• Welcome to BirdForum, the internet's largest birding community with thousands of members from all over the world. The forums are dedicated to wild birds, birding, binoculars and equipment and all that goes with it.

    Please register for an account to take part in the discussions in the forum, post your pictures in the gallery and more.
ZEISS DTI thermal imaging cameras. For more discoveries at night, and during the day.

Audacity and file editing (1 Viewer)

Avetarda

Well-known member
Spain
What kinds of adjustments do you make in audacity? If I use "noise reduction" or "amplify", the results are usually not good ... Probably because I make those resources wrong. Which ones do you use and in what way? Although I have read that the less a file is edited the better.
 
You are quite right, repeated re-editting is not good which is why rule No 1 is to keep your original somewhere safe where you are not going to overwrite it. Copy it into and edit in a different folder.

The functions are accessed through the Effects Tab

Usually the first thing folks do is noise reduction, though this may well prove to be the wrong first step.

If you have a very long recording you will frequently find that editing it into suitable shorter sections makes any corrections easier and more effective.

Amplification for very weak/low amplitude recordings may well give noise reduction a better chance to identify and subtract noise without too great an adverse effect on the recording.

For this one merely selects the whole of the recording or sub-section. You then amplify it with the clipping box unticked. The software will recommend a figure and this will usually prove optimum.



Noise Reduction will only efficiently remove a background hiss or other noise that is fairly constant - clearly anything in the same frequency range as your noise will be degraded to an extent that varies with each recording for some uses noise reduction may be counter productive. Normally for bird identification purposes you can use noise reduction if the noise to signal ratio gives the signal a chance - practice gives you experience.

You need to select a part of the recording that represents a typical sample of the noise you are trying to remove the selection area will have a clear background.

For any recording you should include a few seconds without the birdsong or other extraneous noise - this is the typical sample of noise mentioned below. For long recordings this may occur naturally and one should be included in each sub-file you have created from the master.

You run noise reduction once on the selected ' noise' to calibrate and a second time having selected the whole recording to remove the noise.

It can remove pre-amp noise, noise from the microphone circuitry, cooling fan noise from some video cameras, airconditioning, road hum etc., and if reasonably constant some types of wind noise.

By keeping the original file, you have access to the unedited version if the noise reduction proves detrimental.

There are other approaches to noise reduction, but the simple function will meet most needs for many people.

If you have problem files please attach a sample and I am sure I or someone else will advise you further.

I started this some hours ago and had to come back to finish it so may well be repeating others responses for which I apologise.

J
 
Last edited:
The files that I upload to Xeno Canto are the originals, are in raw.
I upload an original sample of the singing of a Buteo (as an example of a bad recording), another with noise reduction and an amplified one. I don't like the results at all! It may be the configuration of both tools as well, that they are incorrectly configured, although they are with the default measurements of Audacity.

My configuration:

Noise reduction- dB 12; sensibility 6,00; frequency s. 3
Amplif.- 22 dB
 

Attachments

  • original.mp3
    200.9 KB
  • noise_r.mp3
    132.7 KB
  • ampli.mp3
    133.5 KB
I believe the general guidelines for editing files for upload to both Macauley and Xeno-canto can be reduced to - simpler is better. Neither wants heavily edited or processed files, especially when some are being used for science or collections.

Macauley's guidelines are fairly easy to find; X-C's not so much. Here's the link. It's actually buried in the second page of the Upload Sounds process!

Both repositories recommend first trimming handling noise at the beginning and end of the recording - you don't want amplify them! Then, 'normalize' (also in the effects tab) to a peak of -3.0 dB. Brings all recordings in the repositories to a somewhat common level of enough amplification without clipping.

Some gentle high-pass filtering to reduce wind or road noise might be useful, e.g., 3 or 6 dB/octave roll-off below 1KHz, but beyond that, many of the more advanced noise reductions protocols result in some unwanted anomalies.

I'm not trying to contradict Iveljay's guidance - I'm a minimalist at heart and these guidelines resonate with me.

