Messages

1

Bug Reports / 2.2.1 build 20443 - can't decompress depth error

« on: April 23, 2025, 02:23:26 AM »

Hello,
I try to create the mesh from a project which has 545 drone photos and 197 laser scans but no matter what settings I choose (low, medium, high) I always get this error "cant decompress depth". I tried to hit the retry button some times, it looks like it proceeds but it's painfully slow to continue like that. Is this a bug? With this pc I have successfully processed many similar projects without this error. I have also set my virtual RAM manually to 512GB and I have plenty of disk space on my ssd. Does anyone know how to avoid this error?

2

General / Re: Agisoft Metashape 2.2.0 pre-release

« on: January 13, 2025, 10:33:22 PM »

This is not a powerful processing machine, no laptop is. You have bottle necks everywhere, Thermal, power, RAM, GPU.

I neglected to mention processor, of course that is relevant, perhaps the most relevant factor of all.
It is not entirely clear that the system is laptop unless the 13900hx is a mobile only version, or I missed where laptop is stated.

13900hx can only be found in laptops yes, its laptop CPU with very limited thermal room for processing.

The processing power of the 13900hx is not that far off from a 13900K(s) or even a 14900K(s) which is recommended as the best CPU for metashape by most benchmarks on the internet. On cinebench r23 I get 34000 points with -120mv undervolting and I avoid thermal throttling completely. The 13900K on cinebench r23 gets about 39000 points. It's not that big of a difference to describe this CPU as "weak". Also, I have undervolted my rtx 4090 which now boosts at 2200mhz and consumes only 130watts  (from the max 170) and the score I get in Timespy is 22000+ which exceeds a 3090ti desktop GPU which is also the last gen beast of a card for photogrammetry projects. So except for my limited RAM at the moment, I don't see any other reason why this project takes such a long time to process or there is a bug in the software.

3

General / Re: Agisoft Metashape 2.2.0 pre-release

« on: January 12, 2025, 11:19:02 PM »

I have a project with about 7000 photos and 15 laser scans and the alignment process reached pretty fast at 70% but now it takes about 24 hours to move by 1%. It is at 74% now and by that speed it will take a month to finish. Is that normal? The CPU is maxed out most of the time as well as my ram (which unfortunately at the moment is only 32GB but I have a lot of SSD free space to work with virtual memory). Is this maybe a bug of the 2.2.0 version? or is my limited ram the reason for this extreme slow processing?

You need multiple SSDs for best results ... store source data, project files, and OS installation (startup drive) all on separate drives (three drives)
I don't see a way to specify a Temp directory for Metashape, presumably this is stored on the startup drive.

You also need more RAM than 32Gb, virtual memory is almost useless because it will cripple the speed, although separating source data, project files, and OS installation on separate drives might help. That way the system is not reading and writing to the same drive.

The video card is also relevant, ideally you would have an Nvidia card to support CUDA with at least 6 - 8Gb.
If you have less than 6Gb on your video card, or are using an integrated GPU, that will also translate to a reduction in processing speed.

I'm sorry to say if you you don't have slots to allow for 3 SSDs and cannot add more RAM, I think you you will have to suffer through it until you can get a new computer.

I have two fast nvme drives in my PC and I have set the virtual memory to be on the c drive and on the second d ssd drive which is a 2tb I only have this project. My PC is very powerfull with a 13900hx CPU and rtx 4090 16gb vram. At the moment I have 32gb ram but I have ordered a 2x48gb (96gb) and waiting for it to arrive, so at the moment I have to work with the 32gb. My question regarding if this is normal or not is because I see the CPU is almost always maxed out it doesn't wait for data to process from the disks. Now it has reached 75% from 70% after 3 days. My client is waiting for me to deliver some orthophtos and I don't even know to tell him approximately when it will be ready.. Every day it passes the estimate time raises.

4

General / Re: Agisoft Metashape 2.2.0 pre-release

« on: January 11, 2025, 07:50:21 PM »

I have a project with about 7000 photos and 15 laser scans and the alignment process reached pretty fast at 70% but now it takes about 24 hours to move by 1%. It is at 74% now and by that speed it will take a month to finish. Is that normal? The CPU is maxed out most of the time as well as my ram (which unfortunately at the moment is only 32GB but I have a lot of SSD free space to work with virtual memory). Is this maybe a bug of the 2.2.0 version? or is my limited ram the reason for this extreme slow processing?

5

General / Re: AMD Ryzen Threadripper PRO 7995WX

« on: April 24, 2024, 06:40:16 AM »

Is the AMD Ryzen Threadripper PRO 7995WX (96 Cores | 192 Threads | 5.1GHz Boost) a good choice for the CPU to process large datasets??

No, just get a 13900k(s) or a 14900K(s) and 192gb of ram. Thats the fastest hardware for metashape.

6

General / Re: Agisoft Metashape 2.1.0 pre-release

« on: March 19, 2024, 12:45:37 AM »

Hello bassistas,

Can you provide more detailed description of the steps followed starting from model generation (also specify the coordinate system of the chunk, model and CS options in export/import dialog)? I was not able to reproduce any problem with export-import workflow for FBX model in the latest pre-release version.

