I had wondered why the 40mp photos are not better than what my osmo action 4 delivers.
After a few simple tests i know the reason.  What the osmo action 5 delivers as 40mp resolution is only an bad upscaled 10mp resolution. If you take a photo in both resolutions and compare them, you realise that the 10mp resolution produces even better images than 40mp.

Full picture (40mp downscaled)




Osmo Action 5 40MP 100% crop




Osmo Action 5 10MP 200% ! crop




Osmo Action 4 200% ! crop




Osmo Action 5 40mp 200% crop




As you can see, the 40mp 100% crop is already worse than the 200% crop of the 10mp resolution. The 200% crop looks even worse.

Hi there,

We are truly sorry for the inconvenience you are experiencing.

Thank you for your feedback. Please rest assured that we will forward your feedback to the related department for evaluation. In the future, we will optimize and improve the product or service based on user comments.

Thank you for your support and understanding.

Quality of a picture doesn't automatic increase with more MP, the lens is the major factor.

That's why phones produce terrible quality images compared to a cheap 15 year old DSLR with a good lens.

I've a quastion in a similar way: Why does have the RAW (DNG) files such a low resolution: 960*720px ? It means around 0,5Mpx?

tsiwo Posted at 9-21 03:01
I've a quastion in a similar way: Why does have the RAW (DNG) files such a low resolution: 960*720px ? It means around 0,5Mpx?

It is just a preview resolution.

johansenfoto Posted at 9-21 02:00
Quality of a picture doesn't automatic increase with more MP, the lens is the major factor.

That's why phones produce terrible quality images compared to a cheap 15 year old DSLR with a good lens.

Sure it depends also on the lens.  (It seems to be the same as on OA4)

But! DJI explicitly advertises the high resolution, which is currently of no real use.

"Each photo is around four times clearer than the previous generation and has a resolution of 40 megapixels (7294×5472) or 8K (16:9)."

Interesting , if this proves to be true , then my OA5 Pro will be returned to  Amz in no time cause i would consider it as false advertising .

40 MP sensor is used for video to be able to use better stabilization and also the follow me function. But think it's strange that it doesn't have 6K video resolution.

JustName Posted at 9-21 03:03
It is just a preview resolution.

No it's not. If i read the Fileinformation on My Devices (all devices Windows & iOS) every RAW DNG is only the mentioned Size.

tsiwo Posted at 9-23 00:23
No it's not. If i read the Fileinformation on My Devices (all devices Windows & iOS) every RAW DNG is only the mentioned Size.

Yes, you see only the preview resolution in the fileinformation :-)

So it means there is no 'hardcoded' resolution on the RAW Files? Thanks for clarification on this Usually i use MediaInfo (MediaInfo)
Here is the Windows Explorer View:

Each DNG file contains a preview version as a jpg. This preview version has the resolution as you can see in MediaInfo.

tsiwo Posted at 9-23 00:45
So it means there is no 'hardcoded' resolution on the RAW Files? Thanks for clarification on this Usually i use MediaInfo (MediaInfo)
Here is the Windows Explorer View:

RAW files contains only preview of the image and even that is optional. It is sensor read our data it is nothing until you export it with resolution and dpi you set it to be.

When exporting RAW you can set resolution and DPI to be what ever you want. These factors tells nothing about the image quality until a certain point is reached.

Thanks. This was exactly what I was searching for and now I know, I can skip the OA5

Can't wait for my OA 5 to be delivered , so i can test it myself , even if i don't have a OA4 to compare with .

It's odd that DJI's folks respond on other topics when there are some nice videos shot with the OA5, but no comment here about their own advertising :

"Each photo boasts approximately 4x the clarity of the previous generation with roughly 40 megapixels (7294×5472) or 8K (16:9) resolution."

No matter how i read that , it translates with 4 times better  , not 4 times bigger  .

johansenfoto Posted at 9-21 02:00
Quality of a picture doesn't automatic increase with more MP, the lens is the major factor.

That's why phones produce terrible quality images compared to a cheap 15 year old DSLR with a good lens.

Exactly, and more you have pixels, more the lens must be better.
More pixels yes, but smaller pixel (for the same size of the sensor) so worse in low light, and potentially contrast.

Xavier1984 Posted at 9-23 05:31
Exactly, and more you have pixels, more the lens must be better.
More pixels yes, but smaller pixel (for the same size of the sensor) so worse in low light, and potentially contrast.

