Matterport no longer have skilled support staff to stitch models together.

Quote:

Originally Posted by leonherbert
I also want to publish to google street view, how am I to publish that since they are two models on top of each other.

You can try to do it with pano2VR. Publish each model to GSV, then download all 360 into pano2VR from all models at once into a new Pano2VR project. After that find ends of each model and link them together and republish all 360 to Google.

If you have access to local guides connect street view board I have explained joining multiple models there in more details
https://www.localguidesconnect.com/t5/Street-View-Discussion/Virtual-tour-of-Big-area-with-no-concrete-maps-business-page/m-p/2003043#M37982

BTW, that is a good way to get all hi res 360 from Matterport tour without any hacking and risking being suspended from MSP. It costs to pay Matterport for publishing to GSV but it is worth it. You can use them to build a custom self hosted tour from them using Pano2VR. For example you can use a Matterport camera to capture hi-res quality scans, publish to GSV, then get all 360 from Google for virtual staging and present it as a virtually staged self hosted tour.

I am going to use it soon with my new client as they need their apartment tours virtually staged. Since there will be no use for a dollhouse view in their Pano2VR tour I am going to do as less scans as possible to build a Matterport model, push it to GSV and get all panoramas from there into Pano2VR. The idea behind it is to get a 360 quality of a Pro2 and help my client to pay less for staging each tour.

337 Website San Francisco	337 Jwbuckl private msg quote post Address this user	plus 1 plus 1 plus 1 0
	@leonherbert I was able to get some time with one of our stitching experts. This is what was explained and no it is not cut and dry, unfortunately. "In the end, the number of scan positions is only one small factor in a very complex process that determines whether the complete model - stitched or not - can be processed, is of high quality, and can be easily and smoothly navigated to optimize the viewers' experience. What some may not understand is that stitching is not combining two or more completed models together. What we do is combine two or more manifest files together. What does that even mean? When working with Capture to create a model, among other things, there are two types of files being generated. One is the scan data itself. This is the .SWL files within the model data. Every scan position generates a .SWL or what we call a Swil file. The other file being generated is the manifest file. This is a very simple code-based file that records the coordinates - amongst other data - of every Swil and marking captured and set throughout the scanning process. The Swils and manifest files are then uploaded to the processing engine where a complete model is created. When stitching, we pull the manifest files previously uploaded from the two models, combine them and upload them back into the processing engine so that a "stitched" model gets created. While that sounds simple, it's a very manual process with many steps required to properly combine the two manifest files. There are many problems that could prevent two files from being properly stitched - including not having enough overlapping scan data in the models being stitched. The processing engine, being considerably more powerful than any iPad, is used to tweak the alignment between every scan position to guarantee the highest level of accuracy. If there isn't enough alignment data between scans being stitched, the processing engine could throw everything off, thereby requiring another attempt or possibly a failed stitch that simply won't work. In the end, what you end up with is a complete model just as you would had it been scanned all together via Capture. Meaning, the same rules apply. Matterport can still only support models that reach a total of about 200 - 300 scans not only because of what can be captured but more so because what can be processed and even more critical is what can be easily navigated via most computers in the world. It is not in anyone's best interest to max out the capabilities of the platform and tech since if the end-user viewing the model has a poor navigation experience that reflects poorly on Matterport as well as the MSP. Therefore, setting proper expectations with Matterport's customers and for them to set proper expectations with their clients is paramount. Lastly, I think it's important to also understand that while we want to give an easily understood limit such as 200 - 300 scans as the official size limit, the reality is not that simple and the reason we don't have a hard limit when in Capture. The actual limit - and where the problems begin - is with the amount of surface area scanned. Meaning, a single outdoor scan with the BLK, for example, could capture substantially more surface data than a typical Pro2 scan inside a house. With so much more data per scan, you may not get to even 200 scans before the navigation experience begins to suffer whereas scanning large warehouses with hundreds of scan points may not have much surface area at all comparatively and can get to large sizes without much issue. Not the straight forward answer I am sure you were looking for, but at least it is an insight into to the tech logic behind the process.
	5 years ago	Post 32 • IP flag post

WGAN Standard Member 2.87K Website Las Vegas Linked In Facebook Twitter YouTube Instragram Matterport Service Partner Google Street View Trusted Photographer Zillow Certified Photographer We Get Around Network Member	2.87K VTLV private msg quote post Address this user	plus 1 plus 1 plus 1 0
	That's a tough one. You clearly see the scans in nearly identical spaces for floors 3 and 4. Staircase gets blown out from the sun (on floor 3 maybe) and now you have a 4 week wait going on for something Dee could prioritize spin around in a day or so. We have some "Partners" willing to pay a fee on top of hosting 2 models plus a stitching fee to host large models. Then we have legitimate issues like the one @tejas_mithapelli shares with an obvious need for attention getting zero response from his support ticket. Followed by a shove on MOUG and now coming here for help. I appreciate @Jwbuckl coming here to share the backstory as SuperDee cannot simply cannot be replaced. Currently looking at 4 people to do what she was managing. What a poor choice to let her go. Looking forward to the "Where is Dee Johnson" interview on Tuesday for some piece of mind. Frustrating to pay a company large sums of month over the past 4 years. Now knowing a 500% price hike is lingering out there with an inability to support the loyal "Partners" that continue to pay into the system. As Lando mentions in the above video; "This deal is getting worse all the time"
	5 years ago	Post 26 • IP flag post

