Bindings Parameters

Note

picker.* Parameters controlling the second-stage picker for re-picking P phases after an initial detection.

Note

picker.deepc.* Phase repicking and classification based on deep learning by the deepc plugin. Add deepc to the global parameter “plugins” for activating this plugin.

picker.deepc.modelsPath

Default: /home/data/dlmodels

Type: directory

Path to the models collection.

picker.deepc.resampleTimeout

Default: 1

Unit: s

Type: int

The timeout for waveform resampling. When exceeded pick processing is stopped.

picker.deepc.addComment

Default: false

Type: boolean

If enabled adds a comment to the pick with information about the classification result and the confidence.

picker.deepc.sendNoise

Default: false

Type: boolean

Send picks with phase hint classified as ‘Noise’ and evaluation status is set to ‘rejected’. Use this option for testing.

picker.deepc.sendRejected

Default: false

Type: boolean

Send false picks with evaluation status ‘rejected’. Use this option for testing.

Note

picker.deepc.pickFalsification.* Picks result from sudden increase in amplitudes. When this happens due to leading zeros, the resulting pick is certainly false and will not be sent.

picker.deepc.pickFalsification.signalBegin

Default: 0

Unit: s

Type: double

Start of the time window relative to pick time for identifying zeros.

picker.deepc.pickFalsification.zeros

Default: -1

Unit: s

Type: double

The summed length of time windows within signalBegin and pick time containing only zeros for rejecting picks. Negative values deactivate the test.

Note

picker.deepc.deepcEC.* Identification of the event type causing a signal. The type is added to the pick by a comment which can be evaluated by the evType plugin for scevent in order to set the eventType attribute of the event.

picker.deepc.deepcEC.picker

Type: string

Values: AIC,BK

The picker interface for re-defining the pick time before classifications. Empty: The initial time is used. The parameter can be used as in scautopick.

picker.deepc.deepcEC.filter

Type: string

Optional filter applied to waveforms before feeding them into the model.

picker.deepc.deepcEC.modelWeights

Type: string

The model/weights name. Equals the filename stem. For instance, a PhaseNet with Instance weights would be ‘phasenet-instance’, or even shorter ‘pn-in’. The model’s location is expected to be found in ${modelPath}/dlmodels-pick/ or ${modelPath}/dlmodels-class/, respectively.

picker.deepc.deepcEC.threshold

Default: -1

Type: double

Range: -inf:1

The confidence threshold event classifications must exceed to be accepted. Otherwise picks are classified as "not existing".

Negative values: The threshold parameter preferred by the model is used instead.

Note

picker.deepc.deepcPC.* Identification of phase of the pick added to the pick as a comment.

picker.deepc.deepcPC.picker

Type: string

Values: AIC,BK

The picker interface for re-defining the pick time before classifications. Empty: The initial time is used. The parameter can be used as in scautopick.

picker.deepc.deepcPC.filter

Type: string

Optional filter applied to waveforms before feeding them into the model.

picker.deepc.deepcPC.modelWeights

Type: string

The model/weights name. Equals the filename stem.

picker.deepc.deepcPC.threshold

Default: -1

Type: double

Range: -inf:1

The confidence threshold phase classifications must exceed to be accepted. Otherwise picks receive the comment "undefined".

Negative values: The threshold parameter preferred by the model is used instead.

Note

picker.deepc.deepcPR.* Refinement of time and phase name of picks. For full renaming phaseHints from P to S configure “reclassify”. Otherwise just a comment is added to the pick.

picker.deepc.deepcPR.filter

Type: string

Optional filter applied to waveforms before feeding them into the model.

picker.deepc.deepcPR.modelWeights

Type: string

The model/weights name. Equals the filename stem.

picker.deepc.deepcPR.threshold

Default: -1

Type: list:string

The confidence threshold pick classifications must exceed to be accepted. Otherwise picks receive the comment "Noise".

Negative values: The threshold parameter preferred by the model is used instead.

A phaseHint will only change from P to S if "reclassify" is active and a new S pick is closer to the trigger then a new P pick and the confidence of S is higher than of P.

For S classification set individual thresholds as [phase]:[threshold].

