Batch processor: Reconstruction settings?

[zoltan]

I noticed that if I reconstruct and auto-label a take file with my code, the results are slightly different when I export it from Motive.

* When all rigid bodies are fine, then everything is identical
* My code tends to drop the rigid body data easier than with Motive.

I am using the default settings, but I get the same when I load the reconstruction settings from the examples using Motive.Settings.ImportMotiveProfile( reconstruction_settings_file );

Can someone give me an updated ReconstructionSettings.motive file that works the same way as Motive 3 does?

These are the values I have:

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<profile version="1">
    <property_warehouse>
        <properties>
            <property>
                <name>BitDepth</name>
                <value>12</value>
            </property>
            <property>
                <name>QuantumActiveFlickerWeight</name>
                <value>0.200000</value>
            </property>
            <property>
                <name>QuantumResidualThreshold</name>
                <value>90.000000</value>
            </property>
            <property>
                <name>QuantumThreshold3D</name>
                <value>30.000000</value>
            </property>
            <property>
                <name>QuantumThreshold2D</name>
                <value>6.000000</value>
            </property>
            <property>
                <name>QuantumMinRaysBoot</name>
                <value>2</value>
            </property>
            <property>
                <name>QuantumMinRaysClean</name>
                <value>1</value>
            </property>
            <property>
                <name>QuantumMaxFillFrames</name>
                <value>2</value>
            </property>
        </properties>
    </property_warehouse>
</profile>

Or, can someone tell me what settings I should give my Motive.Tracjectorizer object?

It would be a great help, because Motive processes my files in 2-3 hours, whereas my code does it in about 4 minutes instead.

[ish67]

Hi
adjust your motive.Tracjectorizer settings by QuantumResidualThreshold to 90.0, QuantumThreshold3D to 30.0, QuantumThreshold2D to 6.0, and QuantumMinRaysClean to 1.
experiment these modifications, it will solve your problem.
thanks

[Lylawa]

Hello,
Modify your Tracjectorizer settings by adjusting QuantumResidualThreshold to 90.0, QuantumThreshold3D to 30.0, QuantumThreshold2D to 6.0, and QuantumMinRaysClean to 1. Give these adjustments a try, as they should effectively resolve the issue you're facing. Your thanks are appreciated.

[zoltan]

Thanks for these. It seems that I don't check this place often enough!

Will try these settings, and report back.

[zoltan]

Okay, so I pretty much implemented these. They look like the 'default' settings in my processing settings file.

Now it runs about 10 times slower, but the output seems more similar to what I am getting when I do the exporting in Motive.

I know this is super proprietary and probably a trade secret, but can someone explain what each of these values do? I really don't feel like tinkering with production-quality data without knowing what I am doing, because I want to minimise the chance of introducing artefacts.


Anyway, thanks for the help!

[bandknobby]

I've successfully incorporated these settings into my processing setup, and they closely resemble the 'default' configuration in my processing settings file.

However, I've noticed a significant decrease in processing speed – approximately 10 times slower. Despite this, the output now aligns more closely with what I observe during the exporting process in Motive.

While I understand that the details behind these values might be proprietary or considered trade secrets, I'm eager to gain a deeper understanding of each parameter. Having a comprehensive grasp of their functionalities is crucial before delving into manipulating production-quality data. I'm cautious about introducing any artifacts and want to minimize any unintended impacts on the accuracy of the results.

If anyone could shed light on the purpose and effects of each specific value, it would greatly enhance my confidence in adjusting these settings effectively. Your assistance is much appreciated!

[betterwound]

I noticed that if I reconstruct and auto-label a take file with my code, the results are slightly different when I export it from Motive.

* When all rigid bodies are fine, then everything is identical
* My code tends to drop the rigid body data easier than with Motive.

I am using the default settings, but I get the same when I load the reconstruction settings from the examples using Motive.Settings.ImportMotiveProfile( reconstruction_settings_file );

Can someone give me an updated ReconstructionSettings.motive file that works the same way as Motive 3 does?

These are the values I have:

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<profile version="1">
    <property_warehouse>
        <properties>
            <property>
                <name>BitDepth</name>
                <value>12</value>
            </property>
            <property>
                <name>QuantumActiveFlickerWeight</name>
                <value>0.200000</value>
            </property>
            <property>
                <name>QuantumResidualThreshold</name>
                <value>90.000000</value>
            </property>
            <property>
                <name>QuantumThreshold3D</name>
                <value>30.000000</value>
            </property>
            <property>
                <name>QuantumThreshold2D</name>
                <value>6.000000</value>
            </property>
            <property>
                <name>QuantumMinRaysBoot</name>
                <value>2</value>
            </property>
            <property>
                <name>QuantumMinRaysClean</name>
                <value>1</value>
            </property>
            <property>
                <name>QuantumMaxFillFrames</name>
                <value>2</value>
            </property>
        </properties>
    </property_warehouse>
</profile>

Or, can someone tell me what settings I should give my Motive.Tracjectorizer object?

It would be a great help, because Motive processes my files in 2-3 hours, whereas my code does it in about 4 minutes instead.

I've seen a huge decrease in processing speed, perhaps ten times slower. I understand this is very private and likely a trade secret, but could someone explain what each of these numbers do?