If you want to do an in-depth analysis in this link you will find some  comparative shots between my DJI Osmo Action 3 with the latest firmware  installed and GoPro HERO 6, taken this morning at the same distance from  the subjects, about 40 cm, and obviously with the action cameras stuck  on the floor.
This is my humble demonstration that the new firmware cannot fix obvious hardware problems with the OA3 lens system.
This archive was also sent to DJI Technical Support, at their explicit request.
This images show two crops made for comparison, I think the problem is quite obvious.

https://drive.google.com/file/d/ ... Zx/view?usp=sharing

---DELETED---

I chatted a lot with technical support, and in my humble opinion they are just beating around the bush, climbing on the mirrors.
Send me this, give it a try, in short, it seems that they treat us like people who can't do to press a button on the action camera leaning on which side to take a shot.
It is not a complicated operation, and if the result is this they should admit the mistake.
The money used to pay for their product was not defective imho.
An experienced photographer friend after analyzing these shots told me "these are just a bad set of lenses, the sensor has nothing".
I have enhanced the qualities of this action cam in some of my FPV videos, but the problem to the focus on some units is evident, useless to go around it.
DJI should initiate a massive recall of the obviously failed units and replace them, not repair them.
When a product is born badly it must be replaced, even if a month has passed since the purchase.

Are you sure that what you are showing isn’t the result of minimum focus distance and not an out of focus lens. You wouldn’t need a focus chart to see how bad that focus is, you’d see that immediately without any comparison when recording any video.

I’m not saying there isn’t an issue with the 3, I’m just saying that this particular test may not be correct.

---DELETED---

David_Harry Posted at 10-15 02:07
Are you sure that what you are showing isn’t the result of minimum focus distance and not an out of focus lens. You wouldn’t need a focus chart to see how bad that focus is, you’d see that immediately without any comparison when recording any video.

I’m not saying there isn’t an issue with the 3, I’m just saying that this particular test may not be correct.

Exactly, so in this video of mine you see it better?
The minimum focus distance is 30cm, as I wrote above they were done at 40, read well my first post.

Do you know why the gopro images are 4000x3000 and the dji ones are 960x720 in the archive you shared?

osmonauta Posted at 10-15 02:36
He did mention that it was taken from 40cm, the minimum being listed as 30.

Also, not sure if you seen this video, but it clearly shows that it’s not minimum focus distance we8re talking about. And the comparison was made by using a cheaper camera.

I know what he mentioned. It’s clear I’m questioning the position of the cameras and that he’s shot below the minimum focus distance.

This looks to me like a cherry picked scenario.

As for Andre’s example. Given that I raised the same issues with the Action 1 back in the day, I’m fully aware of potential focus issues.

Once again, the example in this post is a bad example. If this truly is an example of bad focus and not a minimum focus distance issue. Then you’d have to ask yourself why someone would need a focus chart to see what is clearly an out of focus lens.

This example is not typical of the “slight” softening that some people are getting with their 3, this example is either a completely f…ed lens or a cherry picked, exaggerated example.

I have no problem with people bitching and moaning about issues, in fact, I positively encourage it as it’s the only way some manufacturers will ever learn to do the right thing. However, if you’re gonna bitch and moan then you have to be fair with it and give good “honest” examples to prove a point.

fish sticks Posted at 10-15 02:53
Do you know why the gopro images are 4000x3000 and the dji ones are 960x720 in the archive you shared?

Do you know how to treat an image in DNG (RAW) format?

https://helpx.adobe.com/it/photo ... erter.html#Appendix

fish sticks Posted at 10-15 02:53
Do you know why the gopro images are 4000x3000 and the dji ones are 960x720 in the archive you shared?

If you clicking the DJI DNG-file you're probably seeing the embedded lowres jpg-thumbnail file. You need to open the DNG-file in a program like Photoshop or Lightroom to see the full resolution and quality.

David_Harry Posted at 10-15 02:58
I know what he mentioned. It’s clear I’m questioning the position of the cameras and that he’s shot below the minimum focus distance.

This looks to me like a cherry picked scenario.

If I were the only one complaining I'd say I'm an incompetent fool.
Then I'm in good company apparently, the web is already full of reports similar to mine.
Do not try to defend the indefensible, it is a waste of time.
Some of these cameras have this evident flaw, and as I have already written several times it is possible that DJI had to compromise with the lens train (7 I think) to have a better focus on distant subjects.
But then they don't have to write in the specs that the minimum focal length is 30 cm because it isn't.
If you want a test like this i can do it without any problem, it would just be a confirmation because there are those who have already done so.