The issue still exists in the latest stable release 2.1.0. The project contains photos from a DJI drone and laser scans imported as e57 files. The coordinate system is WGS 84, it doesn't matter which options I choose during the export or what file format, the exported model always comes up shifted by a few meters from the original position. I can send you an example project  (without the photos and laser scans) to investigate if you want.

Sounds like your problem is because you are using WGS84 as coordinate space, why not use a real coordinate space with accuracy and measurements, not something in degrees.

I don't care about the coordinate space as I don't georeference my model, I use the laser scans as reference which are in local coordinates and I just need to export the 3d model in local coordinates also. That is the workflow I always follow but now with the latest version of Metashape when I export the model it doesn't hold its original position and I cant re-import back to re-texture it after the cleanup in another program.

I just tested the previous 2.0.0 version and the bug does not exist there. I can export/import the model and it stays in the same position as it should

7

General / Re: Agisoft Metashape 2.1.0 pre-release

« on: March 16, 2024, 08:26:33 AM »

Hello bassistas,

Can you provide more detailed description of the steps followed starting from model generation (also specify the coordinate system of the chunk, model and CS options in export/import dialog)? I was not able to reproduce any problem with export-import workflow for FBX model in the latest pre-release version.

The issue still exists in the latest stable release 2.1.0. The project contains photos from a DJI drone and laser scans imported as e57 files. The coordinate system is WGS 84, it doesn't matter which options I choose during the export or what file format, the exported model always comes up shifted by a few meters from the original position. I can send you an example project  (without the photos and laser scans) to investigate if you want.

8

General / Re: Agisoft Metashape 2.1.0 pre-release

« on: November 23, 2023, 06:52:27 AM »

With this release, if I export the model and import it back it comes up in a different position, so its impossible to do any editing on the mesh on another program and then re-import it back for re-texturing. It seems that during the export the model is shifted by a few meters.
Also sometimes the model rotates by itself inside metashape if I decimate it through the decimate model command.

Can you please check, if the issue is resolved in the newly released 2.1.0 update (build 17343)?

Hello, I just tried this build, the issue still exists when I export an FBX file but when I  export an obj file and then import it back again it comes in the same position. So it is partially solved, at least for the obj format and I can continue working with my projects now, thank you!

9

General / Re: Decimated Model not aligned to original

« on: November 18, 2023, 04:26:33 AM »

I produced a model after aligning multiple laser scans to supliment the photogrammetry. I selected the resulting model (with an [R] alignment designation) and decimated it to create a 2nd smaller model. The resulting decimated model is tilted at an angle and not aligned to any other model or point cloud in the project (even in other chunks). This of course makes it impossible to build new textures from the images.

Metashape Professional v. 2.1.0 build 17262 (64 bit)
Windows 10 Pro v. 22H2 OS Build 19045.3693

It happens to me also, using the same version (2.1.0 17262), not always but in most cases 8 out of 10 times. There is a similar bug in this version, sometimes when you export the model it is moved by a few meters from its original position and when you import it back it's impossible to re-texture.
I hope they can fix this soon in the next update as I'm stuck with this version now, I'm unable to open my saved projects with an older version of metashape and continue my work.

10

General / Re: Agisoft Metashape 2.1.0 pre-release

« on: November 15, 2023, 09:08:19 AM »

With this release, if I export the model and import it back it comes up in a different position, so its impossible to do any editing on the mesh on another program and then re-import it back for re-texturing. It seems that during the export the model is shifted by a few meters.
Also sometimes the model rotates by itself inside metashape if I decimate it through the decimate model command.

11

Bug Reports / Re: Texture artifacts after mesh decimation

« on: November 15, 2023, 02:20:29 AM »

Hello MAVtrix,

Can you send the archive with the original and decimated models (for example, in OBJ formats) to support@agisoft.com? It would be easier for us to check the problem, as we weren't able to reproduce the issue on some similar projects.
Additionally please include the PDF report exported from the complete project.

This issue with the texture artifacts still exists in the latest 2.1.0 pre-release version. Are there any plans to fix this or do you have any workaround to avoid having them on my model?

Also at the moment, there is no way to have a smooth texture for indoors when using a flash.  If I choose Mosaic each photo blends nicely in terms of exposure but the texture is not smooth (in some areas is very sharp and high res and in other areas blurry. I understand this is because of the distance and angle I take each photo).  On the other side, If I select average mode, I get a nice smooth texture but with a lot of lines everywhere (check the attached screenshots). On top of these if I have a decimated mesh or if I have re-imported the mesh which I have cleaned and decimated on another program then I get all these triangle artifacts spreaded everywhere in my model and even if I want to fix them manually using, for example, the substance painter then also the UDIM support is still broken in 2.1.0 and substance throws errors when I try to import. It only works when I have a small model/object with just 1 texture.