That's true. But there were so many that complained that OA4 "only" had 10MP, so I think DJI did this to please those

johansenfoto Posted at 9-23 08:26
That's true. But there were so many that complained that OA4 "only" had 10MP, so I think DJI did this to please those

eh eh, nice answer ;-)

What is the real-world resolution of the Action 5's sensor? DJI mentions a pixel size of 2.4 μm. The sensor has a width of about 10 mm. This results in about 4,000 pixels horizontally. So 4000 x 3000 pixels = 12 MPixel. How do you convert 12 MPixels into 40 MPixels? To understand this, you have to deal with the concept of the Quad Bayer sensor:

https://www.sony-semicon.com/en/ ... d-bayer-coding.html

In this sensor, there are 4 photodiodes under each color filter. They can be connected together in poor light conditions or treated separately in good conditions, e.g. for 2 different exposures at the same time, one according to the lowlights, the other according to the highlights. In order to achieve a dynamic range of 13.5 stops (at least that's what DJI says) with the small sensor. OneShotHDR is the name of the process.

And you can use special algorithms to interpolate a higher photo resolution. Insta has been doing this for some time with the Ace pro, and DJI does so with the drones with the 3/4" sensor. Does that make sense? If you follow the pictures shown here probably not. Only huge files are generated and no visibly better images. I don't want to rule out the possibility that DJI has not yet implemented the right algorithms for upscaling and will add them in a later update. In the Action 4, for example, the OneshotHDR mode was only added in an update, at least up to 30 fps.

@JustName thanks so much for the photo comparison, it's really hard to find other comparisons like that. By any chance did you take more shoots, similar results?
@Balistic it would be great if you could give your impression as well once you receive OA5.

I wouldn't expect the Action 5 to take better 40 mp photos than my A7IV does in 35 mp with my Sony 16/35mm.
Do you really return an action camera because of this? The 40 mp sensor is used for pixel interpolation and/or multiple exposure to raise the dynamic range (i guess it's a quad bayer sensor).
40 mp photos is just a collateral effect.

You do realize that a 1/1.3" sensor was never going to take a real 40MP picture?  There are cameras that take genuine 40MP pictures but they cost $000's.  Its just software wizardry in the Action 5

johansenfoto Posted at 9-21 02:00
Quality of a picture doesn't automatic increase with more MP, the lens is the major factor.

That's why phones produce terrible quality images compared to a cheap 15 year old DSLR with a good lens.

I think you need to say this louder...

Not sure of why the majority of people seriously expect quality out of those little lenses...

If you want serious quality, you need to get professional lenses and decent body size to go with it...

It is not a secret... It is common knowledge, and it has been since the first lens was produced..

You can put lipstick on a pig, but it is still a pig!

Burt37 Posted at 9-27 14:30
I think you need to say this louder...

Not sure of why the majority of people seriously expect quality out of those little lenses...

Many people do think that more MP will actually produce a sharper image, just because there will be more pixels.
With video the lens do matter but not as much as with still images.

Steph Jant Posted at 9-27 10:42
You do realize that a 1/1.3" sensor was never going to take a real 40MP picture?  There are cameras that take genuine 40MP pictures but they cost $000's.  Its just software wizardry in the Action 5

If DJI sells me a camera and says that it now delivers an image that is twice as sharp thanks to the higher resolution, then I also expect the image to be better.

And of course I realise that a lot depends on the sensor size, the lens, the lighting situation, etc.

But... I had tested further with my camera.
On the one hand, it seems to have been defective anyway. The camera always had a blurred image on the right-hand side.
Secondly, the OA5 doesn't seem to be perfectly focussed from the factory either, as was extremely the case with the OA3 and slightly with the OA4. And this was the main reason with my camera, that you could hardly see any difference between 10mp and 40mp.

After adjusting the lens correctly, you can clearly see the difference between 10mp and 40mp, especially when shooting at short distances.

So I have to correct my original statement. The 40MP image is not an upscaled 10MP image.

But if your cam is not good focused, you will not see any improvement to 10mp.

Exactly my thoughts , it's not about being on par with a 5k EUR camera, it's about DJI's statement/advertising compared to their previous model , which costs 100 EUR less and it seems that to some extent it's better than the newer model, so i guess many people wonder why spending more on the OA5 .

JustName, so you tweaked the focus on the new OA5 too ?

JustName Posted at 9-28 00:55
If DJI sells me a camera and says that it now delivers an image that is twice as sharp thanks to the higher resolution, then I also expect the image to be better.