MarcoR Posted at 10-15 03:15
If I were the only one complaining I'd say I'm an incompetent fool.
Then I'm in good company apparently, the web is already full of reports similar to mine.
Do not try to defend the indefensible, it is a waste of time.

A fool is still a fool whether in the company of other fools or not…

Plus, you don’t make a good case for your claims when you’re avoiding the questions being asked and when you don’t supply the full details of your testing.

David_Harry Posted at 10-15 03:35
A fool is still a fool whether in the company of other fools or not…

Plus, you don’t make a good case for your claims when you’re avoiding the questions being asked and when you don’t supply the full details of your testing.

Perfect David, then I'm your worker: tell me which tests you want to do and I'll do it.
Tell me what and how to do it and I'll give you all the details and the results in RAW format (DNG).

The shots taken were produced 40 cm away from the sheet of paper, with the cameras resting on the floor, at the same distance.
Both in 4:3, with all automatic (exposure etc.), ie how the average user will use it.
Daylight, around 10 in the morning, with good light conditions, Sun slightly covered by a cloudy haze.
Everything else is displayable in the EXIF data.
If you also want to know what I ate for breakfast before doing these tests i'm sorry, I don't remember.

---DELETED---

MarcoR Posted at 10-15 03:37
Perfect David, then I'm your worker: tell me which tests you want to do and I'll do it.
Tell me what and how to do it and I'll give you all the details and the results in RAW format (DNG).

Now shoot some video outdoors with depth to it.

IF you download the files, pixel size is the same for both (3000x4000) - it seems to me to be a perfectly valid test - now I'm starting to think that there's been a firmware issue affecting everyone, and also a hardware issue affecting a smaller number of purchasers.  People suffering only from the FW issue now seem to be happy, but those with the HW issue too are obviously not going to be helped much by the FW update.

fansfe82067d Posted at 10-15 08:25
IF you download the files, pixel size is the same for both (3000x4000) - it seems to me to be a perfectly valid test - now I'm starting to think that there's been a firmware issue affecting everyone, and also a hardware issue affecting a smaller number of purchasers.  People suffering only from the FW issue now seem to be happy, but those with the HW issue too are obviously not going to be helped much by the FW update.

I think exactly the same thing.
For those who have the lens system in place the old firmware had little sharpness, then adjusted in the new.
But for those with bad lenses, the problem cannot be solved with a firmware update but only at the hardware level.
I compared it in flight with the DJI FPV and I must say that in the remote subjects it did not behave badly, on the contrary, better than the DJI FPV.
Obviously it has no focus problem in distant subjects but under 2/3 meters it does, so it is quite easy to understand that it is a matter of lenses.

David_Harry Posted at 10-15 02:07
Are you sure that what you are showing isn’t the result of minimum focus distance and not an out of focus lens. You wouldn’t need a focus chart to see how bad that focus is, you’d see that immediately without any comparison when recording any video.

I’m not saying there isn’t an issue with the 3, I’m just saying that this particular test may not be correct.

I think thats the problem with this guy, there well maybe a problem with close focus, this test is just too overcooked , in fact his lens is so out of focus a visit to spec savers is in order.

Thanks for sharing.

hallmark007 Posted at 10-15 10:51
I think thats the problem with this guy, there well maybe a problem with close focus, this test is just too overcooked , in fact his lens is so out of focus a visit to spec savers is in order.

I totally agree. I think it's overcooked (Cherry picked).

Also. Given the fact that the Action will have a wider FOV compared to the Hero. To get the same framing the Action will have to be closer to the card. So, just how close?

Also. I'm very curious about the similar geometry between the shots.

In any event, this is not an example of what everyone seems to be talking about, as in, a soft focus. This is an example of a totally out of focus lens, something that doesn't need a focus chart to demonstrate. As you say, a visit to spec savers is in order with this one. Although, my money is on this being a BS example because the shot has been deliberately taken inside the minimum focus distance.

Don't forget. DJI's published distances for minimum focus on their cameras should be taken with a pinch of salt. Anyone familiar with my old videos will know I have shown minimum focus distance issues and soft focus ages ago on the Action 1. Nothing being reported with the Action 3 is new to any of us who've used these cameras from the start. I'm most certainly not making excuses for DJI, they should be hounded for repeating the same mistakes time and time again. I'm just saying that with DJI's track record, these things are of no great surprise. Something that GoPro have been worse with, although the Hero 11 is shaping up to be the action cam that it should always have been.

BTW. MarcoR. If you do a proper video example of these two video cameras outside with proper wide shots and provide some downloadable camera files and your Action is still as out of focus as it is in these stills examples. I will immediately concede that I have made a mistake and will give an apology for accusing you of BS and you'll also have my condolenses for having the worst focus problem that I've seen on any Action.

