Review of Ricoh Theta S spherical camera

This reviews describes the Ricoh Theta S in the context of Mapillary. This will give more info about some aspects and less about others.

Pros

  • Reasonably priced.
  • Good image
  • Best stitching
  • Good form factor
  • Battery lasts for 500-600 images.

Cons

  • Min 8 seconds between images in timelapse mode
  • Slow getting images off
  • Slow deletion process
  • GPS tagging is unusable for Mapillary.

Review

The camera is reasonable priced. The image quality and resolution could be better, but for the price it is very good. The spherical image is stitched from 2 images and the stitch line is not seen in general. There are always issues in the bottom (where the camera removes itself) and often in trees and foliage.

Form factor is nice and fits in a pocket. It comes with a neopren pouch, which will prevent the lenses from being scratched but not from damage. I always put it in a plastic box. The camera can be triggered using a button or an app. It is hard to avoid moving the camera when pressing the button, and use the app when the light is less than perfect. Tip: In the app, disable preview and postview to get it ready again as soon as possible.

At the highest resolution it can be set to take an image every 8 seconds, in a lower every 5 seconds. This is fine for walking, but on a bicycle I would at least use another camera to supplement the images.

Position will be tagged on the images if the camera is connected with the app (and the setting is enabled). Unfortunately the app takes the position raw from the GPS, so any GPS error ends up in the images. In my latest 3 smart phones, I some times see that the GPS jumps between where I am and a point up to 100 meters away. Perhaps one time, perhaps 10 times. The app will also record this, which means you cannot trust the tagging. Use your favorite loggin app (hint: OsmAnd for Android) and geotag the usual way.

When you get the camera, update the firmware. The factory version is buggy and very unstable. It cannot be done using the Android app! Only Windows and Mac. That is dumb, because the only ting the Windows app does, is that it uploades a file to the Theta S.

Getting images off the camera is slow. Wifi is very slow, both to copy and delete. It works better to plug it into a computer.
I would not get it to work reliably under Linux, but it works fine under Windows. Using an USB OTG cable I have end up plugging it into my Android phone. The standard transfer mode is very unstable too (I guess they have hacked their way to get it to run under Windows and Mac) but there is a secret trick: Turn the device off. Hold the wifi-button and shutter button in at the same time. Then connect the camera to the phone/computer. Then it will enter the good old USB mass storage device mode in read-only mode. My Android 6 will only look for new USB devices if I press a special button, so here I can attach all cables, press the key combo on the Theta and it will be found. It will only transfer with 5-6 MB/s (Usb 1.1 can do 10 MB/s) so it is slow.
I delete images using wifi. This is a pain too, because the app will start fetching thumbnails of all images. Then I have to press a select all button (a process that takes along time) and then finally the deletion can start, which can easily be 30 min for 1000 images.

The camera have an auto mode, a fully manual, an ISO priority and shutter speed priority. Be careful using anything but auto. If you e.g. say shutter speed must be 200/s, then if the light is too bright the image will be overexposed. This is not a problem with the camera. Any camera with this setting will do the same. It just does what you tell it: Shutter speed 200/s. The manual modes are great if you knows what you are doing and can use the preview.

The camera holds 1600 images, which is enough for walking for 3½ hours straight. You then have to offload them - an USB OTG cable is good. Remember to offload during any significant break and have a battery power supply for both phone and camera when shooting.

Later I will write some mounting tips.

Uploading to Mapillary

Just as with an action camera. Python scripts or the web. The camera has an open standard API, so perhaps Mapillary will support it some day?

Summary

With the current state of the technology I think this is a great camera for Mapillary. I think it is the best way to take images for Mapillary, because it shows everything. It with 1 images / 8 seconds it will kill your stats brutally, so no more top 5 in your country.

2 Likes

Thanks for the review! We just bought a LG 360 cam, which also supports Open Spherical Camera (OSC), and I have started to add support for it. Hopefully then the Theta S will work from the app as well :slight_smile: The LG seems faster, more like 1 images / 2 seconds, but haven’t measured.

I dissed the Ricoh cam because of the long delay between each picture, and because the Ricoh is somewhat more expensive than the LG.

But I also have low tolerance so I want any cam to work straight with the app. Downloading manually, using scripts or adding gps info by hand will make me skip the 360 cam.

I read reviews and specs on the Samsung, the LG and Ricoh.

I did not select the Samsung because it depends so much on a Samsung phone (which I do not have).
I did not select the LG for two reasons: In Two new spherical cameras you @Patrick describe the app as buggy and you have not had it working with Mapillary. The latter should be fixable, but an app that crashes often is a no-go for me.

In other reviews, I had read, that the Theta had the best stitching by far of the 3 cameras. It helped that I found it at a price that is not too far from the LG.

I needed the camera for earlier this week, so I did not have time to wait for any improvements. Stability and image quality was key.

If the LG has gotten better than your first review, it is awesome, because it is so cheap and should perform pretty well.

@Anders it shounds great. I can push the Theta to be a little faster by triggering each image manually, so perhaps a Mapillary app that does not rely on the built in timelap will do better.

I hope one or both of you will write a review of the LG camera with regard to Mapillary. There is a solid lack of info which is important in this regard.

I’m not a good program developer any more (at least 20 years since I developed last time), but is it possible to code a generic support for any 360 cam?
@Anders mentioned “Open Spherical camera API”, that many cams support.

My dream would be to get hold of an NCTech Iris360 cam in the future, but it’s ~$1600 so it’s unreachable for the moment - but gives extremely good pictures!

A generic support is not possible. Some cameras support Open Spherical API while others (like the Samsung) only have a proprietary API. It is always possible to reverse engineer the latter, but it often takes a lot of work.

Today I just had it take 800 pictures on a single charge. The camera can hold about 1600 images in its 8 GB of internal memory.

I was able to test this cam for a few days.
As @tryl explained the app is pretty slow to transfer the photos. But it is much faster if you just press the trigger raised arm without a smartphone connected.

Here are some of my experiments on a bike:

and this one walking:

1 Like