Background gain

[cmlenell1]

Hi Dr. Coffey,

I've been working with deepsqueak for several studies now. I really like the program. One problem that I constantly run into is that the background "gain" is too dark. I end up having to draw a box around nothing so I can better see the signal. Is there a way to set the gain so that it doesn't auto to the detected signal? I attached pictures to illustrate the issue.

[DrCoffey]

Hey, we have a new test build that has this, and many more features built in. I will be updating the main branch sometime soon, but I would recommend checking it out now, it is pretty stable.

https://github.com/DrCoffey/DeepSqueak/tree/Screener

[cmlenell1]

Great! Thank you!

Charlie Charles Lenell, PhD, CCC-SLP he/they

On Mon, Mar 22, 2021 at 3:24 PM DrCoffey @.***> wrote:

Hey, we have a new test build that has this, and many more features built in. I will be updating the main branch sometime soon, but I would recommend checking it out now, it is pretty stable.

https://github.com/DrCoffey/DeepSqueak/tree/Screener

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DrCoffey/DeepSqueak/issues/104#issuecomment-804369203, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK23P6MUKMQFZKOYX3YFRXTTE6RPVANCNFSM4ZTHHWKQ .

[cmlenell1]

Hi Dr. Coffey,

I used the new version of deepsqueak and I love the changes. Being able to add USVs and change the detection box of calls that were cut off were much needed functions.

I was wondering if you wanted to add "peak frequency" to the measurements and excel export like I suggested on a previous thread. The Cuicci Lab and other researchers like Wohr use peak frequency a lot. I managed to add the code and I think it would help make this amazing program more widely adopted and utilized. Let me know if you want my code again.

Thanks again!

Charlie

Charles Lenell, PhD, CCC-SLP he/they

On Mon, Mar 22, 2021 at 3:52 PM Charles Lenell @.***> wrote:

Great! Thank you!

Charlie Charles Lenell, PhD, CCC-SLP he/they

On Mon, Mar 22, 2021 at 3:24 PM DrCoffey @.***> wrote:

Hey, we have a new test build that has this, and many more features built in. I will be updating the main branch sometime soon, but I would recommend checking it out now, it is pretty stable.

https://github.com/DrCoffey/DeepSqueak/tree/Screener

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DrCoffey/DeepSqueak/issues/104#issuecomment-804369203, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK23P6MUKMQFZKOYX3YFRXTTE6RPVANCNFSM4ZTHHWKQ .

[DrCoffey]

Hey Charlie, I completely missed this. Ya, send me the code and I will add peak frequency to the default outputs!

[cmlenell1]

Okay. Let me know if you have any questions. Pasted the code below.

Add these three lines to the "CalculateStats.m" underneath % Power of the call contour stats.Power = ridgePower;

% Peak frequency of the call contour stats.RidgeFreq = FreqScale*stats.ridgeFreq_smooth + Box(2); stats.PeakFreq = stats.RidgeFreq(stats.Power==max(ridgePower));

Then add peak frequency to "excel_Callback.m" by changing the following lines

exceltable = [{'ID'} {'Label'} {'Accepted'} {'Score'} {'Begin Time (s)'} {'End Time (s)'} {'Call Length (s)'} {'Principal Frequency (kHz)'} {'Low Freq (kHz)'} {'High Freq (kHz)'} {'Delta Freq (kHz)'} {'Frequency Standard Deviation (kHz)'} {'Slope (kHz/s)'} {'Sinuosity'} {'Mean Power (dB/Hz)'} {'Tonality'} {'Peak Freq'}];

exceltable = [exceltable; {ID} {Label} {accepted} {Score} {stats.BeginTime} {stats.EndTime} {stats.DeltaTime} {stats.PrincipalFreq} {stats.LowFreq} {stats.HighFreq} {stats.DeltaFreq} {stats.stdev} {stats.Slope} {stats.Sinuosity} {stats.MaxPower} {stats.SignalToNoise} {stats.PeakFreq}];

Best,

Charlie Charles Lenell, PhD, CCC-SLP he/they

On Wed, May 5, 2021 at 2:34 PM DrCoffey @.***> wrote:

Hey Charlie, I completely missed this. Ya, send me the code and I will add peak frequency to the default outputs!

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DrCoffey/DeepSqueak/issues/104#issuecomment-832953927, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK23P6LRUJS5KTIU6AZ2QFDTMGMSTANCNFSM4ZTHHWKQ .

[DrCoffey]

Thanks for this!

[cmlenell1]

Hi Dr. Coffey,

I've been using this version and was wondering if you were planning on including a delete button to delete detected noise elements and was also wondering why the contour of this version of deepsqueak had such much fewer points than the previous version.

Thanks!

Charlie Charles Lenell, PhD, CCC-SLP he/they

On Mon, Mar 22, 2021 at 3:24 PM DrCoffey @.***> wrote:

Hey, we have a new test build that has this, and many more features built in. I will be updating the main branch sometime soon, but I would recommend checking it out now, it is pretty stable.

https://github.com/DrCoffey/DeepSqueak/tree/Screener

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DrCoffey/DeepSqueak/issues/104#issuecomment-804369203, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK23P6MUKMQFZKOYX3YFRXTTE6RPVANCNFSM4ZTHHWKQ .

[DrCoffey]

Hey Charlie,

You can remove any call by right clicking on it. You can also remove all rejected calls with Tools->Automatic Review->Remove Rejected Calls.

The contour is based on the tonality slider, but I don't think we changed the code or point calculation in the new version. Can you give a specific example?

-Kevin

[DrCoffey]

Hey Charlie,

There doe seem to be something up with the contour extraction. I will investigate and get back to you.

-Kevin

[cmlenell1]

Yes, Sorry for the delay. When you change the display overlap percentage (in display range), the contour and value changes as well.

Best,

Charlie

Charles Lenell, PhD, CCC-SLP he/they

On Mon, Jul 19, 2021 at 2:23 PM DrCoffey @.***> wrote:

Hey Charlie,

There doe seem to be something up with the contour extraction. I will investigate and get back to you.

-Kevin

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DrCoffey/DeepSqueak/issues/104#issuecomment-882798760, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK23P6M72Q4LDBP4UQGMCWLTYR3UTANCNFSM4ZTHHWKQ .

[DrCoffey]

Ya, the contour points are based on the number of pixels in the display. I am working on the edge detection this week, so there may be a couple code updates that improve contour detection.