Has anyone already had experience with the insta 360 X5 and would like to share it with us? it makes a good impression on me, the new, larger sensor should also deliver better image results than all other previous cameras in the class.
I’d also be interested to hear and see raw videos from the camera if anyone can share. Typically the insta360 x3/x4 is not particularly easy to work with because it requires a some manual/time consuming steps to use with Mapillary (compared to say GoPro Max), but the image quality and ubiquity of the camera is attractive.
I recently did a upload with my Insta360 X5. I did use the hyper-lapse mode and needed to manually export the GPX track with third party tools (exiftool). For the example track see: Mapillary
If some developer want to fix the Mapillary desktop uploader i can provide the original video file. The automatic extraction of the GPX track like described by the link of @boris did not work for me.
Still external GPS needed? That’s the biggest holdback to switch to insta360 for me.
Yes. I used my Smartphone connected to the app to get the GPS track into the recording. But I think about record the GPS track separate on my Phone and sync them afterwards. The recording interval of GPS in hyper-lapse mode was quite coarse. I hope they fix that in a firmware/app update. But using a independed recording app on the phone should be a quick fix.
Currently i am planing to set up a little pipeline tool to automate the required rendering, GPS extracting, fixing time stamps, etc steps. I already got access to the SDK of insta360 so that i can do the rendering of the insv
files without Insta360 Studio.
Nice, that’s exciting about the pipeline tool to automate the insta360 work, keep us updated on how it’s coming!
In terms of your capture, what camera settings were you using exactly? It’s a bit hard to understand the transition between one image and the next in your capture. The recommended setting for capturing on foot is “timelapse 360 video at 2s interval” (from the documentation)
When you say the automatic extraction of the GPX track did not work, are you sure you saved the exported 360 .mp4 file in the same file path as the original 00 “.insv” files with the same name as the .insv file? If not, it won’t work. If you’re certain that you did, then yes, please provide a link to the .insv files and your converted .mp4 and we’ll take a look. Cheers!
I will update you (the community) if i have something ready to test!
Regards the automatic extraction:
The mp4 and the 00 insv file are in the same folder. I open the MP4 file with the desktop uploader and get the message that there is no GPS data found in the given sources. If I extract the GPS with exiftool and inject it into the MP4 file the desktop uploader works fine. Boris I will DM you the insv file and the resulting mp4 file. You can share it internally but i would not like to have footage of my face unblurred unblurred on the internet. If you need something for a public bug tracker DM me and I will record something for you! The insv file is that one I used to created the track that i have posted yesterday. The MP4 you will get is a fresh render directly from Insta360 studio without me patching the file.
Excellent, thanks so much. Don’t worry, we don’t share these originals publicly. Appreciate the DM!
Hi @senden9 - I received your DM, and looked at the videos, thank you!
It looks like Insta360 changed some things about the GPS format, and exiftool just fixed it in version 13.28 (Released 3 weeks ago on April 25). We will update to the latest exiftool version with every desktop uploader release, and then it should start working again. For now, the workaround is to use exiftool or the insta360 studio to extract the .gpx file and attach it to the video in the desktop uploader via the overflow menu.
Thanks for spotting this! cc: @nikola
The funny thing is that even Insta360 Studio is wrong with this specific file. The track is way to short. I opened a bug report at Insta360 and they working on that. Only exiftool generates the right track for me. But many thanks for the quick response. Looking forward to the next release of the desktop uploader then!
Based on the first public X5 recordings from a Mapillary user, https://www.mapillary.com/app/?lat=46.619931441906&lng=14.25832692749&z=17&panos=true&pKey=580162391769159&focus=photo&x=0.2923784009266049&y=0.5297478148082716&zoom=0 you can see that the X5 has sphere correction. I am therefore very interested in switching to this camera.
I’ve purchased an X5 recently and have been uploading some stuff which you can check out.
I’ll outline my experience with it so far.
Never done this type of thing before so I don’t know if it’s good or bad.
