Area 51 » Science

192

followers

Follow It!

This proposal is now in the Commitment phase — example questions are locked!

80 Example Questions

active newest votes

up vote 22 down vote

added by Shannon Feb 22 at 19:03

maybe this can be in a especific section about machine learning – Bicudo Mar 3 at 14:56

We may want a 'tag' for Machine Learning – Shannon Mar 3 at 18:52

up vote 14 down vote

added by Kerri Feb 22 at 22:31

suggest spelling out acronyms (NFFT) – Shannon Feb 24 at 3:09

this ties in well with perhaps the creation of a section on 'fourier coefficients' or 'fourier transforms'? Quite a few questions have asked something similar. – Vijay Ramesh Feb 24 at 14:17

We may want a 'tag' for FFT – Shannon Mar 3 at 18:52

up vote 13 down vote

added by MZiegenhorn Feb 23 at 2:18

up vote 13 down vote

added by Thejasvi Feb 23 at 8:59

Good question, but perhaps specify the taxon you are interested in? – EcologyTom Feb 25 at 9:48

up vote 11 down vote

added by macster110 Feb 22 at 8:59

up vote 11 down vote

added by Anne Feb 22 at 18:51

"Best" can imply an opinion-based question, which have typically been frowned upon in SE. See stackoverflow.blog/2010/09/29/… for some historical discussion. It may be better to craft a question that references a specific criterion to optimize, rather than "best". – Bryan Krause Feb 23 at 23:10

Thanks Bryan. How do folks feel about this modification to reduce subjectivity? "Which soundscape metrics describe noise levels associated with shipping?" By referencing noise levels, this is a little more specific and I removed the potentially troublesome word "best". There are multiple ways to measure shipping noise, so I think any subjectivity could promote useful discussion in the community. – Anne Feb 24 at 17:51

up vote 11 down vote

added by chloe Feb 23 at 11:49

up vote 11 down vote

added by etgriffiths Feb 24 at 8:58

	Do you mean how to measure received levels? Or are you asking about methods for source level estimation? – Léa Bouffaut Feb 28 at 18:40
	More source level estimation. I'm thinking of studies that don't have a good idea how far away the animals (unsupervised passive monitoring) or when they only have an estimate (100-150 m away from sensor). Many studies just report how loud a signal is over the ambient, but that doesn't tell us much about how loud the source is. Does that make more sense? – etgriffiths Mar 3 at 7:39

up vote 11 down vote

added by Michael Ladegaard Feb 24 at 10:57

This is an important question, but can really only be answered easily if the question includes information about the recording chain. It would also be good to know if it's water or air, since the reference values are generally different. – Anonymous Mar 9 at 0:03

up vote 11 down vote

added by EcologyTom Feb 25 at 9:58

up vote 11 down vote

added by Carlos Abrahams Mar 1 at 11:45

I think this platform would be a great way to identify and encourage the use of Best Practices/Standard Operating Procetures (if I had more votes!). – Shannon Mar 4 at 14:04

up vote 11 down vote

added by selene Mar 1 at 15:02

up vote 10 down vote

added by Shannon Feb 22 at 3:35

	This would depend on target species call frequencies - so would probably need to include information on the taxa? – Carlos Abrahams Mar 1 at 12:37
	Agree, we should encourage people to say what there species of interest is in all questions. – Anonymous Mar 8 at 23:57

up vote 10 down vote

added by Shannon Feb 22 at 3:41

up vote 10 down vote

added by LizFerg Feb 22 at 3:51

	Possibly define "improved detection" more clearly - better recall, better precision? both? – selene Feb 22 at 18:53
	By asking "How does decimation..." and specifying how this might vary with different decimation settings, then this makes way for a more useful answer. – Shannon Feb 23 at 17:00

up vote 10 down vote

added by macster110 Feb 22 at 9:00

up vote 10 down vote

added by CosmicSpittle Feb 22 at 17:37

	The answer could be a review paper - question would be better if it were "Are there examples of terrestrial and marine species which produce sounds unique enough to identify individuals?" – macster110 Feb 23 at 8:42
	I have worked on wolf howls. They have a unique voice. sciencedirect.com/science/article/pii/… I have published an article on identifying unknown howls nature.com/articles/s41598-021-86718-w Also this paper on gibbon classifies call to an individual besjournals.onlinelibrary.wiley.com/doi/abs/… – Sougata Sadhukhan Mar 10 at 9:22

up vote 10 down vote

added by selene Feb 22 at 18:38

Recommend revising to identify a specific topic within bioacoustics (using acoustics for population assessment, defining sounds associated with a specific species, acoustic behavior, etc). – Shannon Feb 22 at 18:46

I updated to make more specific - sorry I did not preserve the original question! It was something like..."what is a good journal to publish bioacoustic research in?" – selene Mar 1 at 15:01

This is good, but you now have two separate (but related) questions in one. I'll upvote it, but it would be better as one question per question... – Dan S Mar 3 at 15:02

Good point, Dan. I've modified to just be a single question. – selene Mar 4 at 16:03

up vote 10 down vote

added by selene Feb 22 at 18:43

Nice question, but could you specify which type of animal / biome you are thinking of? I suspect this is quite different between marine and terrestrial. – EcologyTom Feb 25 at 9:45

I mean this question to kind of ask about the most basic density or abundance estimating equation, which should be pretty similar across taxa, but envision that similar questions with more specifics to a taxa or study system would also be within the scope of the site. – selene Feb 25 at 14:01

