Suggestion/Feature Requests: RevoScan 5.3.0 -- Raw Data Edit related

I’ve spent some time with the Windows version of the v5.3.0 Revoscan software, and now have some suggestions to make. Overall, I’m quite pleased with how development has improved and refined the software, and really made my Mini a much more effective tool.

I encourage any discussion or input from other users about these ideas. I’ve tried to frame each suggestion so they could be implemented as incremental improvements, not features that remove other functionality.

  • Frame count Include the total frame count in the Raw Data Edit mode when selected. The total frame count is noted in the One-click and other “modes”, but that panel isn’t in view while in the Raw Data Edit mode.

  • Select frame by selecting a point or points workflow with cleaning up isn’t as simple as just choosing one of a dozen frames in some scans.

If there are 137 frames, or the geometry of the scan is complicated, the current way of viewing each frame by frame doesn’t as easily allow the user to pick out the out-of-place frames (a doubled nose, or a frame that is clearly misaligned).

An example of this:

If you could say… lasso or rectangular select an area that’s out of place, and the program can tell you that frames 3, 4, 85, 86, 87 of 137 frames have points in that out of place “selection” that was lassoed.

From the program giving the list of frames that have datapoints in the selection, the user can then check that smaller subset against the rest of the model.

  • A Cummulative Playback Feature when previewing frames in the Raw Data Edit mode, it defaults to showing each frame in isolation. What I am proposing here is allowing the user to toggle the play/pause and forward/back buttons to function more like the same buttons did in RevoScan 4.x

As currently implemented in 5.3.0, longer frame-count scans, over 100 frames scanned in a model, the “play” icon to review which frame is in view is not as effective.

Having the ability to toggle viewing from a starting frame, adding each in the view after a delay will also aid users trying to select the misaligned frames they want to edit out.

This feature could be enabled/disabled by a checkbox. I am fine with it not being the default, and then I have to enable it.

  • Add a frame playback “delay” slider High frame count on a scan, having the individual frames skip by at the current 3-5 per second may work for some situations.

This rate of playback should be user-adjustable, instead of one-framerate-for-all.

Easily implemented with a slider with the number for the FPS playback rate.


I welcome any feedback on these suggestions. Please point out if anything is unclear, or if screenshots / edited screen shots would further explain.

4 Likes

And, after reading my suggestions, I remembered the one I missed putting in the list:

  • Keyboard Shortcuts
    Either via a help file, or other document, or adding to the tool tips on buttons or controls, displaying what keys are mapped to any of the buttons in this mode would help immensely.

For instance:

I don’t know how RevoScan allows us to zoom in or out from a model. And there is nothing in the program that tells me on screen, or via tooltips, or other guides on how to do so.

Adding this to the program would make it much more useful. All the keyboard+mouse movement combos I’m used to in other 3d programs did not work as guesses.

2 Likes

Hi, totally agree with your requests, hope they will improve this feature too :slight_smile:

I think you have missed out some features that are implemented in frame mode.

  • You can edit the points globally if you wish and that will remove the points in all frames that have selected / delete points

  • The option hide unselected frames, if switched off will show all the frames but highlight in green the ones you have selected

  • If you shift + cursor move down/up you get the cummulative effect you are looking for you can even select multiple frames to see what that selection represents.

Be carefull tho, when editing raw frames (lasso / delete), you are incurring in a irreversible data loss that cannot be retrieved. the modification permsnently mangles the raw data on the depth map (where a mask would be more desirable instead)

Another thing that is missing is a way to rearm deleted frames, since that is just a flag in the inf file and no data is truly removes

1 Like

1st: I wasn’t after removing just points by selecting the points … rather:

I was suggesting there be an option to select a point/selection, and … it selectes all frames that have that data-point.

It’s a reverse-work flow of what is the only option currently:

  • I click the scan data in the horribly misaligned frame,
  • the program highlights that it’s in frames 159-165 of 173 frames in my scan,
  • I am saved the work to find which frames I want to remove by scrolling/animating each individual frame in a long 173 frame series

2nd/3rd :
I noticed that feature of it, going back to the program after you mentioned how it shows that cummulative effect.

So, we could call that one solved, but … under-documented.

I’m persnickety about undocumented behaviour. In my books, it’s a bug. This is still considered beta software, so I’ll give ample leeway for such bugs. But undocumented features/keys/buttons all lead to increased, repetitive support work, so documenting features helps reduce the volume of work left for support.

Third edit:
Forgot to acknowledge your point about the inability to “undo” deletes – quite valid,and important to tackle that!