12

General / Re: Circular bands in texture

« on: November 07, 2023, 03:32:43 AM »

hello

I'm doing some test scans of a room and sometimes I see these circular bands appear in the texture. Just now I repeated the 'build texture' step and the bands seem to be gone (see attachments). Luckily I uploaded the 'banded' model to sketchfab so you can still view it:

bands:
https://sketchfab.com/3d-models/corridor-test-312d91081890457db86e7f00c2e02367

no bands:
https://sketchfab.com/3d-models/corridor-test-2-f2a395148dfb4b8e94ec1143547c55f5

At first I thought it had to do with using images with multiple exposures. This resulted in the banded texture. I then used the manual setting on the camera and took all the images with the same settings, then the bands were gone. But as said, I just ran the build texture step again on the images with multiple different exposures (I used a single aperture and automatic exposure time) but now the circles are also gone..

So I was wondering if this could be a bug, or did I forget to change some setting? I used the 'mosaic' setting in the texture dialog in both cases as far as I know.....

thnx!

I have the same problem when I choose Mosaic, in average mode I don't get these lines but I also get other types of lines (check the attachment), which are the edges of each photo due to different exposed areas when I use on-camera flash. I noticed that this bug isn't present in the 1.8.5 version. They seem to appear only on newer versions of metashape.

13

General / Re: Metashape 2.0 Laser Scan Referencing

« on: November 03, 2023, 10:15:19 PM »

I'm hoping to revive this thread, as I attempted to use the script but didn't do so successfully. It's also a topic that multiple folks are attempting to tackle (https://www.agisoft.com/forum/index.php?topic=15939.msg68793#msg68793).

I ran the script with a slightly different workflow and didn't get anticipated results. Is anyone willing to provide some insight into how I might be misusing or misinterpreting the Python code?

• I imported laser scans and independent photos per the workflow documented here:  https://agisoft.freshdesk.com/support/solutions/articles/31000168474-terrestrial-laser-scans-processing-in-metashape-2-0-0#Processing-of-terrestrial-laser-scans-with-images
• I discovered this thread after performing several other time-intensive tasks (point cloud generation, manual classification, meshing trials, etc.). Consequently, I tried duplicating the chunk, re-importing a laser scan in its original position, and applying a transform matrix post-alignment.
• I'll give an example of one of the many trials I ran that correlates with the edited code below. Steps:  1) align laser scans and photos, 2) duplicate the chunk, 3) re-import a laser scan into the duplicate chunk and re-name it so there wasn't a duplicate "camera" name, 4) confirm that the re-imported laser scan loaded in the correct position by reviewing the associated estimated coordinates and rotation values in the Reference Pane, 5) run the script below in reference to the aligned chunk[0] / aligned laser scan and the duplicate chunk[5] / re-imported laser scan.

I won't confuse matters with the myriad of other solutions I attempted - any input is greatly appreciated!

Code: [Select]

# gets chunks to use
non_aligned_chunk = Metashape.app.document.chunks[0]
aligned_chunk = Metashape.app.document.chunks[5]

# function to get camera based on label
def get_camera(chunk,label):
    for camera in chunk.cameras:
        if camera.label == label:
            return camera
    return None

# gets a spherical from a laser scan for both chunks
non_aligned_camera = get_camera(non_aligned_chunk, "Ground-Ext-100")
aligned_camera = get_camera(aligned_chunk, "Ground-Ext-001")

# non aligned camera is in the correct position so we work out chunk transform to equal non aligned camera position
aligned_chunk.transform.matrix = non_aligned_camera.transform * aligned_camera.transform.inv()


What works for me is something very simple, first I mark some of my targets on the laser scans right after I import them in metashape and then I duplicate the chunk to preserve the original orientation of the scans. When the alignment of photos and laser scans from the first chunk is finished I align the first chunk with the second (duplicated) using my previously placed targes (which are now present in both chunks) and using the second chunk as a reference.

14

Feature Requests / Re: Weighted Average Texture Blending Mode

« on: November 02, 2023, 03:31:36 AM »

Actually I might try just masking out the flash area... too simple... doh

Edit: It didn't help greatly, I think the mottled effect is more down to general changes in the lighting, rather than the flash alone.

Have you managed to find a solution to this issue? I'm using the latest 2.1 pre-release now and I still get these lines when I use the average texture mode. If I select mosaic it comes out almost perfect but as I took the photos by free hand in different distances and angles I get some blurry areas with this mode as expected. Only average gives me a smooth texture but I can't get rid of these lines. I tried to mask areas, and readjust photos exposure in Lightroom but its a lot of trial and error until I see some improvement and it's still not good enough.

15

General / Re: Metashape to Substance Painter with UDIMs

« on: July 12, 2023, 06:14:21 PM »

I had this problem and found the solution in editing the obj with udims from metashape in c4D with arnold. (redshift will work as well probably). Then the obj will be imported in Painter correctly.

I'm trying to import an obj which has 65x 8K textures from metashape 2.0.2 to substance painter and I get this error:
''Errors have been found while splitting meshes per UV tile: Mesh 'defaultobject' has faces in out of bounds UDIM indices: [10,5].''

Also I tried to delete all lines with mtl from the obj but still does the same. I can open the obj in 3dsmax or any other viewer without any errors and the model looks fine. I checked also the uv tiles and they seem fine. Can you describe what you did exactly in cinema4d or if there is any tutorial about how to fix udims in C4D with arnold please let me know.