All the best!

Scott
 
Thanks! I had not read / found the indications or advice of XC.

When you say "Some gentle high-pass filtering to reduce wind or road noise might be useful, e.g., 3 or 6 dB/octave roll-off below 1KHz, but beyond that, many of the more advanced noise reductions protocols result in some unwanted anomalies.", what function of Audacity is the filter of wind or road noise?
Although it is clear to me that the best is to record near of the birds in a silence place 😅
 
Clearly you have to take into consideration the final use of your files.

My notes were purely for first timers experimenting with Audacity and being dissatisfied with first results and not intended for folks with more specific needs such as uploading to X-C.
Other software packages may suit different people, but getting the hang of a simple one makes others more accessible.

The requirements for Macauley and Xeno-Canto were brought up a few threads back and clearly it makes no sense to disregard their guidance for that particular usage.

Folks who have invested in audio recording equipment for the first time like to feel they havn't wasted their money.
We have all sat through presentations where photographs, video or audio are presented so that their lack of editting and basic cleaning up has left the audience wondering whether to laugh or walk out.
In the end you sit quietly, appplaud, thank the speaker and make a note not to ask them back.
Unfortunately it will also reinforce the beliefs of those prejudiced against birding and birders in general.
There will always be cases where you need to use sub-standard material as its all you have at your disposal, but especially for non-specialist audiences you are their to entertain as well as disseminate knowledge and inspire one or two of the audience to actually join a local society or contribute to an nationally organised survey.

The most common faults that need the blunt force approach mentioned above, tend to be excessive noise from pre-amp stretched to the limit, microphone noise from inexpensive microphones and recordings where the signal from the microphone is so low that it is a barely discernable ripple. Things like handling noise are down to technique, but again do not use recordings contaminated by such basic faults unless vital and one should at least trying to edit the worst out by cutting or using point reductions in amplitude to minimise its effect.

By actually practicing using the available tools you will be able to understand their limitations and strengths and answer most of your questions. There is rarely a one size fits all answer.

So you need to consider the use of your recordings, perhaps even having different versions for different uses.

Most contributors to this forum have perfectly valid points of view, depending on your interests and needs you will always find that some advice or discussion is more relevent to you than others.
It is one reason why I hang around birdforum - I am always learning something new and interesting - not just from the audio section.
 
one more point: the sound quality of normalized (or amplified in general) files depends on its bit depth as soon as you plan to post-process the recording.

So if your recorder's settings are capable of 24 bit (or even 32 bit floating point), then try to choose that highest possible setting because it is less prone to "clipping" in the case of signal amplification.
 
What kinds of adjustments do you make in audacity? If I use "noise reduction" or "amplify", the results are usually not good ... Probably because I make those resources wrong. Which ones do you use and in what way? Although I have read that the less a file is edited the better.

Audacity --> I use Effect - Graphic EQ (Both for reducing the low frecuencies (noise) and increasing the part of the sound I want.)

I think it is better to record with low level noise and increase the sound you want in post rather than record with more general noise and try to fix it.
 
According to the editing guide of XC and Macauley when I use the filter "Normalize" to -3db, it increases the sound but also the background noise. I'm using medium and low gains so I get low sound recordings even though the bird is near and I wish I could turn up the volume on the recordings. Any tip?
 
Well, normalizing raises the amplitude TO a certain level under maximum (e.g. -3 dB),

The amplifying function increases the level BY a certain value.

So when you normalize, the loudest part of your recording will determine how much the level is raised.

This means that if there is a loud crackling somewhere in the recording, an otherwise very quiet recording will only be amplified slightly or not at all.

That is why you should decide individually whether normalizing or amplifying leads to better results.

Regardless of this, it is important to adjust gain as precisely as possible during the recording, because in the later recording only 16 or 24 bits are available (If you normalize a very quiet recording, only few bits to represent the amplitude changes will be spread to 24 bit > valuable details are lost due to amplitude clipping).

Best way to get the perfect gain is to choose a recorder supporting 32 bit floating point.

Even with the weakest signals or the strongest level amplification, there is never any risk of the dynamics of the processed recording being impaired.
 
Well, normalizing raises the amplitude TO a certain level under maximum (e.g. -3 dB),

The amplifying function increases the level BY a certain value.

So when you normalize, the loudest part of your recording will determine how much the level is raised.

This means that if there is a loud crackling somewhere in the recording, an otherwise very quiet recording will only be amplified slightly or not at all.

That is why you should decide individually whether normalizing or amplifying leads to better results.

Regardless of this, it is important to adjust gain as precisely as possible during the recording, because in the later recording only 16 or 24 bits are available (If you normalize a very quiet recording, only few bits to represent the amplitude changes will be spread to 24 bit > valuable details are lost due to amplitude clipping).

Best way to get the perfect gain is to choose a recorder supporting 32 bit floating point.

Even with the weakest signals or the strongest level amplification, there is never any risk of the dynamics of the processed recording being impaired.
Here, and in other places, the 32-bit floating point technology has been mentioned and even praised. I notice that Zoom now has a very small form factor 32-bit recorder - the F2 - that accepts XLR inputs. 32 float appears to be so adaptable that this recorder doesn't even have a gain control! You record without fear of clipping, and adjust (normalize or amplify) in post This recorder is small enough that it can clip to your belt.

I can't help but wonder if this recorder, married up to a really low-noise pre-amped shotgun, could be very effective package while keeping the size down. Any thoughts?

Scott
 
Here, and in other places, the 32-bit floating point technology has been mentioned and even praised. I notice that Zoom now has a very small form factor 32-bit recorder - the F2 - that accepts XLR inputs. 32 float appears to be so adaptable that this recorder doesn't even have a gain control! You record without fear of clipping, and adjust (normalize or amplify) in post This recorder is small enough that it can clip to your belt.

I can't help but wonder if this recorder, married up to a really low-noise pre-amped shotgun, could be very effective package while keeping the size down. Any thoughts?

Scott
Scott,

heard about the new Zoom F2 recently and that tiny box could be a powerful solution on the lower budget side.

The F2 can be paired with a single Plug-in-power mic via 3,5 mm jack, no XLR, no phantom power, no display for status or browsing files and no stereo recordings - it's mono only. Zoom's website shows no info about the preamp's signal to noise ratio, maybe here's someone with some samples to share with us.

Cheers)
 
Many reviews for the Zoom F2, no definitive s/n ratio, but many people like it. Main noise problem seems to be the input jack on early examples, that was distinctly a problem, that may be fixed by now. Otherwise the only real criticism is of its 'lightweight construction' and inability to format memory cards without using a computer. Reviewers and recording examples are mainly in wedding video production or similar, so definitely a need for someone in the birding world to take one into the field and report back.

Thanks to Stein, for the link to the 32 bit editing guide.
 
Hi everyone,
I see the discussion above is what to do first etc. For most recordings it does NOT hurt to reduce the lowest 300-500 Hz where you find most background noise (distant traffic, noisy shoreline, wind in trees etc.) I stay ab 300 mtr from the sea and most of my recordings made here has a low noise that I always reduce with -6 or 12 dB in Effect>High pass Filter. But EQ works also perfect here! For White noise I seldom use more than -5dB trying in Noisereduction not to reduce the quality of the recorded song/sounds. It probably has an impact but I can not hear it, for those who wants to work scientifically on these recordings I will rather send them the original wav.
I went through some recordings made in Tanzania overnight New Years Eve 2018 and a strong insect-sound buzzing all night at 5kHz were reduced by Equalizer when this was above the birds sounds XC679780 Crested Guineafowl (Guttera pucherani) and https://www.xeno-canto.org/679785
 
Warning! This thread is more than 3 years ago old.
It's likely that no further discussion is required, in which case we recommend starting a new thread. If however you feel your response is required you can still do so.

Users who are viewing this thread

Back
Top