Still not sure what to think about it. I was expecting a little bit more in terms of sharpness. Weirdly it has kind of distance range where it is so so sharp but too close and it isn't (minimum focus distance obviously) but then it's like anything from 3m on is not sharp either. Judge for yourself, shot in 8k30.

as a sidenote this YouTube Short format you posted in is unreadable on my laptop (Firefox) - I get this massive long display which takes ages to load and can't display properly even zooming out.

This is what it looks like zoomed out fully:

Wonder if other non-phonies are in the same boat?

Can't therefore comment on the vid/camera results, though I would have liked to see what it looked like!

cheers

Edited June 29Jun 29 by bghazzal

22 minutes ago, bghazzal said:

as a sidenote this YouTube Short format you posted in is unreadable on my laptop (Firefox) - I get this massive long display which takes ages to load and can't display properly even zooming out.

This is what it looks like zoomed out fully:

Wonder if other non-phonies are in the same boat?

Can't therefore comment on the vid/camera results, though I would have liked to see what it looked like!

cheers

I have the same issue, I believe it's due to the Shorts format.

1 hour ago, RomiK said:

but then it's like anything from 3m on is not sharp either.

I think once you get past 3m, you're limited by resolution. I thought about getting a 360 camera for B-roll, but decided against it because the quality of footage you get is nowhere near an action camera.

4 hours ago, bghazzal said:

as a sidenote this YouTube Short format you posted in is unreadable on my laptop (Firefox) - I get this massive long display which takes ages to load and can't display properly even zooming out.

This is what it looks like zoomed out fully:

Wonder if other non-phonies are in the same boat?

Can't therefore comment on the vid/camera results, though I would have liked to see what it looked like!

cheers

Maybe switching over to YouTube would help? These built in players have their own issues, YT should play fine. Just click.

Hi @RomiK

Many thanks for sharing this. I've just seen it on my phone just fine. It's always great to see other's experiences using these kinds of cameras. Can I ask were you using the PureVision Mode and how deep were you?

These cameras are built more for Content Creators, so it's a whole different system to work with.

I've seen some incredible images and footage with them, but they definitely need the right angle, the correct position and the right light. I've found the same as you in terms of focusing distance as well.

Still loved seeing this though, thanks for posting.

Maria

No I wasn't using Pure Video... honestly I forgot about that mode the entire trip, I got it like two days before. I am leaving for Indonesia next week so I will try it there. It didn't look though as there would be dark, this video was like at 18m at Yolanda reef in Red Sea during the clear day. There were other dives ideal for that I guess like Thistlegorm but again ... We will see next week ...

13 hours ago, RomiK said:

Maybe switching over to YouTube would help? These built in players have their own issues, YT should play fine. Just click.

It's a little better in youtube, at least the vertical format fits...

Yes, not super sharp indeed...

otherwise on YT I get a garbled image up to 0:19 seconds or so - thought it was my connection but it seems to be in the clip compression / player / cache, as it's systematically on the same point.

this is what is looks like:

and it clears up here 0:19 I think

strange beast, these youtube shorts...

cheers

Edited June 30Jun 30 by bghazzal

Hi again @RomiK

It's totally understandable when something is new. I still enjoyed the clip. I'll have a think tomorrow to see if I can come up with something else before you head off to Indonesia.

Thanks again for posting and sharing your thoughts.

Maria

On 6/29/2025 at 11:46 AM, RomiK said:

No I wasn't using Pure Video.

I have had very similar results when using Pure Video. Really not sharp, especially close to the lens. It looks best when it is the invisible stick away or a little farther from the subject.

How did you find the housing? I get a LOT of terrible reflections bouncing around, there's a new one coming that I'll definitely try.

Completely agree about the sharpness at a distance, or lack thereof.

Also, you actually have to be mindful of turtles with it, they can be extremely attracted to it. I've had a few try to bite it.

Edited September 1Sep 1 by lylefk

I've found housing ok. No problem with fogging, good response. Stiching glitches on occasion, definitely avoiding stitch over the subject. So good overall. Just the camera isn't there yet for UW. I also got Osmo360 so waiting for housing for this one...

I'll just add that the main reason i use 360 underwater is for VR using a Meta Quest to let non-divers experience the world down there. I have a bespoke GoPro Max housing that works well for this, but the Insta360 in the underwater housing is not acceptable. Poor stitching in significant number of shots and really soft to the point of unusability. I almost wonder if they are capturing data from a subset of the sensor and then extrapolating leading to the softness. I hope they can firmware up this, but at the moment, I do not consider this an underwater tool (it works ok If you then turn it into a non -360 image... but it is still really soft.)

I hope the Osmo360 in a housing may be better (or really hope for a firmware update that makes the X5 work underwater).

Or if anyone has settings or tricks to make it not look bad - I'm listening!

they release a dive case PRO... its a HUGH dome compare to non PRO

6 hours ago, hellhole said:

they release a dive case PRO... its a HUGH dome compare to non PRO

I just ordered one and will let everyone know how it goes. The dome I have for the GoPro Max is giant and scratches at a sideways glance -so if this works it is still a better option. I am still annoyed that Insta360 would even produce the non-Pro version - or maybe that is for above water only... in reality.

As an FYI - 14 day lead time for shipment of the new housing for the insta360 X5.

To hopefully help people see what normal quality looks like from the X5 and non-pro video housing here are some screenshots. They look fine if viewed on a phone but as soon as one starts to look in closer, they are just very soft and really obvious stitching depending on the sun angle... and maybe if there is something close for it to focus on? Not sure as there didn't seem to be a rhyme or reason. These were shot on pure video. It could also be that my particular camera was not having a good day as it did fry itself half way through the dive (Insta360 did replace it without hesitation, although that was after I returned from the trip).

This is the best of (and we do large scale experiments on the seafloor to understand the impact of nutrient polution and fishing on coral reef ecosystems - those are the cages you are seeing in some of the images):

I uploaded these at full size so you can see how it looks. That is the best one.

Here you can see what it looks like when close to things.

A non-smooth stitch that would sometimes happen.

And how it sometimes looks with the 'invisible' stick.

@RomiK - I hope you don't mind me adding these as further examples to your threads. I do like the story telling in your video.

You can see in all of the 360 promo videos that they play with the 360 aspect by zooming around (which is really fun) but also masks the soft nature of the camera in the housing itself.

Hope this sets expectations correctly for anyone who may be purchasing this camera and housing.

Images are really soft (euphemism). Completely useless except maybe documentation.

Maybe the video is a little bit better?

1 hour ago, Davide DB said:

Images are really soft (euphemism). Completely useless except maybe documentation.

Maybe the video is a little bit better?

The video is not as obviously bad if viewed on a small screen or a corner of a browsing window. It still does not focus close at all and there does seem to be a certain distance that it is able to get a sharper image. When the stitching is off it is just as if not more obvious and in that case it is sort of like two 180 cameras some of the time, and 1 360 cam another time and you don't know until you download the footage at the end of the dive.

Video will be as better as smaller screen to watch will be and as larger portion of 8k footage you can cram on it. I got 360 a bit naive to think it could at least document the dive - and it can if you can bear watching the footage out of it - but now I think its good for like those special effects if you building memories video and nothing much cinematic :-) . Like this instagram bait :-)

On 9/14/2025 at 10:39 AM, RomiK said:

Video will be as better as smaller screen to watch will be and as larger portion of 8k footage you can cram on it. I got 360 a bit naive to think it could at least document the dive - and it can if you can bear watching the footage out of it - but now I think its good for like those special effects if you building memories video and nothing much cinematic :-) . Like this instagram bait :-)

Basically unusable for any work except some social network like...

I saw the new dive housing:

https://store.insta360.com/product/x5-invisible-dive-case-pro?_sc=NzMzMDM3MCMxMTc1MzIy

They write:

Insta360 X5 automatically recognizes the Invisible Dive Case Pro and enters Dive Case Mode, meaning you can start creating in moments.

It's like they have a different stitching algorithm while underwater. And probably it have some impact on image rendering somehow.

Total newbie here.

The actual resolution of a 2D video exported from a 360-degree camera is significantly lower than the camera's total resolution. It depends mainly on the Field of View (FOV) selected during the "reframing" process. In simple terms, the final 2D video is a "crop" of the 360-degree video sphere.

This is because the total resolution of a 360 camera (for example, 5.7K or 8K) is distributed over an entire 360-degree sphere. When you export a 2D video, you are only selecting a small portion of this sphere, so the resolution will be lower.

An X5 has an 8K total 360 resolution, which I don't believe reaches a true 4K 2d frame even with a wide FOV. So, since it's possible to export 4K clips, the software does an internal upscaling. This process seems to be strongly impacted by the distortion that occurs in the underwater housing.

I'm just thinking out loud here, trying to explain myself this drop in quality in underwater footage, which is absolutely not present in land-based footage. There are excellent examples of its use alongside traditional cinema cameras in many situations with no loss of quality.

Looking forward to see some clip shot with this new housing.

2 hours ago, Davide DB said:

Basically unusable for any work except some social network like...

I saw the new dive housing:

https://store.insta360.com/product/x5-invisible-dive-case-pro?_sc=NzMzMDM3MCMxMTc1MzIy

They write:

Insta360 X5 automatically recognizes the Invisible Dive Case Pro and enters Dive Case Mode, meaning you can start creating in moments.

It's like they have a different stitching algorithm while underwater. And probably it have some impact on image rendering somehow.

Total newbie here.

The actual resolution of a 2D video exported from a 360-degree camera is significantly lower than the camera's total resolution. It depends mainly on the Field of View (FOV) selected during the "reframing" process. In simple terms, the final 2D video is a "crop" of the 360-degree video sphere.

This is because the total resolution of a 360 camera (for example, 5.7K or 8K) is distributed over an entire 360-degree sphere. When you export a 2D video, you are only selecting a small portion of this sphere, so the resolution will be lower.

An X5 has an 8K total 360 resolution, which I don't believe reaches a true 4K 2d frame even with a wide FOV. So, since it's possible to export 4K clips, the software does an internal upscaling. This process seems to be strongly impacted by the distortion that occurs in the underwater housing.

I'm just thinking out loud here, trying to explain myself this drop in quality in underwater footage, which is absolutely not present in land-based footage. There are excellent examples of its use alongside traditional cinema cameras in many situations with no loss of quality.

Looking forward to see some clip shot with this new housing.

I think those housings probably have a magnet in the housing that triggers the dive mode like in some other housing. 360° video and resolution has always been a tricky and confusing story in regards to sharpness and video definition. Most people don't do the math of the number of pixels used for each degree of coverage and this value is actually key to understand sharpness.

As you did, if you compare the number of pixels by degree of coverage for a 360 camera and a regular action camera, the sharpness of the 360° is always bad at the same 2D resolution. For reframing/tracking a moving subject, those camera are good with usable results in HD. For 4K work, it's starts to be suboptimal and of course, it won't be usuable for any macro work.

Yes, most of people believe that a 8K 360 camera is a real 8K after 3D reframing 🤡

1 hour ago, eocean-eu said:

As you did, if you compare the number of pixels by degree of coverage for a 360 camera and a regular action camera, the sharpness of the 360° is always bad at the same 2D resolution. For reframing/tracking a moving subject, those camera are good with usable results in HD. For 4K work, it's starts to be suboptimal and of course, it won't be usuable for any macro work.

Yes and no. They are using clips from the 360 to get 4K frames. Original reframe resolution is 2.7K at best, and as you wrote, it decreases narrowing the FOV.
In the meantime I read Insta360 and DJI forums and I confirm that their software upscale the video internally.

As you see in this (and other) review , in good light conditions the results are simply amazing.

I dare you to notice that the UHD clips in this video are upscaled. The math applied to silicon has made a lot of progress lately ;)

So, if these cameras are able to get these amazing results on land and such terrible ones underwater, I wonder if it's a problem with the geometry of the housing or even a problem with the upscaling algorithm itself (screwed clip metadata?).

When I played with Topaz AI, I remember there are different algorithms optimized for the type of image you're processing.

"Upscaling" is some sort of fancy extrapolation: it creates additional information from existing information but this is radically different than a sensor that would actually capture the missing information. Some people might not notice that because their attention is focused on something/one in the center of the frame but there might be something unusual if one looks at the fine details in the corners or when action is fast.

One thing also is that quality of Youtube streams that makes comparison a bit hard but since most people publish their things online on FB, Instagram or YT, sometimes even without the most appropriate encoding parameters, that makes many video look "acceptable". Rendered files before upload would probably useful to watch. I've seen a review between the Insta X5 and Ace Pro 2. 4K X5 is close to the Ace Pro and it's sometimes hard to tell the difference. When the math are done with the same coverage, the Ace Pro is still winning but the gap is small.

About geometry, it's possible that something is wrong or there is a step missing in insta360 studio for the new model. The first generations of Gopro before the 3 had dome housings and the resulting quality underwater was extremely bad untill someone started to make housing with flat port. The OSMO360 is terrible underwater because there is no housing and the 360 lens is simply made for air and not water therefore all the optic is messed up. But there is also one thing. It's inherently different watching a video and watching a photo, those cameras are made for action and the brain will focus where the image is sharp and may not notice the corners. There is a comparison between the pro and non-pro housing underwater: https://www.youtube.com/watch?v=mS0EI9In7yw. I was a bit disappointed by the near lack of differences on the sharpness in the corner and front. The big benefits is the loss of flares.

1 hour ago, Davide DB said:

So, if these cameras are able to get these amazing results on land and such terrible ones underwater, I wonder if it's a problem with the geometry of the housing or even a problem with the upscaling algorithm itself (screwed clip metadata?).

Interesting question indeed. 2nd gen AKA pro version of I360 housing has larger dome right? So the problem might be in the distance of virtual image perhaps? Perhaps it is unsolvable until someone creates diopter corrected lens either for camera or housing which could play with camera and virtual image better... questions no answers...