Examples:

"0.3" / "P:0.3,S:0.3" : Threshold applies to P and S

"0.3,S:0.6" / "P:0.3,S:0.6" : For P and S thresholds of 0.3 and 0.6, respectively, apply.

picker.deepc.deepcPR.reclassify

Default: false

Type: boolean

Rename P phases to S based on confidence and proximity by changing the phaseHint.

picker.deepc.deepcPR.publishConfidenceAsSnr

Default: false

Type: boolean

Publish the confidence of resulting picks as ‘SNR’. This is particularly useful for scolv, and a workaround for the missing possibility to publish the confidence within the GUI. For any other use, this parameter should be set to false.

picker.deepc.deepcPR.P.signalBegin

Unit: s

Type: double

Start of the time window which is considered legitimate for P picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window start is used.

picker.deepc.deepcPR.P.signalEnd

Unit: s

Type: double

End of the time window which is considered legitimate for P picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window end is used.

picker.deepc.deepcPR.S.signalBegin

Unit: s

Type: double

Start of the time window which is considered legitimate for S picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window start is used.

picker.deepc.deepcPR.S.signalEnd

Unit: s

Type: double

End of the time window which is considered legitimate for S picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window end is used.

Note

spicker.deepc.* Phase repicking and classification based on deep learning by the deepc plugin. Add deepc to the global parameter “plugins” for activating this plugin.

spicker.deepc.modelsPath

Default: /home/data/dlmodels

Type: directory

Path to the models collection.

spicker.deepc.resampleTimeout

Default: 1

Unit: s

Type: int

The timeout for waveform resampling. When exceeded pick processing is stopped.

spicker.deepc.addComment

Default: false

Type: boolean

If enabled adds a comment to the pick with information about the classification result and the confidence.

spicker.deepc.sendNoise

Default: false

Type: boolean

Send picks with phase hint classified as ‘Noise’ and evaluation status is set to ‘rejected’. Use this option for testing.

spicker.deepc.sendRejected

Default: false

Type: boolean

Send false picks with evaluation status ‘rejected’. Use this option for testing.

Note

spicker.deepc.pickFalsification.* Picks result from sudden increase in amplitudes. When this happens due to leading zeros, the resulting pick is certainly false and will not be sent.

spicker.deepc.pickFalsification.signalBegin

Default: 0

Unit: s

Type: double

Start of the time window relative to pick time for identifying zeros.

spicker.deepc.pickFalsification.zeros

Default: -1

Unit: s

Type: double

The summed length of time windows within signalBegin and pick time containing only zeros for rejecting picks. Negative values deactivate the test.

Note

spicker.deepc.deepcSR.* DeepC Phase Classifier and Pick Refiner for S Phases. Can be loaded as secondary picker

spicker.deepc.deepcSR.filter

Type: string

Optional filter applied to waveforms before feeding them into the model.

spicker.deepc.deepcSR.modelWeights

Type: string

The model/weights name. Equals the filename stem.

spicker.deepc.deepcSR.threshold

Default: -1

Type: list:string

The confidence threshold pick classifications must exceed to be accepted. Otherwise, picks receive the comment "Noise".

Negative values: The threshold parameter preferred by the model is used instead.

A phaseHint will only change from S to P if "reclassify" is active and a new P pick is closer to the trigger then a new S pick and the confidence of P is higher than of S.

For S classification set individual thresholds as [phase]:[threshold].

Examples:

"0.3" / "P:0.3,S:0.3" : Threshold applies to P and S

"0.3,S:0.6" / "P:0.3,S:0.6" : For P and S thresholds of 0.3 and 0.6, respectively, apply.

spicker.deepc.deepcSR.reclassify

Default: false

Type: boolean

Rename S phases to P based on confidence and proximity by changing the phaseHint. Usually, this parameter is inactive.

spicker.deepc.deepcSR.S.signalBegin

Unit: s

Type: double

Start of the time window which is considered legitimate for S picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window start is used.

spicker.deepc.deepcSR.S.signalEnd

Unit: s

Type: double

End of the time window which is considered legitimate for S picks, relative to the triggering pick. If the resulting window is larger than supported by the chosen model, it will be corrected to the maximum possible length. The parameter is only evaluated when reclassify is enabled. By default, it is inactive and the model’s recommended window end is used.