Softwares and audio processing
To convert your audio files into spectrograms you will need to use a sound analysis software. There are now several good free softwares that can do all what you will need to do. On this page, you will find a few examples of interesting ones I would recommend on birding purposes. I’m not writing a complete article on how to use them because they already include very good helps and online tutorials, so the aim is mainly to show you how it looks and share some advice about basic operations.
- RavenLite
- Audacity
- Sound processing softwares like Tseep and Thrush
- Basic operations in sound processing
1.RavenLite
RavenLite was designed by the Cornell Lab of Ornithology and the Bioaccoustic Research Program. It’s a very comprehensive and easy-to-use interactive software that enables you to visualize spectrograms and do some basic sound processing operations. Its strength are that it is very intuitive and designed specially for birders.
You can download it here
This is how I use it, with the waveform and spectrogram views :

2.Audacity
Audacity is an audio editor that enables you do do many sound-processing operations (actually more than RavenLite, even if you will rarely need to do more…). It’s designed for varied users (not just bird sounds) so you might find it less easy to use.
You can download it here
I work mainly on the Spectrogram view from the Audio Track Dropdown Menu so that it looks like that :

3. Sound processing softwares like Tseep or Thrush
Tseep and Thrushes are softwares that extract the calls of migrating bird from a recording. They detect short burst comprised in the frequencies of bird calls and return the times at which they occur in the recording.
This is particularly useful if you leave your microphone in a given place (for night migration for example) and don’t want to listen to the entire recording.
You will find everything about Tseep, Thrush and their fellow here
4.Basic operations in sound processing
The four basic operations you may need to improve the quality of your sound are listed below . You should use them with caution since sound editing is destructive and may distort your sound or severely impact its quality. Remember that a natural recording will always sound better for a birder than a heavily edited recording, even if there is some background noise on it.
4.1 High-pass filter
This is probably the most useful one. Indeed, the frequencies of a bird songs are commonly comprised between 1 and 10 kHZ, while mechanical noise (the hustle and bustle of the city’s activity) is usually lower pitched (between 0.1 and 1kHz). That is to say : they do not compete on a spectrogram and you can separate them.
To do so, you will need to chose under which frequency you want to cut the background sounds (cutoff frequency), and how strong the noise reduction should be (it’s called the rolloff, which means that the sounds distant from the cutoff frequency are more attenuated)
Check out the following recording of a Black woodpecker’s (Dryocopus martius) flight call and a singing Great tit (Parus major), using the Effects > High Pass Filter with Audacity :

To cut the noise from the city we select a cutoff frequency equal to the lower limit of the birds’ sounds (1700Hz). The rolloff depends on how much you want to reduce the low frequencies, so you can try several values.

So that we get something like that :

4.2 Low-pass filter
The Low-pass filter works exactly like the High-pass filter, but cuts the frequencies above a certain threshold. Imagine now that you want to select only the call of the Black woodpecker, and remove the song from the Great tit :
Once again we use Effects > Low Pass Filter in Audacity. We select a cutoff frequency equal to the upper limit of the Black woodpecker : 2000Hz

And here is the result :

4.3 Noise reduction
Okay, first let’s say it : Noise reduction is almost never a good idea. But if you absolutely want to get rid of background noise you can try the Noise reduction tool in Audacity.
Select a part of the recording with only background noise, and go to Effects > Noise reduction :

Press the “Get Noise Profile” button.
Then select the entire recording, do the same thing but play with the cursors in the “Step 2” section. Listen to the preview and correct the cursors until you’re happy with the result.You should keep the noise reduction ans sensitivity as low as possible in order not to ruin the quality of your recording. Then press “OK”
You can do this several times in a row, which will give you somehing like this (here I’ve done a very strong noise reduction – that’s really bad for the sound quality)

4.4 Amplification
One final thing you can try is to amplify the sound, so that it sounds louder. With the Effect > Amplification tool, you will get a louder sound (but also a louder background noise !)
Don’t expect miracles – A good sound recording is a sound recorded for which no editing is required !
Read the next page of “The quick guide to soundbirding” : Where and when to record birds ?
Bonjour,
Est’il possible de m’expliquer comment utiliser les logiciels Thrush-x?
Par avance merci.
Maël
Hi, is it better to amplify before doing other manipulations or amplify as a last step?
Thank you.
Hi Damien,
I think it is better to amplify the purest sound, that is to say, at the very last step, after other manipualtions.
Stan
Ok, thank you.
Hello,
I have another question. What is a good sample rate to record birds in my garden? I am currently setting my recorder at 48 KHZ…
Thank you
48kHz is fine !
The basic rule is that the sample rate must be at least two times higher than the maximal frequency you want to record. So if we admit that we record sounds between 20Hz and 20kHZ, your sample rate must be at least 40kHz. If you go below (like 16kHZ), the highest frequencies will disappear or be strongly altered.
You can go over 48 kHZ (like 96 kHZ), but you won’t hear any difference and will have heavier files.
Stan
Thank you.
Hello! And thank you for this site.
The example spectrogram from RavenLite that you show above is a good illustration of a recording distorted by amplitude clipping. There’s a video that explains what amplitude clipping is and how to recognize it in animal sound recordings here: