Sound quality for iNat observations

Seconding this comment. Even though its just a workaround for the bug you’re reporting, you should seriously consider converting your WAV file(s) to MP3. MP3 files seem to be working fine on the iNat website (including with Firefox). Consider also that the higher fidelity of WAV files comes at a cost of 1000% more bandwidth and storage.

1 Like

Off topic I know, but seriously this is bad advice. Always upload the best quality audio and photos the site will stand, why degrade the quality of the evidence, storage and bandwidth are becoming cheaper by the second. When people come to use the observations in years to come they will not thank you for scrimping on quality.

1 Like

I respectfully disagree. iNaturalist is not about building or showing off a portfolio of studio-quality media. iNaturalist is about recording/reporting occurrences of flora and fauna and corroborating them with photo or audio evidence. That evidence only needs to be good enough for another user, now or in the future, to confirm your record. Sure, higher quality is nice, but its entirely superfluous.

Here are all my audio-based records: https://is.gd/3zysUp. They’re ported from Soundcloud (which, btw, has a bandwidth limit on free accounts) and are all stereo MP3 files @ 192kbps replete with a fade-in and fade-out. Please let me know which ones have been degraded to the point of being unusable for corroborating an occurrence record.

3 Likes

Nobody suggested it was :-)

Yes, the distribution of data points created by iNat observations is one use for the iNat observations. Despite the extra-ordinary collection biases created by the very non-randomised spacial and subject selection, the occurence data has some value for research.

But I don’t see that as being stated as the main goal or even primary goal of iNat, although it may be the impetus behind many contributors efforts. We’re creating a vast (unprecedented?) database of observations, worldwide, that will no doubt find many uses in the future (especially if we give an open license to all our observations and media). Why supply data that is just sufficient for one purpose when it is probably less work to supply the full fidelity data?

Yes, you curate your recordings well :-)

Done well mp3s are fine, but no actual advantage on this site unless a WAV file is simply too large to upload. And once iNat natively accepted sound files, I stopped using soundcloud (I too have no need to create a portfolio and don’t want to restrict people’s ability to download files).

1 Like

This tangent was straying from the bug report so I moved it to its own topic.

1 Like

Good decision.

The issue in regard to sound quality that I’ve observed is not the file format but that observer/recordists/submitters do not edit their sound clips in software–by that I mean “normalize” or amplify the sounds and trim out long intros. I can barely hear most of the submitted audio and need to gain up my speakers. Seems that observers are uploading straight from their phone apps.

I looked around iNat and can’t find any guidelines for submitting quality, audible audio. Perhaps I missed something here. eBird has an excellent tutorial for this using the free Audacity software.

https://help.ebird.org/customer/en/portal/articles/2159648-best-practices-for-audio-upload-to-ebird

N. L.

Just upload what works for you and don’t police other people’s contributions. If you can’t tell what it is mark reviewed and move on. A lot of us have super slow internet connections and limited storage space on our devices. Bandwidth isn’t really increasing for some of us. That stuff follows money.

2 Likes

This post (deleted here) was in error. No replies please.

that was in response to Tony’s comment.

I did scribble a comment about sound here - https://forum.inaturalist.org/t/good-voice-sound-recorders-for-the-field/1123/14 - but it’s not really in the forefront of the site ;-)

Perhaps you could add a tutorial to the forum?

I have actually tried to find ways to convert various file types to others (after reading something about file quality relating to recording of bird calls), and gave up after an hour or so of searching for a free software I could download either on my phone or computer that wouldn’t involve a considerable learning curve. I was specifically looking for programs what would produce sonograms (like what eBird does), but couldn’t find anything. My auditory processing is… uh… bad, so I need the visual form of a call or song.

I will continue to upload whatever quality audio files I can get, in whatever file format they come in. Until editing audio files become more accessible, the standard iNaturalist user is unlikely to do it at all.

There are some creative workarounds–I wasn’t able to get any IDs off my sound recording of a Gray Hawk on iNat, but after uploading the same exact, unedited file to eBird to get the visual sonogram, it resulted in a clearer audio file as well, and we were able to identify the bird: https://www.inaturalist.org/observations/19992924

I have used the Cornell Lab’s Raven Lite for this, e.g. here. It’s free, and the learning curve is minimal if you have ever used any waveform-based sound software before. I can’t say how useful these spectrograms are for others, but they do help me visualize the sound and then hear details that I didn’t hear before.

2 Likes

That’s PERFECT, thank you! Just downloaded it!

If you are interested in recording WAV, a free, basic app worth checking out is Røde Reporter. I’ve been using it for birds and frogs without an external microphone with good results.

WAV is a better choice for producing spectrograms. As I understand it, MP3 uses compression which can compress harmonics that are inaudible to humans but can be important for other species.

1 Like

Thanks! I’ve been using SimpleMic but it doesn’t give me any control over my recordings. Cullen Hanks recommended it back when I attended a BioBlitz he ran in 2015, and I’ve stuck with it for lack of any obvious better options.

1 Like

One compromise option would be lossless compression like FLAC, but I’m not sure how well that’s supported by this site or others. If you’re trying to analyze a sound file for sonogram or spectrogram purposes (e.g., bats), you absolutely would want to use the highest quality audio you can get for that analysis. I imagine though that most audio being uploaded here is just straightforward calls (bird, herp, etc…) for which an MP3 would be perfectly adequate. Personally, I just use my point n shoot camera to take a video and use VLC to extract the audio. Works well enough. If I wanted to invest in a dedicated audio recording device for use in the field, I’d probably get a ZOOM (https://www.zoom-na.com/products/field-video-recording/field-recording/).

For editing and file conversion, it looks like there’s plenty of freebies out there to do the basic stuff: https://www.makeuseof.com/tag/free-online-audio-editors/. Also http://www.wave-editor.com/.

No one is talking about policing other peoples contributions here (except you :-). I was just disagreeing with the advice to compress sound files.

I probably didn’t make it clear, I was talking about iNats bandwidth and storage. The point was don’t degrade your submission just to conserve iNats bandwidth and storage (unless of course they request it!), in the long run it is not a problem (as capacity increases with technology).

1 Like

Oooh sorry. I agree with you about inat bandwidth.

Yes that would be good, but last time I checked it wasn’t supported here.

One thing that occured to me the other day when listening to a Song Thrush (an introduced species here), is how it had picked up elements of other bird songs (I thought I could detect Eastern Rosella and Kaka in there). And wondered about analysing recordings of this common bird to detect the presence of other species, and changes over time. Totally pie-in-the-sky musings, but something that automated signal analysis processing could do - of course I don’t know why there would be recordings of Thrushes and not the other species directly ;-)

I use Audacity, but appreciate the other great sounding suggestions from others, I will have to check them out.