Scribd Logo
Upload

Welcome to Scribd!

  • 50 views

    Dither Explained

    Uploaded by

    Jose Silva
    The document discusses dithering sequences for overlapping seismic shots when deblending multiple sources. It notes that using a random dithering sequence can lead to clustering that is not ideal. A better approach is to use a low discrepancy sequence that distributes values more evenly. The key observation is that when aligning shots from one source, the effective dithers of the other sources come from summing their individual dithers. To achieve a uniform random distribution of effective dithers, the document proposes using an "Inverse Irwin-Hall" sequence that compensates for the triangular distribution that results from summing two random sequences. This leads to a low discrepancy distribution that is close to uniform random.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Dither Explained

    Uploaded by

    Jose Silva
    50 views5 pages
    The document discusses dithering sequences for overlapping seismic shots when deblending multiple sources. It notes that using a random dithering sequence can lead to clustering that is not ideal. A better approach is to use a low discrepancy sequence that distributes values more evenly. The key observation is that when aligning shots from one source, the effective dithers of the other sources come from summing their individual dithers. To achieve a uniform random distribution of effective dithers, the document proposes using an "Inverse Irwin-Hall" sequence that compensates for the triangular distribution that results from summing two random sequences. This leads to a low discrepancy distribution that is close to uniform random.

    Original Title

    dither_explained

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses dithering sequences for overlapping seismic shots when deblending multiple sources. It notes that using a random dithering sequence can lead to clustering that is not ideal. A better approach is to use a low discrepancy sequence that distributes values more evenly. The key observation is that when aligning shots from one source, the effective dithers of the other sources come from summing their individual dithers. To achieve a uniform random distribution of effective dithers, the document proposes using an "Inverse Irwin-Hall" sequence that compensates for the triangular distribution that results from summing two random sequences. This leads to a low discrepancy distribution that is close to uniform random.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    50 views5 pages

    Dither Explained

    Uploaded by

    Jose Silva
    The document discusses dithering sequences for overlapping seismic shots when deblending multiple sources. It notes that using a random dithering sequence can lead to clustering that is not ideal. A better approach is to use a low discrepancy sequence that distributes values more evenly. The key observation is that when aligning shots from one source, the effective dithers of the other sources come from summing their individual dithers. To achieve a uniform random distribution of effective dithers, the document proposes using an "Inverse Irwin-Hall" sequence that compensates for the triangular distribution that results from summing two random sequences. This leads to a low discrepancy distribution that is close to uniform random.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    You are on page 1of 5

    Overlapping (blended) shooting with

    dithers:
    How to do better than a “random sequence”….

    Thomas Elboth (telboth@shearwatergeo.com) – April 2021


    Random is perhaps
    Assumption about dithering “too much” random.
    We get natural
    “clustering”
    An ideal dithering sequence for deblending is:
    1. Uniform random
    • Meaning that all values have the same
    likelihood. (the histogram is reasonably flat)
    • The normal ‘rand()’ function on a computer
    will give such a sequence. Insert a picture Low discrepancy
    random seems
    intuitively
    better…
    2. Low discrepancy
    • Meaning that values are well distributed –
    and not clustered.
    • This means that we need to put additional
    constraints on the dithering sequence.
    Dithering in practice: Observations:
    1. Dither all shot-times (from Sources S1, S2 and S3) by [0-x] ms.
    Shot number
    2. Sort to common-offset for deblending
    S1(i)

    Time
    S2(i)
    S3(i)

    • When we align for S1, S2 and S3 will be distributed over [-x to x]ms S1(i)-S1(i)
    S2(i)-S1(i)
    S3(i)-S1(i)

    • When we align for S2, S1 and S3 will be distributed over [-x to x]ms S1(i)-S2(i)
    S2(i)-S2(i)
    S3(i)-S2(i)
    • When we align for S3, S2 and S1 will be distributed over [-x to x]ms S1(i+1)-S3(i)
    S2(i+1)-S3(i)
    S3(i)-S3(i)

    Key observation:
    When aligning the shot times for source N the effective dithers from source N+1 becomes the sum (difference) of the dithers from source
    N and N+1. We want the effective dither (S2-S1) to be uniform random – and low discrepancy.
    Using random dithers

    Adding two uniform random


    sequences gives a triangular
    distribution (Irwin-Hall).
    The result is also high
    discrepancy

    That is clearly NOT ideal!


    By doing some math-magic (inspired by, but not a copy of Patent US10649108):
    Low discrepancy – Inverse Irwin-Hall
    • The key idea is that the effective dithers
    are the sum of the dithers from two
    sources.

    • These effective dithers should ideally be


    low discrepancy – and close to uniform
    random.

    • Adjusted functionality is available at:


    • https://acqplotter-web-app.azurewebsites.net/

    Inverse Irwin-Hall.
    Notice clustering at Low discrepancy Not 100% uniform random,
    the edges… but close!

    You might also like