As a portion of my masters thesis I needed some decibel readings of different bird songs at various frequency ranges. To get these I was using the plot spectrum tool to analyze some bird song recordings to determine decibel levels at various Hz ranges, but am a little confused as to how to get positive decibel values. I have read online that decibels are a relative measure, but am still confused as to how to convert them to positive values especially in audacity. These are not sound files that I recorded myself they were downloads from the all about birds website so I do not have any information about the equipment that was used to capture the audio.
The software I am going to be putting these values into does not seem to like the negative values and from what I have been reading it is not as simple as for example taking my 2kHz measure of -64db and simply making it a positive 64dB.
I am using version 2.2.2 on windows 10 home.
Any and all help would be greatly appreciated.
Basically, you don’t. You need to use a calibrated “Sound Pressure” meter.
The best that you can do with a recording is to say that one signal is xx dB high or lower than another signal. Unless you have a reference level in the recording, and know the distance between sound source and microphone, there is no absolute measure that you can make.
A simple way to think of it, is that a recording of sound is like a photograph. How would you convert the size of something in a photograph, to it’s actual size in real life?
I have read online that decibels are a relative measure
Just adding to what Steve said -
The digital 0dBFS (zero decibels full scale) reference is the “digital maximum”. i.e. It’s the maximum you can “count to” with a given number of bits.* So digital dB values are normally negative. (We usually leave out the “FS” as long as we know we are in the digital domain.)
Acoustic loudness is measured in dB SPL (sound pressure level). The 0dB reference is approximately the quietest sound that can be heard so dB SPL levels are usually positive. Again, we often leave out the “SPL” as long as we know we’re talking about the loudness of sound in the air.
There are also electrical dB references for voltage & wattage.
Usually, there is no calibration so if you record something with your computer and microphone and I record the same thing with my computer & microphone (or some other digital recording device) we’ll get different digital levels.
The same goes for playback. My amplifier & speakers may be bigger than yours, or you may have your volume control turned-up higher than mine, or you may be sitting closer to the speakers, etc.
Movie theaters are calibrated, but it still depends on where you are sitting.
However, there is a direct correlation. If you are playing a digital file and it’s coming-out at 90dB SPL and then you reduce the digital level (or the electical level) by -3dB, the acoustic level will also be reduced by -3dB to -87dB. (That’s assuming that everything is linear… i.e. You aren’t driving the amplifier in to distortion, etc.)
\
The binary numbers in a 24-bit file are bigger than the numbers in an 8-bit file, but everything is automatically scaled so a 0dB 8-bit file plays just as loud as a 0dB 24-bit file.
Floating point audio uses a different reference and it can go over 0dB (there is essentially no upper or lower limit). Audacity uses floating-pint “internally” so it can go over 0dB. But, your digital-to-analog converter (playback) and your analog-to-digital converter (recording) are integer-based and hard-limited to 0dB so you have to “be careful” to avoid clipping (distortion).
It dips at the low pitch and high pitch ends leaving only the middle. You know what that middle pitch is. That’s baby screaming on a jet. Your ear works super well right there. “A” is also the weight used in Health and Safety laws.
“C” Weight is more or less flat across all tones and only dips at the extremes of low and high pitch. This one peels off right around humming bird conversations at the very high end and earthquake and thunder tones at the very low. The low tones feature events such as: for some reason the windows in your house are shaking and the cat is going nuts. C Weight won’t measure those.
“Z” Weighing is dead flat across all pitches of sound.
So pick your poison. What are you comparing the birds to?
When there’s the actual execution. Typically, the sound under measurement is there for a relatively long time because the meters take a while to register. I have no idea how you’re going to get a bird to perform long enough to measure, and I have no idea how you’re going to generate a comparison to convert a sound recording into Sound Pressure Levels.
It is possible to convert, but all the conversions I know of start the sentence with “Set up your SPL meter…” Expose both the meter and your recording equipment to the same calibration sound and write down the readings. Then record something, but don’t change anything.
Also as above, it is possible to not use an SPL meter, but you need a Standard Bird.
Make a recording of the Standard Bird and then, without changing your recorder at all, measure the Mystery Bird. You can make statements such as: "The Mystery Bird is 12dB louder than our Standard Bird. That’s a relational measurement and doesn’t get into SPL or anything else.
How are you recording your birdsong? I would use a stand-alone sound recorder, not a voice recorder and not the recorder in the computer.
Since you are particularly sensitive to quality of sound, you can set up a sound recorder to not apply effects, corrections, equalization or automatic processing to the sound. Just record this thing and don’t help me. I have a modest Zoom H1n that works pretty well. I have a Zoom H4 that used to work OK. I think we’re going to hold funeral services…
Voice Recorders assume you’re interested in the words and clear speech and you don’t much care how you get there. I have a voice recorder that has corrections and automatic settings I can’t turn off. The last thing you want is the recorder changing volume automatically.
The forum is crawling with people struggling to turn off all the Windows “enhancements” so they can get a clean recording. Most succeed, some don’t. If you really offend the sound angels, Audacity won’t record new work in your Windows at all.
