TDA4VM: May I ask what tools can be used to generate a pod model? Similar to "sedan_generic. pod".

Hello Yueqian,

The POD format is a proprietary format used by Imagination Technologies. There is an extension you can use with Blender if you wish to continue using the POD format. Please see this link:

https://developer.imaginationtech.com/downloads/

See the PVRGeoPod tool. Let me know if you run into any issues using it.

Thanks,

Erick

Hi Erick，

1.What do these parameters in the code mean?

		foldername   : "sedan_generic",
		filename     : "sedan_generic.pod",
		scene        : NULL,
		vbo          : NULL,
		indices      : NULL,
		texture_ids  : NULL,
		xrot_degrees : 90,
		yrot_degrees : 0,
		zrot_degrees : 0,
		scale        : 20.0f,

2.If I want to replace the car model of "app_srv_camera", do I just need to replace the pod file?

3.Is the "SedanV4_white. bmp" file mandatory? If it is necessary, how can we understand this document? Are there any special requirements for the layout of the car models in the diagram? How should I generate this BMP file?

Regards,

Yueqian

qian yue said:

1.What do these parameters in the code mean?

These are configuration parameters for how the framework in the code to render the car models. I think the parameter's are self-explanatory, but I've listed what I believe are their uses.

foldername: In the car model folder, the top-level directory containing the files required.

filename: The filename of the POD file in the "foldername" directory

scene/vbo/indices/texture_ids: Alternative parameters to "filename", but these seem out-dated and not used in the current demo.

xrot/yrot/zrot_degrees: Rotation degrees

scale: Scaling factor for the model

qian yue said:

2.If I want to replace the car model of "app_srv_camera", do I just need to replace the pod file?

Yes, that should contain the information required for the car. However, the code has not been updated in a long time, and POD files are deprecated by IMG on newer GPUs, so we don't expect bug support if there are issues with POD compatibility with the GPU.

qian yue said:

3.Is the "SedanV4_white. bmp" file mandatory? If it is necessary, how can we understand this document? Are there any special requirements for the layout of the car models in the diagram? How should I generate this BMP file?

I don't think this file is used, I don't see it referenced in the code. Perhaps it is a remnant of older code in the project. The POD file might contain all the information needed, but we would need to see if the textures are part of the POD file itself.

Regards,

Erick

Hi Erick,

Thank you for your reply.

You mentioned "However, the code has not been updated in a long time, and POD files are deprecated by IMG on new GPUs, so we don't expect bug support if there are issues with POD compatibility with the GPU.".

How can I replace the current POD with an IMG in the code to achieve the same effect? I think generating IMG may be simpler than generating POD files. But I am not sure how to make the corresponding modifications in "app_srv_camera".

Regards,

Yueqian

qian yue said:

How can I replace the current POD with an IMG in the code to achieve the same effect? I think generating IMG may be simpler than generating POD files. But I am not sure how to make the corresponding modifications in "app_srv_camera".

This can be a complicated issue to replace the POD framework being used, since it would require to re-write parts of the car model rendering in the demo. Some of the utilities being used for the POD model rendering will need to be replaced with the code to read and parse the new model framework you decide to use. There will also be a need to rewrite the code that renders the model in the scene, moving away from the utility functions and replacing with other implementations.

In summary, it won't be a trivial task to go back and remove the POD model dependency from the demo and replace it with a new model. My recommendation is to try and generate a POD model for a car using the legacy add-on to Blender and then see if that works.

Regards,

Erick

Hi,

May I ask how to configure the conversion using "Convert a Stand Alone File to POD Using" in PVRGeoPOD software?

I tried to convert the dxf file to pod format according to the operating instructions, but the information in the above figure was not modified. Then, I replaced the original "sedan_generic. pod" file with the exported pod file, which caused segment errors during code execution.

Regards，

Yueqian

Hi Erick,

Can I understand it as freely setting these parameters? Among them, "xrot_degrees: 90, yrot-degrees: 0, zrot-degrees: 0," represents the angle of the vehicle body presented in the surround stitching view, "scale: 20.0f" represents the scaling ratio of the model, which can be adjusted to the size of the vehicle presented in the view?

Regards，

Yueqian

Yueqian,

qian yue said:

I tried to convert the dxf file to pod format according to the operating instructions, but the information in the above figure was not modified. Then, I replaced the original "sedan_generic. pod" file with the exported pod file, which caused segment errors during code execution.

There might be allocated memory for some of the models, and perhaps your POD file is larger than the one used by the original models. Can you please try a very simple model to start?

qian yue said:

Can I understand it as freely setting these parameters? Among them, "xrot_degrees: 90, yrot-degrees: 0, zrot-degrees: 0," represents the angle of the vehicle body presented in the surround stitching view, "scale: 20.0f" represents the scaling ratio of the model, which can be adjusted to the size of the vehicle presented in the view?

Yes, these should be able to freely set these parameters. You can test them out to see how they change the view of your rendering.

Regards,

Erick

Hi Erick,

OK, may I ask if it is possible to use a 2D dxf file for conversion? Because I found that in "app_srv_camera", the car model here should be three-dimensional because it can present views from ten angles.

In addition, does the content in the above figure need to be appropriately configured? If necessary, how should I configure it?

Regards,

Yueqian

Yueqian,

I'm not sure about the 2D dxf file for conversion. It might looks strange with no depth in the model, hence can lead to not seeing it on the rendering.

I've not tried the PVRGeoPod tool directly, I've only used the Blender extension. Did you use blender to generate your model?

Regards,

Erick