Lastly:

Also, as of this writing, I need to re-evaluate suggestions, because 5.3.1.902 was just released. Maybe the dev team solved some things without having to hear from me? :smiley:

1 Like

I would add: every frame displayed resets the camera. Please add a way to lock the camera if resetting every frame is a desirable behavior. If you’ve zoomed in to try to see detail on a larger scan, having it zoom back out every time you advance the playback is not fun.

3 Likes

I think the most important suggestion here is selecting the points to reveal the misaligned frame , that would be the most easy way to delete it , because right now you can scroll as much you want but finding the right frame is not an easy and time consuming task .

I knew it will be like that and I said it so many times …

The editing by selecting the Raw points and the ability to remove it from multiple frame cells before fusing at once is a big bonus .

But if I have to scroll over 4K of frames to find the bad one , I would prefer to make a new scan , it would be much faster .

Remember one frame cell shift meaning the rest of it is shifted even if you don’t see it straight away… banana bends happening and are hard to spot until you ready to do some reverse engineering and realize what is going on . Of course it is Algorithms issue in most cases … especially when using marker mode with big objects …

Always scan with smaller sections if your object is long , you will get lower chances with banana bends .

4 Likes

This somewhat reminds me of how the <<Beginning Forward >>to the end “buttons” that RevoScan 4 used to have were crude, coarse controls, but … effective.

Old 4.x workflow during the scan: If I, or the software tracking messed up, I could pause, “rewind” to a rough point where the scan was still clean, and resume.

Having the granularity to select individual scan frames is so much more fine-grained, and detailed. But the interface lacks the coarse but effective controls that the prior version had in it’s place.

With a part I’m trying to integrate as a project now using my Mini (half of an air filter housing), I’m currently up to 35 scans, with 28 merges interspersed between them.

I am going the route of small scans for this like you mentioned, PUTV, and now have to resort to a “crib sheet” … aka, spread sheet of the scans, the merges, which scan are in each merge. There are also 8 “merge strings” where I am glomming more and more scans into subsequent merges.

Why would I be doing all this? Because: I wanted to keep from getting to a point where I had bad frames in one of my 35 scans prevent me from finishing the merge of the inside and outside shells!

No surprise, I have one of my later merges showing a “jitter” where the scans did not feature-merge perfectly.

My current workaround: Now I need recheck up to 28 different merges, comparing against my spreadsheet, so I can check which of my 35 scans are a part of it, and do a “Preview” in the Merge and see which of the scans has the handful of bad frames that needs editing out.

I don’t want to ask for better tools to manage this huge list of scans than the current scrollable list. Further improvements with the tracking/aligning will solve this out better. So would improving the frame-editing tools. Or, just re-implementing the 4.x feature while scanning, where the user could “Rewind” to the point in time where the scan was clean, and resume from there, forgetting the bad data.

More that I wanted to point out how an interface that is lacking can really, really affect workflow: Having better tools in the frame-editing portion of RevoScan 5 would help users in all manners of situations, dropping the complexity to hunt out misaligned scan frames out of hundreds.

afik you can go back frames while scanning and also navigate the frames with up/down arrow after scanning in raw edit like a movie with or without all frames showing also you can delete multiple frames by shift move up/down and clicking in one delete button of one frane to delete the selected lot

Thanks for pointing this out!

This behaviour of the program needs to be documented, right in the program itself, if possible.

It’s still a very clunky manner to search and through anything more than a few dozen scan frames. The programmers should be showing the frames selected with, say, a bar on the bottom of the model viewport, where you can visually see that

  • frame 0 is the first frame,
  • frame N is the last frame.
  • highlighted/selected frames show in the bar.

The listbox with 12-15 entries is a pinhole view of the entire set of 187 to 453 frames per scan I often have. I feel that we need frame editing to be more intuitive than this.


Edit:

I realized I should’ve stated this - it’s in my mind, didn’t express it well.

The downfall of the shift key selected frames method, to me:

It’s either those selected frames show in the viewport, or an individually selected frame, both in in green. Or, the entire scan, in shown in blue.

This either-or choice doesn’t display enough info to me, to effectively tell if I have selected the correct frames that I want to delete, or selected enough of them. Selecting automatically shows me a subset of 1+ frames instead of all the point cloud, AND the frames selected in green that need to be deleted.

I completely agree with you. The ability to edit raw data can be incredibly beneficial. It allows for the removal of unnecessary information right from the start, and any errors can be corrected immediately. This not only improves the quality of the data but also enhances the efficiency of the workflow. It’s great to see that you’re also in favor of this feature! :blush: