I use ocensudio (also free), and process is fairly simple. Load full wav file, identify the location of the audio you want to keep and ditch rest. If there is “noise” and you can remove it without changing the fundamental audio (for example, if you had recorded a bird which made rhythmic calls, to remove noise in between calls would change the periodicity of the call, so you should leave it in) then do so. Now find the normalize (NOT Effects –> Normalize, there is another section under Effects (which I cannot recall) which also has the Normalize so select this. The recommendation is set to -3db but you can play up to -5db to get best output.
Then use the option to add blank section to start and end - recommendation is 3 seconds for each.
Now save and upload your targeted recording.
Hope this helps. I am sure there are technicians on here which will have a better method, but this works for me.
I’m surprised I haven’t seen controversy yet – if we were discussing doing things like this to images, some participants would object that it is too much alteration.
I believe it is on par with cropping an image and perhaps raising the exposure
The quoted examples are more like image denoising and sharpening. Both of which can lead to artifacts when overdone. And people do complain about over processed images.
The few audios i took i exclusively took with my cellphone.
I cut them to length with the built-in audio app, check sound, upload, check again.
If i had any doubt on quality, i just delete.
Most of my audios were positively identified even in cases where i was not knowledgeable enough to ID the bird on my own.
Taking sufficiently good audios with a cellphone is possible.
And quality is a habit of people more than a property of a device.
For the habit part, i understand the suffering.
If audio recordings are your focus area, just stop reviewing any of them and develop a positive list of observers who deliver quality recordings worth listening ?
This is a big part of why I rarely record sound. Sometimes it seems like somethings going to be nice and loud but then I record the observation and well, it’s not. You can barely hear it
I meant when I’m identifying other people’s audio. So I can see if there’s so much wind or overlapping sounds that I don’t want to waste my time. And also prepare myself for possible much louder sounds. I hate when I crank the volume up to hear a quiet bird and then suddenly there is someone talking right next to the microphone or a much louder bird.
For my own recordings while out in the field, I will replay my file (with ear bud so I don’t disturb the bird) to see if it turned out okay. Though I’ve kind of gotten a feel now for how loud the bird has to be to get a good recording using my phone and don’t usually check the files while in the field anymore.
That’s because our minds unconsciously filter between what we are focusing on and what we are not. That’s how people can follow a conversation in a crowded room.
Very true.
Following the discussion around pictures showing no hint of the claimed taxon and the reluctance to invest in systemic quality measures against that, the only answer today can be:
Either join the suffering, or simply be more selective and do not.
Welcome to the club!
I don’t think you can take a philosophical objection to normalisation. If you’d got the recording level right it wouldn’t need it. 32-bit recorders probably have to do it by definition. You can take issue with high-pass filtering I suppose, although most of what’s on https://xeno-canto.org/ has been hit hard by high-pass filtering. That’s within the groove according to Cornell’s Macaulay submission guidelines. I will take note of their pushback against filtering above 250Hz for birds. What they would declare totally out of order is using compressed formats like mp3, which seems fine on here. I don’t struggle to ID birds I know if the mp3 is clear enough, but perhaps AI scanning wouldn’t like the artifacts and it does change the sonogram sometimes. So the controversy is about different things
There is a spectrogram extension that you can add to iNat, kindly coded by @japh - see here - https://forum.inaturalist.org/t/spectrogram-browser-extension-for-inaturalist-feedback-welcome/67415
It’s brilliant!
Reading this thread out of general interest. I totally sympathize with @mustela-erminea because I have very sensitive ears and it would be too painful to listen to bad recordings. On the other hand, as a non audio educated person, none of the terms or suggested edits make any sense to me given that I don’t know what anything means. Obviously I can take the time to learn, but what you are calling easy edits are currently meaningless. I can clip the beginning and end of audio, but that’s about it. I guess it is good I don’t try to do audio very often.