Hey everyone! I made a quick web tool for cleaning up audio observation files I wanted to share: https://normalizer.pages.dev/
If you’re like me, you’ve recorded lots of phone videos of animal sounds with the plan that you’ll later go onto your computer, convert it to audio, load it into audacity, normalize and maybe add a highpass filter, export and upload. That’s kind of a pain! And if you skip the audacity part, the audio is often rumbly from wind or cars and quiet.
The idea with this tool is you’ll be able to visit it on your phone, upload a video or audio file, select a frequency to filter and get a wav file quickly. I can only test on my Mac OS and Android phone, so if you run into any problems let me know!
One thing to note is that you probably want to include “audio normalized” or something like that in your observation description. I’ve recieved a comment on an obs from someone who was very upset that a birdsong recording I uploaded was normalized to -3db because most audio they were reviewing was not normalized, so their volume was turned way up and the background noise from my recording hurt their ears. Since then, I’ve always noted that the audio was normalized in my own observations.
Good point! I’ll start doing that in the future. I always have the opposite experience, but on the other hand I don’t often mass id audio observations.
I always start out with volume lowered a bit when I play each recording the first time because you never know how loud it will be. I wish people would put time stamps if the call/song is not right near the beginning. Many times I have cranked the volume up trying to ID quiet (background) calls and then suddenly there is a very loud bird and I rip my earbuds out. In the same vein, if the call/song is relatively quiet and there is a loud part of someone talking or some mechanical noise, please make that part quieter, clip it out, or give a warning in the notes (with timestamp).
PS I record audio with my phone using an app called Voice Recorder. I picked this one because it makes files with the name as YYYY_MM_DD_hrs_min_sec.mp3. For example if I did a recording right now it would be called 2025_05_21_16_42_03.mp3. If the recording is decent I will upload from my phone. Usually though I transfer to my computer and edit with Audacity. Spectral delete is my favorite tool on Audacity.
Similar to cropping photos for the subject, I think it’s courteous to try your best to make sure the observed sound is at the beginning of the uploaded clip and make sure it’s the loudest sound. A high pass filter does a lot for wind and car sounds; spectral delete is a lot better but also more fine grained and outside how complex I want the tool to be (at that point you’re better served by Audacity). But cropping the audio is something it should support! My hope is to make it easy for anyone to submit high-quality audio, regardless of their tech literacy (or laziness in my case :-)
What a weird thing to complain about a birdsong… I’d assume any birdsong upload from a non-casual user will be at -3db since it’s what eBird requests the users to use. I’ve upload probably hundreds of audios normalised at that level and never received a complaint about it (However, I make observations in Argentina, where only about 3-5 people check bird audios regularly)
I always use Merlin for any recording (even voice notes) because it does “YYYY-MM-DD hrs_min.wav” so I get both the initial ID if the audio is birds, a nicely labelled timestamp, a quick shortcut to add the recognised bird to an open eBird list, and wav files :p
Is there a reason .wav is preferred over .mp3? I guess what I also like about the other recorder is that it is a black screen with a big red button. I mostly ignored Merlin because it felt like cheating, but looking at it now, it is nice to see the spectrogram as I’m recording. I might have to try recording with it for a while.
I used this on my recent Baltimore oriole recording. I thought my recording was pretty good (from my Panasonic Lumix FZ-80 camera converted to mp3). This made it brighter. I could also hear the other birds singing in the area a lot better.
I know close to nothing about audio file types, but eBird prefers .wav, so I try to do .wav lol
That said, why would Merlin be “cheating”? And cheating at what exactly?
As long as you are not blindly trusting it and therefore documenting bad data, it helps a lot to give you pointers to go and compare when you’re not sure about what you’re hearing.
When I started birdwatching I would probably have never really gotten into identifying birdsongs if I didn’t have those pointers to start somewhere instead of having to compare every single little pip against the entire list of species found in my area (I still had to do it a lot, but at least I was not driving completely blind!)
