Sound tips and tricks?

Hi all
Recently, I’ve developed more of an interest in both making and trying to ID sound observations. The problem is that there are a few barriers that make dealing with sound more work than dealing with pictures, so I kind of hate doing it. I have a huge (for me anyway) backlog of sound files I need to turn into observations, and I practically never even bother listening to other people’s sound observations. As a complete noob when it comes to anything related to sound, I’d therefore like to ask a specific question, and also just start a general conversation about sound observations to encourage others (and myself) to both make and ID more sound observations. Do you make sound observations? do you ID them? do you know of any good resources to ID sounds? any tips and tricks you can share? etc…

The specific question- Is there a good way to deal with playback volume in general? I’ve always noticed that sound files in iNat are REALLY quiet, which necessitates me cranking up the volume every time I want to listen to a sound in iNat. Is there some way to make your device (I almost always use a windows 11 laptop) play sounds from different sources at the same level? I am practically always listening to something through my computer- podcasts, videos, music- and having to pause, crank up the volume, listen to something I most likely won’t be able to ID because my brain sucks at sound recognition, and then having to crank down the volume and resume playback is just too annoying for me to bother trying to ID other people’s sound observations.
Similarly, I’ve noticed that the sound level is totally different depending on which device I use. If I open the exact same file in the exact same program (say, adobe audition) but in different devices (laptop 1 and laptop 2), the sound volume is completely different. In laptop 1, I generally increase the sound of my recordings by about 10 decibels in adobe auditon. In laptop 2, I have to increase the sound by at LEAST 20, most often even 30 db to get the same effect. I worry that I I may be making people deaf (or at least very angry, lol) by uploading files that are way too loud when people play them on THEIR devices. Is there any way to standardize/control for this variability?

The more general discussion: to avoid this being too long, I’ll just mention a couple of resources I’ve used (I’m located in the eastern US, so this only applies to that general area):
Sounds of Insects is the best (well, only) resource I’ve found to help identify insect sounds. The problem is that it’s just a bunch of files and it’s very impossible to search. I started to make a spreadsheet with the pulse rates and frequencies of each species along with other notes so that it would be possible to quickly focus your search to only a handful of species, but I think I’ve only gotten to about half way down the page after a number of weeks.
BirdNet is a very useful tool that lets you upload a file and it tries to ID it. It often doesn’t give you any useful results, and is useless for anything other than birds, but it can be really great. Merlin Bird ID seems to be better at making the actual identifications, but you first have to download the Merlin app and then you have to download the sound package.
Herps of NC has some great sound files of frogs and toads, although once again it’s a matter of just blindly going around clicking on things and hoping you stumble upon the species that you heard. As such, it’s much more useful when you already have an idea of what it is that you found.

1 Like

eBird asks that all audio uploaded be normalized so that the maximum volume in a recording is -3 dB. Of course a lot of people don’t but it helps.

But yeah I agree this is frustrating, like when I’m listening to a quiet unboosted audio recording and then compare it with an eBird recordings and forget to adjust the volume and it’s SO LOUD in the headset. And yeah my different headsets all have completely different volume levels. Fortunately my computer automatically recognizes that, but in-browser audio settings don’t.

Yeah I think this system can be helpful if you’re serious about identifying solely from recordings. Let’s say I want to identify bird singing in an area, I’d want to make a list of species that do trills, and separate ones that do simple trills, slow trills, 2-parted trills, etc. Then when you come across a trill song you can search up recordings of each species and compare them.

Eventually you get an idea of what warblers tend to sound like, what finch calls tend to sound like, etc. so you can narrow down what you’re searching for. But it’s definitely more challenging than in the field where you have habitat and height and other clues for context, plus a chance of seeing the bird.

Having a spectrogram can be really helpful since you can often see differences that aren’t apparent by ear.

I sometimes make audio observations, and I do have a bit of a backlog. I just record video with my camera, then use the free version of Zamzar to convert it to audio files. I use Zamzar because it was recommended on another thread (I forget which) but the free version only allows 2 conversions a day and has a size limit, which is occasionally annoying. Other than that I don’t have much information.

We use Audacity to make .WAV files. You can convert video to WAV using this method:

In Audacity, we always use “Ctrl + A” (to select all), Effects, Normalize, and check “normalize peak amplitude” set, as mentioned by others, to -3.0 db.

Even poor video recording of far-away target on a phone can be improved to help with ID using this method (ie

Ah! that’s a great tip. I just tried it on a couple of files, and it seems to be a good maximum.

Yes! I love spectrograms, and that’s one of the main reasons why I want to do more audio observations. As a very visual person, spectrograms are amazing.

1 Like

As mentioned by wildwestnature, you can use Audacity… but if you want to convert a bunch of files, what I often use is AnyVideoConverter. It’s free, works offline, and doesn’t have any limits on the number of files you can convert. You can also choose the sampling rate and a number of other things.
A web tool that is just as easy to use as Zamzar but which doesn’t have conversion limits (I think?) is Online converter. Even if you prefer Zamzar, you can try Online converter when you’ve reached your Zamzar limit :)

1 Like

I do try to identify sound observations when I can. I am good with the frogs but only know a few insects and struggle with birds. But it is tricky when people upload such poor quality recordings with the note “you’ll have to turn your volume way up to hear it.”
If it is a species I believe I can recognize (a North American Anuran or a few other anurans), I will download their audio file and increase the gain in Audactiy. Sometimes I have to do a bit more filtering/cleaning up before I can hear what they were hearing. It is a bit more work that just looking at a photo, but it is the nature of having people with limited experience upload audio.
I’d rather have to put in a bit more effort myself than discourage people from posting audio just because it isn’t ready for identification yet.

Depending where you live, I have a bunch of amphibian calls on my blog (, but again you would have to have an idea of where to start? ( It mostly covers Texas, but has a few calls from various places I’ve traveled and recorded). Maybe let iNat suggest a few local species in your area. I’ve thought about making a “frog call key” on my blog sometime that would help inexperienced people get the call into the right group. One day…


Thanks! I’ll try that out.

I am a sound recordist with much experience of recording on location. It’s really great that so many people are discovering ‘listening’ - honestly, I find sound to be the most evocative of senses for more reasons than I care to mention right here!

Playback volume is tricky - there are lots of variables. It is why online streaming services will turn up/turn down song volumes automatically according to LUFS measure of loudness - it provides consistency for the end user. I generally use the same measure when editing audio. Normalising to -3dB will certainly make sounds louder, but not in a consistent manner.

My main interest is underwater sounds. There is so little information out there about these incredible soundscapes. Over the last 5 years or so I have collected a decent library of sounds. I’m not sure how to categorise them in a way that was useful to other people, so I would be interested to know your thoughts…

My favourite free audio editor is Ocenaudio - any file, handles large files well, and includes spectrogram

You can hear some of my recordings here:

I also started a freshwater listening group on iNat, but not really invested much effort yet - maybe now’s the time!

All the best


To avoid that problem, I listen to the iNat obs on my laptop and pull up the bird songs to compare to on my phone

For insects of Ensifera in North America, has more comprehensive coverage than songsofinsects.

I find the BirdNET app much more useful than the web page. It allows you to select portions of audio to submit. When used like this, I generally find it more accurate than Merlin, but Merlin will identify multiple species and highlight them as they are heard again. Merlin works on its own and needs the appropriate bird pack, but the BirdNET uploads for analysis on the server, so it doesn’t require much space.

Personally, I prefer audio to not be modified other than trimming. Too long is better than too short, some recordings have part of the song trimmed out, some would benefit from being long enough to include multiple call types instead of a single note. For ID with multiple species present, noting timestamps or giving descriptions of the sound of the target species is often helpful.

Attempts to clean audio often remove nuance and sometimes make things sound distorted. If there is noise interfering for me, I can filter in various ways on my own, but I can’t return sound that has already been removed. The majority of observations have unmodified audio and are fairly quiet, so I tend to have the volume jacked up by default. When I’m not prepared, normalized audio just makes it startlingly loud, especially with high frequencies in the background noise or many insect sounds. When going to Macaulay or other references, as most are normalized, I am prepared to adjust the volume of playback. Many audios have loud noises at the beginning or end as the device is handled. If that can be clipped out, that can be nice. I’ve read there are things that automatically adjust volume on a computer, but I don’t know anything about them.

It helps to be mindful of the noises made by your own movement. Recording while trying to get closer often generates a lot of masking noise. Remaining motionless further away is often much better. High audio compression can make things very difficult. Sometimes this is a setting in the software, sometimes it happens in format conversion, or some editors. Wav and high bitrates are great, especially with faint sounds, but 128 or even 96kbits/s is usually sufficient. Lower bitrates intended for memo taking are often too distorted for me. Many sound like they were recorded over a bad cell connection and can be hard to recognize as even an animal.


In the case of grasshopper and cricket songs, for a good identification you need to look at the spectrosonogram and the oscillogram. With the former you can see the main carrier frequency. With the second one you can count the number of syllables per minute, the number of echemes per minute, and so on. Fortunately they are more accurate than birdsongs.

1 Like

I’ll have to read up more on LUFS, that sounds like something I want my sound driver or whatnot to do. I’ve tried changing my sound settings for “loudness equalization” and whatnot, but nothing really works.

As a big fan of ASMR, I love your underwater asmr album! it’s very relaxing.

1 Like

Like @djcdelahaye, we have used Ocenaudio and found it very effective. We have since switched to Audacity but found them comparable for sound editing and conversion to WAV.

1 Like

I love birding, so I remember the sound of bird and the behaviour they display when making sound, like some birds sing continously, and some birds sing same notes in repetition, and in audio I can also see number of birds, so it is also helpful to id a birds, and some bird changes place while singing, so there is some change in frequency of sound(or whatever they say in physics) so that might help,
my point of saying if you know the bird you id (like 100%) you can id them well, because then you don’t rely (not sure about spelling, just it means depend) only on sound.
and yes I also put sound observation, when I am recording them, I check the wind speed, and I never face the recorder towards open, because a little breeze can blow my sound observation, I record them in early morning like 1 hour before sunrise to sunrise after that I don’t record, I also make some visual notes like what was bird like, If I can see, if it is hiding in the bushes, I try to guess the size
my sizes are
sparrow size (12-15cm) or smaller
feral pigeon size (32- 35 cm) or little smaller
crow size ( 40 cm) or little more or less
hawk size (45 to 80 cm)
a very big bird( beyond that) which is rare

other thing you can tell when it is making sound, like red wattled lapwing make sound while flying.

Some birds make sound while sitting in on place

some birds make sound like while flying but in groups
they are communicating in groups

some birds make sound in groups but on land, these birds can fly but jump and climb on bushes instead, they are doing communication

some birds come flying and then make sound when they sit

some birds have more than one call like

Pigeons and doves have soft cooing like sound
little twee sound is the sound of bird green bee eater, which travel in groups and have distinct flying style, the fly like flap there wings and ready to fall, flap there wings and ready to fall,
this bird make sound while flying
this bird make long calls with same sound
this bird makes call when it perches on something and also while flying , these birds fly short distances
and for audios like this, you should be patient because the call of the bird is from very far away like half of kilometre I can hear this bird call but when I try to record voice becomes low, so keep calm when recording or idying these obs.
My point of saying is that make notes of birds around you, about everything, if you can’t describe the bird by and description, describe them by gut feeling. I am sure you can describe, any problems?, I am here to help
sorry for sticking to birds, but there is nothing else hear in cities except dogs and squirrels.
regards and happy birding, I hope it helps

Thanks for checking the ASMR album out :grinning: There are some truly spectacular soundscapes out there - trying in tune in!

Xeno-canto is a great resource for bird songs.

I process bird recording in Audacity which is free. My process is:

  • cut out short patches of excess noise which prevent effective normalizing (such as dropping the lens cap or treading on branch)
  • normalize
  • select part of recording which does not have the sound I am focused on and use that for setting the noise reduction filter
  • do noise reduction
  • If there are patches of wind noise I find using “bass and treble” to set bass to -30 (the minimum) does a very good job of removing it without doing much damage to most bird song (which tends to be in the treble range).

Cheers Andrew from NZ


Thanks for the bass and treble tip! I’ll have to give that a try on some of my recordings that I’ve deemed unusable due to wind noise. The best way I’ve found so far to cut down on wind and camera noises is to use an external microphone with a wind muff but that of course does not help with recordings already done without that equipment.