I will note that there is quite a bit to the workflow but I knew that was always going to be the case and especially since I want to utilise RTK/PPK in the near future for more accurate GPX traces, so that’s why I decided on this camera as I will need to do a bit of post processing anyways.
Imagery Recording
I use the video function to do my imagery as I found the timelapse mode captures images in a poor quality I’ve found and that’s not what I wanted to achieve with this piece of equipment.
Timelapse Mode:
Very low quality, not what I was quite expecting so I didn’t really spend much time on this mode at all.
Above is a snip from the timelapse mode, I can only attach one image as a new user but it is of a much worse quality when compared to a snip from the 5.7k+ video mode.
Interval Photo Mode:
It has an option for every 3 seconds but in reality, it’s more like 5-10 seconds so that throws that option out the window
Video Mode:
I’m trying to go for highest quality so the main options for me here is either the 8K or the 5.7K+ mode.
8K:
Records at 7680x3840 and bitrate of around 188 Mb/s
5.7K+:
Records at 5760x2880 and bitrate of around 126 Mb/s
From searching online it seems like it’s recording in 8k and then brought down to 5.7k resolution, don’t know how true that is.
The lowest framerate for both is 24fps.
I have a hard time trying to find a noticeable difference between the two modes visually, so I opt for 5.7k+ as it produces a smaller file size, thus less time rendering for me when I have to run the video through Insta Studio to give me an mp4.
Plus the images get compressed when viewing them in the browser I don’t feel that it’s worth the extra time required for such a minimal improvement.
Examples below of what I’ve uploaded to Mapillary already:
5.7K+ Example
8K Example
Workflow
I’m looking to automate this a bit more with the MediaSDK that Insta360 has, since I’m running on Linux and their programs seem to not utilise AMD GPU’s at all, it’s either NVIDIA or CPU rendering. So everything runs just that bit slower for me.
Current workflow looks like:
- Insta360 Studio to render to MP4
- Use video_process with mapillary tools to extract frames and attach separate GPX file to the frames
- Fix metadata on images to include 360 exif tags
- Upload
I process the videos locally as my upload speed isn’t the greatest and saves space on storage because I plan on keeping just the images and deleting the videos generated.
I haven’t timed how long it takes to render because I just leave the computer running while I do other things, maybe check up on it every once in a while. Maybe I’ll time this somewhat in the future. But given 8k videos have a much bigger file size it take much longer and as mentioned previously, it’s barely noticeable (at least to me)
GPX Traces
I did notice that exporting GPX traces from Insta Studio has slightly different result to what is actually recorded but I haven’t gotten around to comparing them. At the moment I capture GPX traces from a separate logging app on my android, have the insta app running so it can attach a trace straight into the video and a Garmin Forerunner 955 watch (Multiband GNSS). I’ll post the difference in traces when I get around to it.
That’s my limited experience with it so far. I just saw that the GoPro Max 2 just released and now I’m hoping that it doesn’t blow this camera out of the park quality wise.
Thank you, it very useful and I have the same experience with this camera.
Photo interval is to long to use it on vehicle.
Embeded GPX track via android Insta application is not usable, it is smoothed and incorrect. I have created Insta support ticket but it useless. I am wondering how other people use this, it is fully wrong!
@Oxeelix how do you merge external gpx with video? Do you start it at the same moment or do you find a way to choose where is first video frame on the track? It is most time consuming operation for me.
GPX Comparison
I have done a quick comparison of GPX traces on one of my routes within a forest, unfortunately I didn’t have a longer track with more bends in it with GPX traces from all my devices.
As you can slightly see, the Purple Insta360 Studio Export line has a smoother trace around the corners for some reason. The blue line is what was actually recorded on my phone which I used a separate GPX logging app while have the Insta360 App running at the same time.
The Yellow line is from the Garmin has multi-band which uses both L1 and L5 satellites. Would consider this my reference as it seems pretty accurate.
I did see methods to extract the GPX from the .insv file, Github - Exiftool Insta360 GPX, but I was unable to get any GPX tracks from it with this method. Not sure exactly how Studio gets the GPX information for it to be exported but when I run exiftool -m -ee -api largefilesupport
on the .insv file, I don’t see anything mentioning about location.
Yeah doesn’t make much sense as to why they smooth it out a bit, who knows.
./mapillary_tools-0.14.4-linux-x86_64 video_process VIDEO.mp4 --geotag_source "gpx" --geotag_source_path GPXTRACE.gpx --video_sample_distance -1 --video_sample_interval 2 --video_start_time 2025_09_24_10_12_51_050 --overwrite_all_EXIF_tags --interpolation_offset_time 14
This is the command I use to get the frames and attach the GPX to those images.
The X5 has a timecode feature which adds an additional track to the video which you can see with a tool such as MediaInfo which makes this process easier in my opinion, there’s probably a way to automate this but haven’t looked into it yet.
MediaInfo Output
Other
ID : 3
Type : Time code
Format : QuickTime TC
Duration : 26 min 52 s
Frame rate : 24.000 FPS
Time code of first frame : 11:12:51:05
Time code of last frame : 11:39:43:10
Time code, stripped : Yes
Language : English
Default : No
Encoded date : 2025-09-24 10:12:51 UTC
Tagged date : 2025-09-24 10:12:51 UTC
The process I take is:
- Start recording GPX
- Start recording video
I get the video start time of first frame and then subtract the start time of the GPX and then just input that difference beside interpolation_offset_time
might need to adjust depending how out of sync with time.
The timecode records the local time meanwhile the GPX traces record in UTC+0 so make sure
you adjust the hours and input the video_start_time
in UTC+0 as well.
It helps to try and sync the times as much as you can and try to start recording the video shortly after starting the GPX so less calculations are involved. You could fix this issue altogether by exactly pressing recording on both things at the same time but well, I could never get it right first try.
From there, everything should line up and attach the correct location to the images.
Simple instructions for GPX synchronization:
Open the Insta 360 GPX track and your external GPX recording together in the JOSM editor.
Now trim the external GPX track to the same length as the Insta GPX track. Export the shortened GPX recording and upload it along with the MP4 or MOV file.
Tip: I let the external GPX recording run all day and only extract the required portion.
This example is an early version that I no longer use myself. I’ve since edited the Insta 360 GPX track so that I get good results with this recording, and the recording also works inside buildings and even when driving through tunnels. However, if you’d like to integrate external GPX recordings, and there may be good reasons for doing so, this method is a good fit.
I achieve good rendering times by using an Apple M1 computer. The computer gets a bit warm, so I place my MacBook on top of an old PC fan. These days, the time it takes to prepare my recordings is similar to how long it takes my 125 Mbps internet connection to upload .mov files. This is very practical, and especially with Insta Studio, I can edit my recordings to the desired length. This significantly reduces the stress involved in preparing and finishing recordings.
To obtain a reasonable GPX track, the Insta Remote is required. I have it permanently attached to the bottom end of a second selfie stick. This selfie stick rests on the dashboard and is in a loop near the left mirror window; on the right, it rests on my smartphone mount. I can easily adjust the length using the telescope.
The Insta Remote is powered by a USB-C cable and, in addition to the left mirror window, has a stable Bluetooth connection to the camera on the roof. This position also provides good satellite reception—even in the city. In summer, the remote is cooled by the left ventilation grille of the car’s air conditioning. This keeps the remote display and control button in a stable position while driving, and I can easily read the display.
When I get out of the car, I remove the selfie stick, fully extend the telescope, and screw the Insta camera onto the stick.
In just a few simple steps, the Insta Remote can be converted from vehicle use to selfie stick use. The remote remains in the same position on the pole and is simply rotated.
Something is very annoying, and I haven’t found a solution yet. Before I start recording, I have to restart the Insta remote, and thus the camera. If I don’t do this, I get a GPX track with strangely sprawling curves. However, after restarting immediately before recording, my settings reliably produce a very good GPX track.