Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

CQT filter normalization revisited

See original GitHub issue

@ejhumphrey pointed out that the CQT filter bank is not scaled by the length of the filters when norm=None. (When norm!=None, this doesn’t matter because length scaling washes out.)

This should be easy to fix.

Issue Analytics

State:
Created 7 years ago
Comments:10 (6 by maintainers)

Top GitHub Comments

1reaction

ejhumphreycommented, Sep 28, 2016

good point / generally aware of that fact … wasn’t going to b4b it, the meat of the work is adding the scalar and updating corresponding tests.

On Wed, Sep 28, 2016 at 9:50 AM, Brian McFee notifications@github.com wrote:

BTW, I think you’re not going to get bit-level comparison to the dl4mir implementation because librosa’s filters are fractional-length, and dl4mir had rounded-length windows (as far as I can tell from a skim of the code).

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/librosa/librosa/issues/412#issuecomment-250171904, or mute the thread https://github.com/notifications/unsubscribe-auth/AA4iq5q0qxPWX15RlotUerzPdYigJ8rpks5qunCKgaJpZM4KIvfj .

0reactions

bmcfeecommented, Oct 11, 2016

Fixed by merging #417

Top Results From Across the Web

huangleiBuaa/NormalizationSurvey: This repo is for ... - GitHub

Survey on Normalization Techniques. This repo is for our paper survey paper on normalization techniques in training deep neural networks:.

Divisive normalization image quality metric revisited

In this work, the classical divisive normalization metric [1] was revisited to address the criticisms raised against error visibility techniques.

Normalization Techniques in Training DNNs - arXiv

Abstract—Normalization techniques are essential for accelerating the training and improving the generalization of deep neural networks.

CG‐SENSE revisited: Results from the first ISMRM ...

Purpose The aim of this work is to shed light on the issue of reproducibility in MR image reconstruction in the context of...

Representative Batch Normalization With Feature Calibration

3.1. Revisiting Batch Normalization. We first revisit the formulation of BatchNorm. Batch-. Norm is composed of the feature centering, feature scaling,.