search

smousavi05 / Denoising-BTwavelet

This repository contains MATLAB scripts and sample seismic data for appying seismid denoising proposed in: "Hybrid Seismic Denoising Using Higher‐Order Statistics and Improved Wavelet Block Thresholding"
https://www.researchgate.net/publication/303849872_Hybrid_Seismic_Denoising_Using_Higher-Order_Statistics_and_Improved_Wavelet_Block_Thresholding
34 stars 22 forks source link

I want to konw something... #7

Open jzy0023 opened 7 months ago

jzy0023 commented 7 months ago

Dear Mousavi,I am a PhD in seismology from Chengdu University of Technology. I am running this program very slowly and the computation time is very long. My SAC file takes about 1 hour. Do you have any suggestions for improving the speed? In addition, thank you very much for the ideas provided by your code!

kindly Jiang zhiyu

smousavi05 commented 7 months ago

Thanks for your email. Unfortunately, this process is computationally expensive for long duration traces and I haven't spent much time optimizing the code. So it is mostly useful for shorter windows. One option for you would be to break your 1 hours data into smaller 1 to 1.5 min windows with some - e.g. 30 % - overlapping. Denoise these smaller pieces - which should be much faster - and then put them back together.

Alternatively, you can usehttps://github.com/smousavi05/FastDenoising Which is a simplified denoising code I wrote for processing of continuous data.

Good luck, Mostafa

On Wed, Dec 6, 2023, 10:24 PM jzy0023 @.***> wrote:

Dear Mousavi,I am a PhD in seismology from Chengdu University of Technology. I am running this program very slowly and the computation time is very long. My SAC file takes about 1 hour. Do you have any suggestions for improving the speed? In addition, thank you very much for the ideas provided by your code!

kindly Jiang zhiyu

— Reply to this email directly, view it on GitHub https://github.com/smousavi05/Denoising-BTwavelet/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADHWAPEUVTA7GYEJQOAT3C3YIFOKBAVCNFSM6AAAAABAKQ23JOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGAZDSOJZG42TQNY . You are receiving this because you are subscribed to this thread.Message ID: @.***>

jzy0023 commented 7 months ago

Thank you for your reply. Your suggestion is very good. In addition, I have some issues with your other program, Denoising Custom master. When processing some data, such as the 40 line if (AICd>0&difpack>=0.02 * na) statement in the segment, when outputting as disp ('Bimodal '), the program can run smoothly. If it is else, it always reports an error and the array exceeds the limit. I have studied your program and it seems that I have not found a solution, If I comment out the if now, everything can run smoothly. I hope to get your help

------------------ 原始邮件 ------------------ 发件人: "smousavi05/Denoising-BTwavelet" @.>; 发送时间: 2023年12月7日(星期四) 下午4:50 @.>; @.**@.>; 主题: Re: [smousavi05/Denoising-BTwavelet] I want to konw something... (Issue #7)

Thanks for your email. Unfortunately, this process is computationally expensive for long duration traces and I haven't spent much time optimizing the code. So it is mostly useful for shorter windows. One option for you would be to break your 1 hours data into smaller 1 to 1.5 min windows with some - e.g. 30 % - overlapping. Denoise these smaller pieces - which should be much faster - and then put them back together.

Alternatively, you can usehttps://github.com/smousavi05/FastDenoising Which is a simplified denoising code I wrote for processing of continuous data.

Good luck, Mostafa

On Wed, Dec 6, 2023, 10:24 PM jzy0023 @.***> wrote:

> Dear Mousavi,I am a PhD in seismology from Chengdu University of > Technology. I am running this program very slowly and the computation time > is very long. My SAC file takes about 1 hour. Do you have any suggestions > for improving the speed? In addition, thank you very much for the ideas > provided by your code! > > kindly > Jiang zhiyu > > — > Reply to this email directly, view it on GitHub > <https://github.com/smousavi05/Denoising-BTwavelet/issues/7&gt;, or > unsubscribe > <https://github.com/notifications/unsubscribe-auth/ADHWAPEUVTA7GYEJQOAT3C3YIFOKBAVCNFSM6AAAAABAKQ23JOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGAZDSOJZG42TQNY&gt; > . > You are receiving this because you are subscribed to this thread.Message > ID: @.***> >

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

smousavi05 commented 7 months ago

I suggest to ignore custom denoising ones. As I've mentioned in an warming on its github repo, that code needs lots of work to be operational as the method is very complex and I wrote the code naively (that was the first program I wrote during my grad school and after that I learned to gradually make my codes simpler)

On Thu, Dec 7, 2023, 1:48 AM jzy0023 @.***> wrote:

Thank you for your reply. Your suggestion is very good. In addition, I have some issues with your other program, Denoising Custom master. When processing some data, such as the 40 line if (AICd>0&difpack>=0.02 * na) statement in the segment, when outputting as disp ('Bimodal '), the program can run smoothly. If it is else, it always reports an error and the array exceeds the limit. I have studied your program and it seems that I have not found a solution, If I comment out the if now, everything can run smoothly. I hope to get your help

------------------ 原始邮件 ------------------ 发件人: "smousavi05/Denoising-BTwavelet" @.>; 发送时间: 2023年12月7日(星期四) 下午4:50 @.>; @.**@.>; 主题: Re: [smousavi05/Denoising-BTwavelet] I want to konw something... (Issue #7)

Thanks for your email. Unfortunately, this process is computationally expensive for long duration traces and I haven't spent much time optimizing the code. So it is mostly useful for shorter windows. One option for you would be to break your 1 hours data into smaller 1 to 1.5 min windows with some - e.g. 30 % - overlapping. Denoise these smaller pieces - which should be much faster - and then put them back together.

Alternatively, you can usehttps://github.com/smousavi05/FastDenoising Which is a simplified denoising code I wrote for processing of continuous data.

Good luck, Mostafa

On Wed, Dec 6, 2023, 10:24 PM jzy0023 @.***> wrote:

> Dear Mousavi,I am a PhD in seismology from Chengdu University of > Technology. I am running this program very slowly and the computation time > is very long. My SAC file takes about 1 hour. Do you have any suggestions > for improving the speed? In addition, thank you very much for the ideas > provided by your code! > > kindly > Jiang zhiyu > > — > Reply to this email directly, view it on GitHub > <https://github.com/smousavi05/Denoising-BTwavelet/issues/7&gt;, or > unsubscribe > < https://github.com/notifications/unsubscribe-auth/ADHWAPEUVTA7GYEJQOAT3C3YIFOKBAVCNFSM6AAAAABAKQ23JOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGAZDSOJZG42TQNY&gt;

> . > You are receiving this because you are subscribed to this thread.Message > ID: @.***> >

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/smousavi05/Denoising-BTwavelet/issues/7#issuecomment-1845020278, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADHWAPANW2KQMMYGUHKK5I3YIGGG3AVCNFSM6AAAAABAKQ23JOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBVGAZDAMRXHA . You are receiving this because you commented.Message ID: @.***>