I will proceed in this direction as well, but I still think the best way is to put it in webcam mode to find the right closed focus at about 50 cm, the correct distance in my opinion for someone using it as a vlogger, 35 cm is really unnecessary.
DJi claims 30cm minimum but even at 40 it is not properly in focus, in normal use below 40/50 you never go there, so better to be more conservative to avoid blurring in distant subjects if you set well in focus at 30cm.
I will maybe make a video.
How much does the part that su pick up with tweezers tick by when it is about in right focus?
The idea is to design and print in TPU a ring nut that can then be left inside so that calibration can be done without the tweezers.
MarcoR Posted at 11-29 23:26
I will proceed in this direction as well, but I still think the best way is to put it in webcam mode to find the right closed focus at about 50 cm, the correct distance in my opinion for someone using it as a vlogger, 35 cm is really unnecessary.
DJi claims 30cm minimum but even at 40 it is not properly in focus, in normal use below 40/50 you never go there, so better to be more conservative to avoid blurring in distant subjects if you set well in focus at 30cm.
I will maybe make a video.
Focus it at 60 cm should give desent focus at 30cm and all the way to infinity if you check my bigger post further up here
johansenfoto Posted at 11-30 05:01
Focus it at 60 cm should give desent focus at 30cm and all the way to infinity if you check my bigger post further up here
If you set the focus at 60cm, so regarding my tests you will get also blurry image at the bigger distance.
I printed out a test sheet to determine the focus. Considering that the lighting in the room is not the best, in my opinion the focus looks good. (date 9/2022 standart combo)
Ilya187 Posted at 12-1 22:06
I printed out a test sheet to determine the focus. Considering that the lighting in the room is not the best, in my opinion the focus looks good. (date 9/2022 standart combo)
Wow, those are really good - you lucky bastard! But did you use a letter or tabloid size paper? That printout seems to be bigger. I have printed a smaller version which is letter size (8.5" x 11"). Can you share a link to the chart you used? Thanks.
Ilya187 Posted at 12-1 22:06
I printed out a test sheet to determine the focus. Considering that the lighting in the room is not the best, in my opinion the focus looks good. (date 9/2022 standart combo)
osmonauta Posted at 12-1 23:35
Wow, those are really good - you lucky bastard! But did you use a letter or tabloid size paper? That printout seems to be bigger. I have printed a smaller version which is letter size (8.5" x 11"). Can you share a link to the chart you used? Thanks.
I printed a matte sheet in A3 format, as recommended. But i will try do as you said.
Thnks
MarcoR Posted at 12-1 23:38
Imho it is not good but very good!
You still have to try taking pictures, in videos OA3 is digitally pumped in detail.
And this is precisely what is missing here compared to the GP, to choose the sharpness, because on contrasting scenes, as someone has already suggested, the AO3 is digitally over-sharpened (background). Well, it's only a sw thing, so one time later fw...
StevoB Posted at 12-1 23:46
And this is precisely what is missing here compared to the GP, to choose the sharpness, because on contrasting scenes, as someone has already suggested, the AO3 is digitally over-sharpened (background). Well, it's only a sw thing, so one time later fw...
DJi listens to user requests in part, has enabled 10Bit but no ability to set sharpness and no gyro data.
We have to make do with what comes in.
MarcoR Posted at 12-2 00:00
DJi listens to user requests in part, has enabled 10Bit but no ability to set sharpness and no gyro data.
We have to make do with what comes in.
And we can be happy that the new fw are free. The time will come when we will have to pay extra for "improvements", or you will buy a bare body with max 2k and you will pay extra for 4k, 10bit, extra stabilization, voice control, gyrodata... like buying car. Then you will have everything that another manufacturer has, don't worry .....an economist must have plus numbers, and more and more...
Ilya187 Posted at 12-1 22:06
I printed out a test sheet to determine the focus. Considering that the lighting in the room is not the best, in my opinion the focus looks good. (date 9/2022 standart combo)
There is a sure reason why we ask for saved gyro data in the movies.
I did some testing with the new DJI O3, which is basically the same cam installed in the Avata.
I installed it on a racer that I intentionally sent out of tune by exaggerating with the derivative on the roll in order to generate those fast oscillations.
Gyroflow stabilizes in post production much better than RockSteady in realtime (it is patently easy to understand why), which in the same situation generates small abberations on the extreme sides of the footage, I don't have the video with RockSteady here but I will post that as well.
In addition RockSteady generates "too stiff" stabilization, whereas in Gyroflow you can set the amount of stabilization, those violent roll commands RockSteady doesn't dampen them well, look instead at what Gyroflow manages to do.
That is why we have been asking for gyro data on the Osmo Action 3 for months, so that we can have a better result, or at least have a choice.
