Provide feedback

How to share your feedback

Feedback, findings, or examples of beta feature usage can be shared with us on this Beta program community page for the Arkio team and other Beta users to see. If something is marked as WIP or “Known issue” in the release notes we might be working on a solution but we’d still love to hear what you think and how we can improve these features further.

Share feedback personally

If you want to share your findings personally you get in touch with us by Email or Slack (ask us to get invited) You can also make recordings directly from the Quest with voice and submit them to us using Meta’s “Submit Playtesting Feedback” feature. We also love to jump on a live call to go over your ideas together as using a meeting in the beta version is the most direct way to share your findings, Arkio scenes, and issues with us.

Reply with your feedback here

Everyone is welcome to share feedback with the Arkio team and other beta users in this forum. You can reply to this post or make a new post about a topic you would like to discuss.

As a guideline here are some questions that you can answer:

  1. Are you missing anything that should be clearer in the app or that we should describe better in this forum?
  2. Did you encounter any problems or bugs while trying these features?
  3. Is there anything in particular that you love about a feature?
  4. Are there any areas where you think these features can be improved?

When describing a bug/issue it would help if you could include the following info so our team can more easily find and fix the issue :

  1. Title
  2. Environment (device, version)
  3. Steps to reproduce
  4. Expected result
  5. Actual result
  6. Visual proof (screenshot, video)

We’ve been hearing great feedback from some of our beta users so far mostly focusing on the new AI camera improvements.

Since 2.0 will focus on overall usability of Arkio we’d also love to hear your feedback on other features that are in this beta release or your ideas for other improvements that are might be include for our 2.0 roadmap

You can reach us by mail or leave a message here to open a discussion with other beta users!

1 Like

These are some early stage images generated from the model illustrated above. The model was built on a sketch and you can see how the AI interprets the sketch lines as various landscaping elements. I plan to test along the same lines in sketchup and see if control of AI results can be improved with more concise environmental [model] input. Having an Archicad/Arkio connection would stream-line this further so am excited to see how it develops!

Some feedback so far on Sketchup integration - with cloud. When you place the model as ‘fixed at origin’ it imports the model at the ground plane. This means that if you have geometry below the ground plane in sketchup it will disappear into the table. Given that most sketchup site models will have some ground level geometry, this could be possibly optimised - maybe with an additional input option that allows the ground to be established - or better still, with the lowest geometry from the sketchup model establishing the point of contact on the table?

Also noted that the table doesnt resize evenly around the imported model. for both of the above points, I have found it easier to not use the ‘pin at origin’ and just place the model as a moveable object on the table instead [and resize the table manually].

On the positive, the model update feature is excellent and very workable with ongoing development of the sketchup model.

1 Like

Thanks for sharing these @Greg! It’s great to see the results and we’re curious to hear how your other experiments with imports go.

Regarding the model placement in the scene. We load models at the origin for round-trips to the same coordinate (also below your origin) to align well with your source software (eg Sketchup). Would you consider these round-trip alignments to be important? Would you prefer more import/placement settings during the import or better ways to adjust and align a model once it was placed in the scene?

1 Like

Hi Johan, thanks for coming back to me on this. I’ve experimented a bit more and have produced a more controlled set of images as posted above by:

  • creating a clearer context for the AI to understand via Setchup
  • keeping the architectural language of my Arkio model simple [ie windows, characters, trees, cars etc as basic scale prompts]
  • refining the word prompts for the image production [ie modern apartments trees in foreground at night street view… etc]

With regard to sketchup placement, I’ve played around with it some more and the origin doesnt matter as much as for example an image file pinned to table, because the cloud file automatically updates the model when its re-exported from Sketchup.

The opposite is true of image file placement - this had been my previous workflow with Arkio using sketches produced in TracePro on Ipad [or CAD drawings as PNG files] and brought into the Arkio table as a base for a volumetric model. With that set-up I would need to start a new Arkio project with each iteration to make sure the scale was consistent. However, with the cloud import on SKP, this process is automatic, so the ‘pin at origin’ option is moot.