(I also love to see the spectrogram while I record, especially because I just use my phone and it makes it much easier to realise if I’m catching too much wind or background noise)
I use Merlin as a tool to use to alert me to possible birds in the vicinity, to verify what I am hearing or as a small compact field guide. My hearing is not as good as a lot of birders in my area. So, it is helpful at times. I won’t use what I see on Merlin unless I see or hear the bird myself. When I started birding back in around 2014 I listened to CDs of bird songs/calls and learned from other more experienced birders. And, I still review calls/songs on the CDs or online to get familiar with them in the spring.
@Naelin I meant this as a personal thing, not a comment other people’s use of the app. I want to ID them myself. I do the same with iNat. I try to ID something myself (with memorized knowledge, books, and online illustrations and keys) before turning to using the CV for suggestions. Merlin pops up the names as you’re recording. I can always post photos to iNat later and not see what the CV suggests.
@ken_ohio When I first started birding I didn’t bother with learning to ID by song. Then when I took Ornithology in college (2006) I learned about eBird, the breeding bird survey, and the Christmas bird count and that you can record birds IDed by sound only. I used various online sources to listen to recordings, but found that I learned more easily if I saw the bird in the wild while it was doing the singing/calling. I was given a copy of Stokes Field Guides to Bird Songs CDs the next year. I like to study the recordings to refresh my memory when breeding or winter birds come back, but I also keep the mp3s on my phone, to double check (played on low volume so the birds can’t hear) when I’m not sure I’m remembering correctly while writing down my tally in the field.
My hearing of high pitched songs/calls is better than my audio recording equipment. I have some low pitch hearing loss, but I usually go birding with my mom and she can hear the low pitched birds and frogs for me (and I the high pitched for her). Other people might find it exciting to try to find the bird and train their binoculars or camera on, but I frequently can’t look up for long periods of time without getting dizzy, so more often trying to ID the song is the fun part for me.
It is best NOT to alter ultrasound files for bat calls. Important ID data is often included in the files, which some sound editing programs may interpret as noise.
I like walking through the woods and identifying birds by song/call. Sometimes, in late spring and summer when the leaves are full, you get most of your birds by sound only.
“My hearing of high pitched songs/calls is better than my audio recording equipment. I have some low pitch hearing loss, but I usually go birding with my mom and she can hear the low pitched birds and frogs for me (and I the high pitched for her).” That’s really fantastic. I can’t hear the higher pitches anymore. But, I seem to be able to see movement better than a lot of birders. So, I really like going with a couple of friends who have good hearing. We find birds both by sight and sound.
I have the Stokes field guide and CD, too. I originally started with CDs from the Ohio Department of Natural Resources - which are very good. And, I use Merlin as a guide when I’m out in the field and need a sound reference.
WAV files are “lossless” and thus all sound recorded information is in the file. MP3 files are compressed using “lossy” compression (similar to JPEG for images). Basically, to save space and make the file smaller, the MP3 algorithm removes sound that humans supposedly can’t hear. Lots of controversy as to whether people can tell the difference or not which I don’t think we should get into here (do some internet searches if you like to read lots of angry arguments).
When it comes to something like animal calls, quite a bit of the calls are outside the range of human hearing and I suspect would likely be removed by MP3 compression. Thus they wouldnt show up in a spectrogram, etc.
I used this Audacity Tool on my past iNaturalist audio files to make them better.
You can right-click on them in the observation page to download them. Then, you add the new version and delete the old one in the Edit page. There’s a tab for Sounds.
I was going to do this for my existing eBird audio files. You can download your eBird audio files.
You log in and find the audio file on a checklist. Or, you can go through My eBird and list your species with audio.
There is a link to the Macaulay Library version in the bottom right part of the web page with your audio recording in a checklist. Click on that link.
On the lower right side of the new Macaulay Library web page that opens, there is a link to download it. You can then add the new recording to your checklist and delete the old one.