It depends on the species, your vocal behavior, vocal types, social behavior, etc. For example, you will apply different approaches to study howler monkeys and maned wolves. – Bruna Campos Paula Mar 3 at 20:19

The basic answer here may not be helpful to most folk because it would just be a basic density estimation equation or pointing to a method like SECR. On the other hand a long answer with lots of links to further reading would be a good reference for folk... – macster110 Mar 7 at 8:51

up vote 10 down vote

added by Anne Feb 22 at 18:50

Would this be looking for a broad overview type answer or specific for a certain system? That would require more detail on the individuals set up. – selene Feb 22 at 18:56

up vote 10 down vote

added by Anne Feb 22 at 18:50

Unclear what is meant by 'calibration value'. Possibly replace by more specific term, as there may be multiple values that go into producing a calibrated measurement (mic sensitivity, saturation voltage of ADC, etc). – Thejasvi Feb 23 at 7:36

up vote 10 down vote

added by Anne Feb 22 at 18:52

up vote 10 down vote

added by Anne Feb 22 at 18:52

up vote 10 down vote

added by selene Feb 22 at 19:01

I think the idea could be broader, so knowing when sounds (not just human-made) cause harm to animals in general (not just marine animals). But perhaps need to define what you mean by "harm"? – Pralle Mar 1 at 14:48

Agreed I feel like this could be a whole series of questions and maybe this is too broad or doesn't capture that idea well here. Maybe I could make it more specific and say, "At what sound pressure level does a permanent and/or temporary threshold shift occur in pinnipeds?" – selene Mar 1 at 15:00

up vote 10 down vote

added by selene Feb 22 at 19:02

Do you also want to know how to set it up or just it's functionality? – Cory Hom-Weaver Feb 23 at 21:31

I envision asking a more specific troubleshooting type question, but an overview of functionality would maybe be helpful too. I've edited it to be more specific, but still short since it has to be. – selene Feb 23 at 21:49

I realize this is not an option in this format, but for future consideration, we will want to be sure people have done their homework first! Appreciate the troubleshooting Q for scope! – Shannon Feb 24 at 3:14

I think this is a reasonable question. Clearly the user has got the module working and configured. All they need to do is set it to display marks that have already been made by right clicking on the display and selecting from the drop down menu - display of a mark being made and an existing annotation are not quite the same. – Anonymous Mar 8 at 23:56

up vote 10 down vote

added by Samara H Feb 22 at 22:06

How do you define "best"? Highest number of species? Most complete metadata? Are you hoping for auto-detection? – Sarah Vela Feb 23 at 12:37

Maybe this can be a more general question, relating to fishes in general – Bicudo Mar 3 at 14:57

up vote 10 down vote

added by Henry WH Hack v3.0b Feb 22 at 23:04

up vote 10 down vote

added by Thejasvi Feb 23 at 7:27

perhaps specify through 3D space, and specify whether the software you want is to just calculate it, or to also visualize it – chloe Feb 23 at 11:52

up vote 10 down vote

added by Thejasvi Feb 23 at 7:55

up vote 10 down vote

added by chloe Feb 23 at 11:51

Perhaps state the actual kit in use - as this is likely to affect the requirements? – Carlos Abrahams Mar 1 at 12:36

up vote 10 down vote

added by chloe Feb 23 at 14:06

up vote 10 down vote

added by chloe Feb 23 at 14:13

This is quite a poor question and should be answered by posting links to several scientific papers on the subject. – Anonymous Mar 8 at 23:59

Here we are asked to vote/comment on the SCOPE of these example questions (we do not have the character count to prepare these questions according to expected Stack Exchange protocol). Thanks! – Shannon Mar 9 at 1:02

up vote 10 down vote

added by Sarah Vela Feb 23 at 17:28

I actually wrote a paper about this (doi.org/10.1002/ece3.8042)! The short answer is it depends + adjustments can be made. In general MP3 compression with higher bitrates or lossless compression algorithms are better (: – Becky Heath Mar 3 at 15:43

These questions are EXAMPLES only to determine the scope of the Bioacoustics site (what questions are 'on' vs 'off' topic. We will save answers for a later stage in the proposal process! – Shannon Mar 3 at 18:52

@Becky Heath, I will definitely take the paper regardless, as I actually did need to know. Thanks!! – Sarah Vela Mar 8 at 16:03

up vote 10 down vote

added by LizFerg Feb 26 at 16:51

Could perhaps clarify your hoped-for outputs, as this would help define the pros & cons of different software. Are you hoping to separate all species in one pass, build individual species recognizers, and are you looking to develop numbers or just a species list? – Carlos Abrahams Mar 1 at 12:33

up vote 10 down vote

added by Kevin James Feb 26 at 17:41

Great question and we aren't allowed to answer - but there is! Will save that for the next stage. I would suggest providing a reference for what exact package you are looking for or where you first read about it (I'm guessing PAMpal?) – selene Feb 28 at 15:30

up vote 10 down vote

added by Kevin James Feb 26 at 17:48

up vote 10 down vote

added by Kevin James Feb 26 at 17:55

up vote 10 down vote

added by Léa Bouffaut Feb 28 at 19:06

You could perhaps state which software you intend to use, as this could be done with packages in R/Matlab, or using proprietary software such as KalPro, PAMguard? – Carlos Abrahams Mar 1 at 12:29

PAMGuard is not proprietary - it's open source :-). SNR is not a standard measure so it can be calculated in multiple different ways. Would suggest that the question should refer to specific type of vocalisation e.g. whistles. – macster110 Mar 7 at 8:48