WGAN Fan Club Member 4.18K Website Queensland, Australia Matterport Service Partner Google Street View Trusted Photographer We Get Around Network Member	4.18K Wingman private msg quote post Address this user	plus 1 plus 1 plus 1 0
	Quote: Originally Posted by leonherbert I also want to publish to google street view, how am I to publish that since they are two models on top of each other. You can try to do it with pano2VR. Publish each model to GSV, then download all 360 into pano2VR from all models at once into a new Pano2VR project. After that find ends of each model and link them together and republish all 360 to Google. If you have access to local guides connect street view board I have explained joining multiple models there in more details https://www.localguidesconnect.com/t5/Street-View-Discussion/Virtual-tour-of-Big-area-with-no-concrete-maps-business-page/m-p/2003043#M37982 BTW, that is a good way to get all hi res 360 from Matterport tour without any hacking and risking being suspended from MSP. It costs to pay Matterport for publishing to GSV but it is worth it. You can use them to build a custom self hosted tour from them using Pano2VR. For example you can use a Matterport camera to capture hi-res quality scans, publish to GSV, then get all 360 from Google for virtual staging and present it as a virtually staged self hosted tour. I am going to use it soon with my new client as they need their apartment tours virtually staged. Since there will be no use for a dollhouse view in their Pano2VR tour I am going to do as less scans as possible to build a Matterport model, push it to GSV and get all panoramas from there into Pano2VR. The idea behind it is to get a 360 quality of a Pro2 and help my client to pay less for staging each tour.
	5 years ago	Post 27 • IP flag post

31 Penebel IDN	31 WayanMatt private msg quote post Address this user	plus 1 plus 1 plus 1 0
	@wingman .... this is a great and so obvious idea ... (why didn't we thinks about it before ? ) thanks for sharing, as this is the only solution to provide a quality "fully immersive experience" and avoiding this stupid advise to play with mattertags, which basically is not what clients want ! you're a legend
	5 years ago	Post 28 • IP flag post

WGAN Fan Club Member 4.18K Website Queensland, Australia Matterport Service Partner Google Street View Trusted Photographer We Get Around Network Member	4.18K Wingman private msg quote post Address this user	plus 1 plus 1 plus 1 0
	you are welcome but remember it will only work on Google Maps. Your Matterport models will stay disconnected. Also be careful with floor allocations. That was not a problem with my case as I had a few big matterport models with all floors that needed to be joined with an outdoor path done with a Theta V. Plus if you read my post on localguides connect you have probably noticed it is not an easy task when you have a few models to join together and they have 1000s of panos in total. Pano2VR will be slow and buggy with this amount. They all need to be loaded from Google Street to your PC into a new Pano2VR project in order to join tours together and they will take a lot of computer resources to handle this amount. If you load just ending panos from each tour that needed to be joined together you will lose all other links. I believe though that Pano2VR developers can make it super easier if they add some checkbox and a code in their publishing tool that will tell Google to add new links and left intact all other existing links. That way you can just find ends for each tour, load only them into Pano2VR link them and only this new link will be added and nothing existing will be overwritten.
	5 years ago	Post 29 • IP flag post

903 Cincinnati, OH	903 leonherbert private msg quote post Address this user	plus 1 plus 1 plus 1 0
	@Jwbuckl Thank you
	5 years ago	Post 31 • IP flag post

WGAN Fan Club Member 32 Website Pune, Maharashtra, India	32 tejas_mithapelli private msg quote post Address this user	plus 1 plus 1 plus 1 0
	@Jwbuckl Thanks for the insight. Went through each detail you mentioned, especially the setting expectations one. The support ticket i raised for is a simple 137 scans project for a 3 storey office space where i had no alignment errors while scanning since doing this for more than 5 years now. While the processed model i received missed an entire floor which i found to be merged with another floor all together. Where do i stand in terms of expectations ? There still isn't any positive response form the support. My mail was replied once saying they're looking into it. This tour was to be showcased at a business event in DC all arrangements have already been made for and now we are here taking it from the client. Now please explain how do I deal with this situation.
	5 years ago	Post 33 • IP flag post

Matterport no longer have skilled support staff to stitch models together.10231