Contributing to Mapillary with GoPro Hero 11

Does it work with the Mapillary Desktop Uploader ?

I found a manual via Reddit
https://gopro.my.salesforce.com/sfc/p/#o0000000HJuF/a/3b000000NSpR/sWOFUJHFwK2PyJA0yZsLKU8HmjjoS9Gx2FYAkMNxlDg
The timelapse max speed is still 0.5 seconds.

I bought one and a new SD card.
I have no more errors before the actual upload.
I have put the sharpness (= the level of detail captured) to high.

I am happy with the picture quality.
I wonder if it will help with the partially clouded sky.

Here we have a procession of Echternach.

This back and forward seems a big problem. Four years ago I had that with the Garmin Virb XE.

Sorry to disappoint you, but setting the sharpness to high does not capture more detail, at all. In fact, it just overprocesses the image and makes everything look like an oil painting when zooming in, losing detail in the process.

The GoPro is heavily limited by its tiny 1/2.3" sensor, or 1/1.9" on the Hero 11 which comes from more area vertically. As a result it will in general struggle with the dynamic range, which is why capturing partially overcast skies is tricky.

2 Likes

I can only agree with that. @filipc, try to find as neutral settings as possible, even at the expense of beauty. Mapillary’s goal is to capture ground truth, so to speak. Or, maybe rather the truth on the ground, not heavily processed glitter. Anyway, you get the gist. Thus, it is best to refrain from digital processing as much as possible. If your images turn out too dark when capturing under an overcast sky it is usually far better to compensate exposure time by statically adding a few stops (do not increase sensitivity unless you really have to).

Addendum: Sometimes, although unfortunately rarely on action cameras, you can configure the “brightness metering mode” (or sometimes just called “metering mode”). For Mapillary purposes, it is usually best to set it to “average”. In this mode, the camera’s auto-exposure logic measures brightness linearly over the entire image for best exposure. Often enough, cameras are preset to the “center weighted average” or “spot” modes. These modes, as their names suggest, overly concentrate on measuring the brightness at the center of the image, and care less about brightness at the edges. Concentrating on the center (only) when measuring brightness will usually give you widely fluctuating brightness levels in your sequences, especially when your camera is even slightly tilted to sort of point at the sky at the center of the image, even on cloudy days. So, you may want to check that first. If you are unsure what metering mode your camera is using then you should be able to verify it by looking at the value of the MeteringMode Exif tag.

Have you tried to upload with GoPro captured videos with Desktop Uploader? @filipc

No, I can do it tomorrow. What format should I try ?

1 Like

Whatever comes out of the camera, I suppose, which would be .mp4 files. Or did you perhaps mean resolution/aspect ratio?

There is :

Standard
Full frame
Activity
Cinematic
Ultra Slo Mo

Going from 1080 to 5.3K

I don’t have a GoPro myself so I don’t know exactly what the interface looks like, but it’s probably best to leave it on Standard and edit the preset to suit. Switch resolution to 5.3K, 24fps, 16:9, WB locked, sharpness Low.
If shooting in constant lighting, use manual ISO and shutter, dropping the ISO as low as possible without making the required shutter speed too long. In bright daylight, ISO100 whenever possible.

1 Like

Nice of you to test video for us, @filipc. :+1: However, in my experience, I would not bother with video in the Mapillary use case because there are imho two huge downsides to video.

  1. Geo tagging and time stamping a video can easily become a pain, especially when something goes wrong, even with good tools at hand. One bad sector on the SD card can be enough to make a complete and potentially long video file unusable/unprocessable. On paper, geotagged video should be great for Mapillary but in real life; meh, not so much.
  2. Video is not designed for good static photography (what Mapillary imagery actually is). Basically all common digital video codecs are optimized for motion rather than static expositions. Thus, static imagery bitrates and encoders are usually vastly superior to video bitrates and encoders in this scenario, even when comparing to video key frames only.

Switch resolution to 5.3K, 24fps, 16:9, WB locked, sharpness Low.

I would rather go with the sensor’s native aspect ratio here, which if I am not mistaken is still 4:3 (I would assume “Full frame” here but maybe also “Standard”) is 8:7. But otherwise yeah; maximum resolution, lowest fps, no undistortion, and maximum bitrate is the way to go.

The aspect ratio of the Hero 11 Black is an exotic 8:7 :smiley:

Right, I guess it is a tribute or foresight to 360° pano camera setups which basically produce 2:1 images with equirectangular projection.

Version 206
MP4 gives

You need the latest Desktop Uploader…

Done and done.
The updates used to be automatic and they show no version date.

:+1: Give us an update when the video has been processed. I wanna see how Mapillary processes it.

And the next three sequences captured by “none none”.

1 Like