MarcoR Posted at 10-15 03:13
Do you know how to treat an image in DNG (RAW) format?

https://helpx.adobe.com/it/photo ... erter.html#Appendix

I didn't Thanks.

I have another remark: the dji images look like they have been taken in Standard (dewarp) mode, rather than Wide.  You will get a bit extra detail where the chart is in Wide.

David_Harry Posted at 10-15 12:10
I totally agree. I think it's overcooked (Cherry picked).

Also. Given the fact that the Action will have a wider FOV compared to the Hero. To get the same framing the Action will have to be closer to the card. So, just how close?

No problem Dave, I don't need an apology of any kind, we're here to expose our problems and considerations.
If I can I will produce those videos and provide the sources, first for DJI.
Today we did similar tests with GoPro HERO11 and it's not that different, strangely, the HERO6 performed much better at relatively short distances.

I have the Action 2 and Action 3 here and did some simple tests, but  sufficient to determine, the much better focus on the Action 2.


My Action 3 will go back :-(

Action 3

Action 2

---DELETED---

JustName Posted at 10-15 14:04
I have the Action 2 and Action 3 here and did some simple tests, but  sufficient to determine, the much better focus on the Action 2.

Smart move.  That is exactly what I did a while back.  My Action 2 never had issues with close focus.  It was always sharp enough for photos and video.  The Action 2 clearly showed luminance grain so was plenty sharp enough for me.  Everyone has their own opinion on their definition of “sharpness” and that is perfectly fine.  Blurry is not in focus.  Simple!

The images posted by @MarcoR always have seemed legit to me.  So am somewhat surprised anyone feels he is “cherry picking” anything.  Instead of questioning his credibility - upload your own OOC samples to "prove" your point.  He has uploaded many samples and that should be appreciated.

As noted before - energies would be much better spent focusing and putting pressure on DJI Support vs haranguing forum users like @MarcoR  who have visible focus “issues”.

DJI did this.  Not forum users.  They have been given the right to simply suffer.  While DJI smiles and keeps the cash firmly in their bank account.  Personally - have not been following these "focus threads" much anymore - as have been putting the Insta360 X3 under a battery of real world tests.  So far so good.  Back to completing more "work".

10-Bit Posted at 10-15 17:21
Smart move.  That is exactly what I did a while back.  My Action 2 never had issues with close focus.  It was always sharp enough for photos and video.  The Action 2 clearly showed luminance grain so was plenty sharp enough for me.  Everyone has their own opinion on their definition of “sharpness” and that is perfectly fine.  Blurry is not in focus.  Simple!

The images posted by @MarcoR always have seemed legit to me.  So am somewhat surprised anyone feels he is “cherry picking” anything.  Instead of questioning his credibility - upload your own OOC samples to "prove" your point.  He has uploaded many samples and that should be appreciated.

Thanks for your words.
In all honesty I don't give a damn about anyone who questions the veracity of the test, as long as DJI believes me.
I simply wanted to report my negative experience so that others could check if their cam had also failed.
Then the owners of this cam will do as they want, I do not intend to keep a product that has certain specifications and does not respect them.
Each has its own evaluation parameters, as I have mine, and from the tests I have done I have ascertained without a shadow of a doubt that this OA3 has serious out-of-focus problems with subjects in a range where other action cameras work much better.
Just take two comparative shots, it doesn't take long.
Even a video of mine clearly demonstrates it, foolproof.
I looked for confirmations from other users on the web and found them.
I always evaluate the negative constructive comments, even the adverse ones, as long as they are exposed with a logic that aims to solve the problem, not to denigrate the pseudo inability of the other.
I remind you that DJI is a publicly traded company, and the shares go up and down if even one product is good or bad.
So maybe it is permissible for the holders of these shares to praise DJI and its products regardless of the various web channels, they must safeguard their investment money!
So my filter for some comments always remains active...

I post some RAW shots we took with GoPro HERO 11, one always at those 40 cm from the paper and the other closer, daylight but not excessive, trying to put it in difficulty.
The sharpness as you can see it is obvious that it is horrible up close but at those 40 cm it certainly works better, this confirms that in the case of my OA3 the sharpness is not even in the objects that exceed the hyperfocal distance.
So in the product in my hands the problem is unfortunately evident.
I don't care if anyone here believes me or not, I just hope that DJI confirm what I reported, as I hope it will take into consideration maybe the same problem reported by other users.

https://drive.google.com/file/d/ ... O7/view?usp=sharing

This shot was taken at close range with the Action 2.  Stitch detail is there.  All shots were able to resolve IQ like this.  Low light in shadows.  Which looks reasonably sharp.  No real complaints ever.

