Error in Network, Skiping Audio Chunk about deepsqueak

[DrCoffey]

Hey Bavo,

The update seems to be working on all of our audio files, can you send me an offending file?

Thanks,
Kevin

[BavoKempen]

Hey Kevin,

The file that I shared with you some time ago on dropbox should recreate this error :)

Kind regards,

Bavo Kempen

[DrCoffey]

I just cloned the Master Branch and detected AN1.wav and AN12.wav and they both worked on a couple computers with all of the networks. Can you open SqueakDetect.m and under the "Run" drop down menu hit "Pause On Errors". Then can you re-create the error, save and send us the workspace?

[DrCoffey]

Also. when you pause on error, you can change the workspace in a drop down menu. Make sure it is on the SqueakDetect workspace.

[BavoKempen]

Dear Kevin and Russell,

I sent the workspace .mat file through email, thanks for taking a look at it!

Kind regards,

Bavo Kempen

[DrCoffey]

Hey Bavo,

I should have been a little more clear with my instructions. The workspace you sent is from the wrong function. You need the workspace from SqueakDetect. There are 2 ways you can do this.

1. Pause on errors, and then change the "Function Call Stack" drop down to SqueakDetect

2. Put a debug dot before the offending line. I have attached a picture for reference.

I also detected the file in the drive link fine over here, and uploaded the .mat. We have an experimental branch we are working on called "detect as tables" that might just fix your problem.

-Kevin

[BavoKempen]

Hey Kevin,

While implementing the above mentioned workspace extractions, peculiar enough everything ran smoothly again, thus no error anymore to report. Sorry for the inconvenience, very weird though since I did not change a thing. Anyways, thanks as always for the rapid and good help!

Kind regards,

Bavo Kempen

[BavoKempen]

Hey again,

Another related problem I think is that now it takes really really long to merge boxes during detection (up to 10 minutes) for audiofiles that are approximately 5 minutes long, whereas this was not the case prior to the update.
I believe and correct me if I am wrong but my specs should be able to handle that or atleast they could before:

1. Processor: Intel(R) Core(TM) i7-6500U CPU
2. 8,0 GB RAM DDR3
3. GPU: NVIDIA GeForce 940M
4. SSD harddrive

Any idea why this might be the case? Is it a combination of the length of the respective files concurrent with the change in merge_boxes.m function that substracts from mean?

Kind regards,

Bavo Kempen

[DrCoffey]

Hey Bavo, We have solved that problem and will upload a major update today. We are just trying to test everything before the push.

[BavoKempen]

Hey Kevin,

That is great news, thanks for the fast updates!

Kind regards,

Bavo Kempen

[BavoKempen]

Dear Kevin and Russell,

Thanks a lot for the great update, detection runs a lot smoother again!

Kind regards,

Bavo