hero 8 128 GB 1080p
30fps 7h 36 m 456 min
60 fps 5h 04 m  304min
DJI Osmo Action 128 GB card 1080p
30 fps  7h 59 min 479 min
60fps  3h 29 min 209 min

GoPro Hero 8 128 GB 2.7K
30fps 5h13m 313min
60fps 5h 13m 313 min
DJI Osmo Action 128 GB card 2.7K
30fps 3h29m 209 min
60fps 2h47m 167min

GoPro Hero 8 128 GB 4K
30fps 5h 13 min 313 min
60fps 4h 09m 249 min
DJI Osmo Action 128 GB card 4K
30fps  2h47m 167 min
60 fps 2h47m 167 min

I have had the Osmo Action for a few months. I recently got a GP Hero 8 as well.
The file size of 60fps videos on the Osmo Action are huge.
I would like an option for lower bitrate/smaller files. Below the comparisons on how much footage I can put on a Hero 8 vs the Osmo Action at different res and fps.

Most of my content is output at 1080p, I am interested in 2.7k and 1080p (by the way does the battery at 2.7K lasts longer than when shooting at 4K? and 1080 vs 2.7K?)

I noticed the file size at 1080p 60 fps and 2.7K 60fps is huge.
I would be happy with a lower bitrate option to save drive and card space.

I put the same newly formatted 128 GB memory card and I checked the recording time on both the GPH8 and the DJI OA.

Any chance that DJI may provide more compressed/smaller files at 1080p and 2.7K 60 fps?
GoPro Hero 8 files at 4K 60fps are smaller than Osmo Action at 1080p 60fps.
Osmo Action at 2.7K 60fps does make sense, 167 min, same as 4K.


That would be amazing.

GoPro Hero 8 128 GB 1080p
30fps 7h 36 m 456 min
60 fps 5h 04 m  304min
DJI Osmo Action 128 GB card 1080p
30 fps  7h 59 min 479 min
60fps  3h 29 min 209 min

GoPro Hero 8 128 GB 2.7K
30fps 5h13m 313min
60fps 5h 13m 313 min
DJI Osmo Action 128 GB card 2.7K
30fps 3h29m 209 min
60fps 2h47m 167min

GoPro Hero 8 128 GB 4K
30fps 5h 13 min 313 min
60fps 4h 09m 249 min
DJI Osmo Action 128 GB card 4K
30fps  2h47m 167 min
60 fps 2h47m 167 min

i dont know if anyone knows, but when i walk the camera is really shay in rocksteady.. i need to run to make beter results.. can you fix that so also when i walk i ave a nice vieo? or is that already in the settings and i dont know about it.. rock steadey isnt good while walking.

JanCus Posted at 2-14 05:31
i dont know if anyone knows, but when i walk the camera is really shay in rocksteady.. i need to run to make beter results.. can you fix that so also when i walk i ave a nice vieo? or is that already in the settings and i dont know about it.. rock steadey isnt good while walking.

All cameras shake when walking or jogging, and they all need a gimbal to fix the worst issues.
Osmo Action with Rocksteady is as good as gopro 8 or even 9 in some cases.