DJI please, give us this gift for Christmas, you've never done it on the DJI FPV, at least do it on your flagship action cam, damn!
Rockel83 Posted at 12-2 05:57
So, here we finally have these "internal parameters" where DJI is this verry mysterious about for 3 months already
I actually have to say that I find the ability to rotate the lens afterwards and adjust the focus actually much, much nicer than the fixed focus like OA1 and OA2. I'm even considering buying a second OA3 and setting it to focus at about 10-15cm, so I can take good macro photos with it, or just deliver better shots at close range.
The turtle here in the photo, for example, is not in focus, because much too close. Taken with the O2.
Dave Harry did this with his old Yi 4k. He put markers on the camera body for distance and macro. (I think he also had a close focus marker) meant that he could dial it in. The macro shots that he did with it were amazing, but he had to get the lens right on top of the subject!
JustName Posted at 12-2 07:01
I actually have to say that I find the ability to rotate the lens afterwards and adjust the focus actually much, much nicer than the fixed focus like OA1 and OA2. I'm even considering buying a second OA3 and setting it to focus at about 10-15cm, so I can take good macro photos with it, or just deliver better shots at close range.
The turtle here in the photo, for example, is not in focus, because much too close. Taken with the O2.
I also checked the other videos on your channel, they seem to look pretty good now. I was thinking a sec. trying the same you did before returning mine. But I found it a bit much of a risk for €300 hahaha
So great job!
Only too bad DJI couldn't had arranged something like this. This is not a good advertisement for DJI...
Instead they've choosen to hold people on the line with their "internal parameters" and let them wait for firmware updates. There're even people sending in their camera's, hoping their next cam will finally be in focus.
I was planning to RMA my unit as many of you have done. But unfortunately due to poor lighting in my home, and overcast conditions outdoors I was only able to make minimum judgement.
I feel that it is certainly out of focus, for example the DJI OSMO ACTION 3 text on the box is certainly not sharp even for this lighting.
Please tell me am I simply judging too harshly or is this indeed not in focus:
My test https://www.youtube.com/watch?v=dc1jTI0_aAA
Certainly compared to another user's images at 60cm, my unit looks like the original RMA unit they posted on the left:
Other example at 60cm
https://cdn.discordapp.com/attachments/645883410803458067/1051717103482314762/HIrX36k.png
I don't think batch has anything to do with quality. My first unit was 08/2022 and my second (replaced) unit is also 08/2022. My second unit had nearly identical image quality to the first one (meaning, not so good close focus). However, with the latest 11/30/2022 firmware (v01.02.10.10) the quality significantly improved.
This is a crop from a 2.7k standard video recording (not 10-bit) from about 20cm. No sharpening. Crop was saved at 95% JPG compression. The focus chart is printed on A3 size paper.
sonoyuki Posted at 12-11 21:53
I was planning to RMA my unit as many of you have done. But unfortunately due to poor lighting in my home, and overcast conditions outdoors I was only able to make minimum judgement.
I feel that it is certainly out of focus, for example the DJI OSMO ACTION 3 text on the box is certainly not sharp even for this lighting.
Please tell me am I simply judging too harshly or is this indeed not in focus:
Your box is the standard box. The other user's box you posted is the combo box. I believe the combo box is bigger. The outlet on the wall certainly does seem to look sharp.
You could use the focus chart for test if you want.
osmonauta Posted at 12-12 14:21
I don't think batch has anything to do with quality. My first unit was 08/2022 and my second (replaced) unit is also 08/2022. My second unit had nearly identical image quality to the first one (meaning, not so good close focus). However, with the latest 11/30/2022 firmware (v01.02.10.10) the quality significantly improved.
This is a crop from a 2.7k standard video recording (not 10-bit) from about 20cm. No sharpening. Crop was saved at 95% JPG compression. The focus chart is printed on A3 size paper.
It is possible, but until now I have not come across a person who complained about a unit produced in November, which is also strange....
Yes. Same FW and settings.
And this is just a Screenshot from my Macbook.
I was wondering about the poor the quality. I compared it with my friends OA3 while riding MTB. Same settings. And the difference was noticeable. So i ordered a new OA3. And here we are.
i think its pure luck what you get. The Batch doesn't matter. My friends cam is Batch 09/22 ,too. Like my first cam.
Great, thanks for the feedback.
It's been a matter of luck or bad luck since I opened this post.
The production date matters little, instead it matters how they worked at DJI that day.
I don't think it is a machine that places the lens at the correct distance from the sensor, but perhaps simply some kind of mechanical lock, and the operation will be done manually.
And it only takes a very few degrees of rotation of the lens group to send it out of focus.
The only one in my opinion is to proceed to focus calibration independently, if you ask DJI for repair they don't actually touch the lens assembly, well they replace it with another product that always falls into the red or black roulette.