And of course I realise that a lot depends on the sensor size, the lens, the lighting situation, etc.

You do realise that you are dealing with a company that wants to sell quickly large quantity and most of you rushed to buy the first batch...

Advertisement and reality, very rarely go hand in hand...

Don't get me wrong... I believe you are right in expecting what it is advertised, but I mean, is this your first DJI purchase?

I see a recurring pattern here, and yet people are always surprised when a NEW product has bugs and/or missing features...

DJI will improve and fix what they can while they can until the new product get released, and the wheel goes around and around...

Balistic Posted at 9-21 04:55
Interesting , if this proves to be true , then my OA5 Pro will be returned to  Amz in no time cause i would consider it as false advertising .

You seriously didn’t think you were getting a full 40mp sensor ? You will find it hard to get this type of sensor for less than 2k. What you have is simply a quad bayer sensor which allows division of mega pixels IE 1 MP split in 4 even sized pixels and yes you can see a change in quality but really only in good light but I have found you can manipulate pixels a bit more to get cleaner images and larger prints are possible. But its regulations that allow for something to be advertised with for instance 48mm an effective 12mp sensor evenly divided in 4. With the pocket 3 you have 9mp photos and comparing them to a quad bayer 48mp there is little difference as bigger pixels will also give a clearer photo. Below is basically stretching the 40mp of the OA5 .

Burt37 Posted at 9-28 02:14
You do realise that you are dealing with a company that wants to sell quickly large quantity and most of you rushed to buy the first batch...

Advertisement and reality, very rarely go hand in hand...

This is not something specific to dji, many camera companies sell their cameras these days with beata software in fact customers are demanding this because it simply means two things software was not quite read or companies are future proofing their products. I remember the days you bought an SLR it was a finished product nothing could be changed and yet many things went wrong and many were returned for fixing forums particularly camera forums lit up with hundreds of complaints. So all that has happened is problems can now be sorted by way of downloading FW/SW and many unfinished extras can  be passed on by simple updating.
I don’t think the way tech is now we will ever buy tech that is finished without any problems and won’t be updated and I also think if that was the case many would by crying because no updates are coming.

Balistic Posted at 9-28 01:48
Exactly my thoughts , it's not about being on par with a 5k EUR camera, it's about DJI's statement/advertising compared to their previous model , which costs 100 EUR less and it seems that to some extent it's better than the newer model, so i guess many people wonder why spending more on the OA5 .

JustName, so you tweaked the focus on the new OA5 too ?

Yes, i adjusted the lens.

I already have a good camera for very good photos, which cost considerably more than 5k with the right lenses etc. :-)
Nikon Z6ii with some good lenses from Nikon.

Of course you can compare the Osmo Action 5 with the Nikon.
But I don't want to mount the Nikon on my motorbike or go diving with it.
Nevertheless, I will be pleased if the new action cam is a little better than the previous one.

Burt37 Posted at 9-28 02:14
You do realise that you are dealing with a company that wants to sell quickly large quantity and most of you rushed to buy the first batch...

Advertisement and reality, very rarely go hand in hand...

Have a look on the left under my profile name. You will see that I have quite a few DJI products :-)

I fully realise that what is advertised and the reality never match perfectly.  Still, you more or less expect that it won't be completely made up.

That being said. DJI products are never perfect from my point of view, but they are currently the best on the market.
And of course there will be some bugs, especially right after the release. I can deal with them and if I find any, I'll report them and share them with the community.

JustName Posted at 9-28 03:25
Yes, i adjusted the lens.

I already have a good camera for very good photos, which cost considerably more than 5k with the right lenses etc. :-)

" I will be pleased if the new action cam is a little better than the previous one."

As it should be...

JustName Posted at 9-28 03:34
Have a look on the left under my profile name. You will see that I have quite a few DJI products :-)

I fully realise that what is advertised and the reality never match perfectly.  Still, you more or less expect that it won't be completely made up.

It was a rhetorical question..

As you have been purchasing DJI product before, you are not new and you should be familiar with the way DJI does run their business...

Something that you have now confirmed...

"And of course there will be some bugs, especially right after the release. I can deal with them and if I find any, I'll report them and share them with the community."