One thing that could make life easier would be the ‘snap’ availability of the imported skp model. When I place my SKP model on the table , it would be great to have an option to:

  • lock it in position and
  • make it ‘snappable’

As mentioned above, to get a better handle on the AI output with magic camera, the environmental prompts can heavily steer the generated images. It would stream line the whole process toward image generation if the imported model [ie the environmental context generated in outside software eg SKP] could be snapped to with haptic feed back on its ground planes so that models created within Arkio can be placed directly on its surface. On the above generated images, I was not getting haptic feedback on model placement so some buildings are sitting below the visible ground plane [an environmental element brought in from SKP].

Final notes, and these relate more to the commercial release rather than the beta [from where I was modelling]:

When I used too many of the ‘building’ components within the Arkio library on an apartment building, things went a little awry. I tried to group and copy the building I had created in session and suddenly all my in session geometries disappeared! I am certain this is not a PC hardware issue but may be related to the headset or software. [Running session on Quest 3 via Link Cable on a high powered PC]. I was able to recover things by going into ‘flat mode’ on PC and opening a history file or re-opening the base file. However, the issue would repeat each time I tried to copy or group too much f my generated models within the VR session. Maybe this is a ‘me’ thing but thought it might be worth flagging !

Second, I noticed from SKP import that I had to switch off certain components [ie railings] to make the session workable. When I had railings on [imported randomly rom 3D warehouse] the 3D session was jittery and it took a few minutes to export from SKP to Arkio. As soon as it turned off the layer in sketchup with the railings, the export took one second and the VR session was smooth.

I’ll revert with other findings as I go. The above images shared are from an ongoing design process they are both 'beta tested and a useful part of my current workflow! G

This is great feedback @Greg thanks for sharing these thoughts and nice examples of your projects and workflow here! We will collect some of these findings and best practices for working with the generative AI camera and have some updates for the beta coming soon that will make it easier to work with the camera.

The feature you are describing for making mesh shapes “snappable” is something we’ve been discussing internally and involves introducing mesh face normal calculations to Arkio’s modeling engine. We keep investigating that further as we see more use cases for this and we will also look at what we can do to simplify the import placement and pinning workflow.

Thank you for the opportunity to test the beta version of Arkio 2.0. I’m excited about the potential of this platform and appreciate the hard work you’ve put into this upgrade. Here are my initial thoughts after a brief testing period. These are just initial impressions, and I apologize if some of the issues are due to my own mistakes and not related to this new version.

Panel Size: The size of the panels seems too small, making the text difficult to read. This may be due to using the Oculus 2 headset, or not.

Revit Integration: The process of accurately modeling in Revit and then reviewing and sketching corrections at real scale in Arkio is excellent. The cloud-based data transfer is very convenient and simplifies the updating process. However, the timestamp of each file differs from my local time, making version control challenging. I also hope for future multilingual support for file names.

Controller Vibration: Occasionally, the VR controllers vibrate continuously for no apparent reason.

Navigation: Moving around in the virtual environment is currently the most cumbersome operation. It would be great if walkthroughs and other navigation features were more user-friendly. When I invite collaborators, they struggle to keep up.

Magic Camera: The Magic Camera is a lot of fun and feels like a glimpse into the future. Will it support other languages in the future? It’s challenging to convey subtle nuances in a language other than my native tongue.

“neutral” mode: While navigating around a model, I often accidentally move, transform, or paint objects. It would be helpful to have a “neutral” mode where no commands are selected, preventing accidental modifications. Otherwise, it’s cumbersome to have to lock objects that are still under consideration every time.

Thank you for your good work. I look forward to seeing the continued development.

1 Like

Hi Jin,

Thanks for sharing your feedback here. We will take into account the different time zones and support for other language characters as we were not aware of these causing some difficulties.

It is already possible to speak another language when making your AI prompt in Arkio but we don’t show and type non-ASCII characters in our UI yet. But this is something also other non-English users can try out.

Neutral mode is currently possible but a bit hidden. You can deselect tools in Arkio by clicking the active tool icon on your bracelet or by wiggling your hand left and right quickly to disable the active tools on that hand! :grin:

We’re also planning some updates for the navigation using a smooth navigation and a new presentation and view-only mode that makes it easier to automatically to show around people new to VR in and without a meeting.

I hope some of this feedback also resonates with our other users. Let us know if other things come to mind for us to consider!

1 Like