Posts etc saying that the focus and IQ are similar to the OA3 - make little sense.  If you have a Action 2 that cannot focus close like this - you simply have a defective Action 2.  If your OA3 cannot focus like this - you have a defective OA3.  Simple.  No reason to debate or argue.  Just tell DJI to fix your OA3 - to at least focus as well as this.  No sensor size change.  No FOV change.  

Most know by now.  The reason I bought my OA3 - was my understanding that it could focus like this.  Plus was able to remove the selfie stick skiing and snowboarding like the Insta.  Unfortunately the OA3 has a problem because very few if any can focus like this.  And if any can - and yours cannot.  Your OA3 is defective.  Hmmm?

I recorded two comparison videos as it was requested.
I share some extracted frames at various distances, always a comparison between Osmo Action 3 and GoPro HERO 6.
20, 30 and 40 cm, same subject filmed, low light conditions, cloudy weather.
4:3 format, D-Cinelike (Flat for GoPro), 30 fps, RockSteady off.
I hope that the matter is now more evident.
On the GoPro I can read the S/N of the hard disk at any distance, with the Osmo Action 3 I struggle at any distance.
In the Osmo Action 3 you can see very well that with the latest firmware they have increased the sharpness but it is of little use with the focus busted.
If you want I also share the original videos of both the action cam, but I think it is useless since the result is the same.

https://drive.google.com/file/d/ ... XQ/view?usp=sharing

+ 4,0 diopter and NO PROBLEM. And how does DJI do it? Front protective lens cover?
https://www.youtube.com/watch?v=Qrfu4jn5CdE

---DELETED---

osmonauta Posted at 10-16 01:30
Here's my focus chart in full sun from 40cm. This is a cropped version and saved at 90%. Shot on a tripod. Used a 2 second delay.

Originals are on GDrive. Both JPG (auto) and RAW (ISO 100-200) provided. Both have 2 versions. The letter at the end of the file names: W = wide; N = normal (aka de-warp)

Same issue of mine.
Production date of your OA3?

---DELETED---

Hi there. Thank you for reaching out and we're sorry for the trouble. Our engineers are aware of the issue, please be advised that it is found that this problem is caused by the large internal parameters of some cameras. In response to this problem, you can provide our technical support with your SN and DJI account, original photos, and videos through Chat, email, and other channels, and we will help you to solve this problem asap.

Here is the link to contact us: https://www.dji.com/support

DJI Paladin Posted at 10-16 03:05
Hi there. Thank you for reaching out and we're sorry for the trouble. Our engineers are aware of the issue, please be advised that it is found that this problem is caused by the large internal parameters of some cameras. In response to this problem, you can provide our technical support with your SN and DJI account, original photos, and videos through Chat, email, and other channels, and we will help you to solve this problem asap.

Here is the link to contact us: https://www.dji.com/support

Proof is needed to confirm a serial or do you have the serial numbers affected to provide. seems people do not do exactly as requested to resolve the issues,  we al lwant to know if our cameras are at risk and how will we know. tocall support and ask?

DJI Paladin Posted at 10-16 03:05
Hi there. Thank you for reaching out and we're sorry for the trouble. Our engineers are aware of the issue, please be advised that it is found that this problem is caused by the large internal parameters of some cameras. In response to this problem, you can provide our technical support with your SN and DJI account, original photos, and videos through Chat, email, and other channels, and we will help you to solve this problem asap.

Here is the link to contact us: https://www.dji.com/support

Thanks for the reply.
I have already sent everything you have requested to technical support, in this thread there is still everything I have already sent, but keep replying to contact the technical support.
For what? It looks a bit like the cat biting its own tail.
We turn around the problem but no solution is seen.
What should I do with this cam? Can you tell me once and for all?
Regards,
Marco

---DELETED---

osmonauta Posted at 10-16 02:57
2022/08

There's not much choice. The cameras are either August or September - at least what I've seen so far. I already have my RMA label printed. It's going back on Monday. They gave me a case number and RMA label on Thursday but I was holding it back to test the new FW. Of course I'm over the 14 day so it's going to be a repair or replacement or who knows. Maybe they will send it back with "no issue found" written on it.

Same of mine, 2022/08.

Did some further tests with Action 2 vs. Action 3. Action 3 is clearly defocused.

What a shame, had been really looking forward to the cam because of better battery life, internal memory and bigger screen. Now the Action 3 will go back to Amazon.

Action 2

Action 2

Action 3

Action 3

Action 2

Action 2

Action 3

Action 3

---DELETED---