Hallmark007 Posted at 9-28 02:58
You seriously didn’t think you were getting a full 40mp sensor ? You will find it hard to get this type of sensor for less than 2k. What you have is simply a quad bayer sensor which allows division of mega pixels IE 1 MP split in 4 even sized pixels and yes you can see a change in quality but really only in good light but I have found you can manipulate pixels a bit more to get cleaner images and larger prints are possible. But its regulations that allow for something to be advertised with for instance 48mm an effective 12mp sensor evenly divided in 4. With the pocket 3 you have 9mp photos and comparing them to a quad bayer 48mp there is little difference as bigger pixels will also give a clearer photo. Below is basically stretching the 40mp of the OA5 .

[view_image]

Of course i wasn't expecting that .
But like i said, this is my first DJI action cam (which i bought to shot videos but also stills during family trips) and i don't have the previous model to compare with , so when DJI says that still shots are much better than the previous model , then my decision is also made on that  ( otherwise why paying more for a subpar model ) .
And i think this is a valid point for both new and old customers of DJI's products ( btw, have you seen any comments from DJI on this thread , cause i did not and for me not saying a word ,sometimes has more meaning than saying 1000 words )

Burt37 Posted at 9-28 03:49
It was a rhetorical question..

As you have been purchasing DJI product before, you are not new and you should be familiar with the way DJI does run their business...

Sure, I know how it works.
But as I said, DJI is (from my point of view) still much better than the competition and the Osmo Action Cams in particular cost me no more than pocket money, so I can play around with every new version.

JustName Posted at 9-28 03:56
Sure, I know how it works.
But as I said, DJI is (from my point of view) still much better than the competition and the Osmo Action Cams in particular cost me no more than pocket money, so I can play around with every new version.

Do you have any experience with the GoPro?

Burt37 Posted at 9-28 04:02
Do you have any experience with the GoPro?

Sure, i had a couple of them.

Especially since version 10, there has been hardly any recognisable progress at GoPro.
And GoPro is pretty well known for crashes and various problems with the software.

I'm also using Insta360 X4 cam. Is currently my favourite cam.

Yesterday i did several tests outside , during a sunny day , with both videos and stills.Overall i wasn't pleased with the results , so today i will do more tests , to make sure i wasn't doing anything wrong .

However , i decided to do some indoor tests ( but with plenty of light ) , comparing the OA5 against a 2017 Huawei Mate 10 Pro smartphone . This is the result :


https://ibb.co/54ZzXvq


https://ibb.co/2gSjxch

Shooting distance was 50 CM away from the object , so it's more than the minimum focus distance of the OA5 .
OA5 was set to L 4:3  (40mb) , Standard Dewarp , AUTO .Smartphone was in Auto Photo mode .

Btw, i did this test mostly because during the outdoor tests with the OA5 i noticed that when shooting both videos and stills ( selfies) the people face was lacking details .

Let me know what you think after cropping or zooming on both pictures ( i named the pictures accordingly , so you will know which is from OA5 and Mate 10 pro ).

Balistic Posted at 9-29 00:35
Yesterday i did several tests outside , during a sunny day , with both videos and stills.Overall i wasn't pleased with the results , so today i will do more tests , to make sure i wasn't doing anything wrong .

However , i decided to do some indoor tests ( but with plenty of light ) , comparing the OA5 against a 2017 Huawei Mate 10 Pro smartphone . This is the result :

The results from the OA5 don't look too bad for an action cam and the wide angle.

But actually try to do the same from a distance of 35cm. The focus should be sharp there.

I simply use 2-3 water bottles with text and barcodes on them for such tests. The barcode in particular is a good way of recognising when the focus is off.

Here are two photos at 200% crop, for example. First photo before the focus was adjusted, i.e. as delivered, and second photo after adjustment.

Balistic Posted at 9-29 00:35
Yesterday i did several tests outside , during a sunny day , with both videos and stills.Overall i wasn't pleased with the results , so today i will do more tests , to make sure i wasn't doing anything wrong .

However , i decided to do some indoor tests ( but with plenty of light ) , comparing the OA5 against a 2017 Huawei Mate 10 Pro smartphone . This is the result :

The first one has slightly more detail this can be seen when zoomed in on the wheels and handle of the van . They are very close but lenses are different sizes in both so one photo is more zoomed in than the other this will mean things like auto shutter speed iso and aperture will most likely vary.

They are both auto so you are also comparing the processor’s in each camera and how they handle the jpegs , i also presume the Huawei doesn’t shot raw so if you shoot raw on the osmo you are likely to record more detail on the raw file.

If you just shoot jpeg they are pretty close and I’m sure one could make a case for either.