I have Windows XP, and installed 2.0 from the .exe installer…
I am trying to use Audacity for a different purpose – to determine decibel levels of recorded sounds at 8000 and 16,000 Hz. Can you teach me how to click my way to these determinations for my recordings?
The latest analyzer looks better than this. This is an early one. The dB values are down the left and the frequency is along the bottom. The higher the sampling size, the finer the detail. If you hover the cursor over a peak, it will read all the info on that peak.
I thought I was on the right track when I pulled down the “Plot” option – now, is there a way to enter a “zero” number/tone, and/or a “reference” into the data, so that I can see a “positive” spectrum? Is there a way to overlay the reference with the new sound so that I can see the net difference? Can I arbitrarily enter -80 db as a reference or zero that will give me a new plot for comparisons? Thanks for your help!
Nope. You get one display and it chooses its own reference based on the best display for the available data. That’s why I said it would be good to put a known reference tone in the show somewhere. My last complaint about Spectrum was you couldn’t ever see two of them on the screen at once without screen capture and Photoshop.
I think I get most of what you are saying…now at what point in the show is the Spectrum being shown, anyway? Is the Spectrum/Plot a composite of data over the entire period of the recording? The plot doesn’t seem to be somewhere that I could call up at a particular point in the recording, but if it’s not, I don’t see the purpose in the plot in the first place…thanks! Lou
The reference level for Plot Spectrum is the level of a 0 dB sine tone.
If you generate a full-scale (full height of the audio track) sine wave, it will show as 0 dB in Plot Spectrum. That is the reference, so all signals within the normal operating range will be below 0 dB, hence the dB scale showing negative numbers.
The spectrum is formed from an analysis of the entire audio selection, up to a maximum of about 10 million samples (a bit under 4 minutes at 44.1 kHz).
If the track is stereo, the the spectrum is based on the sum of the two channels.
If multiple tracks are selected, the spectrum is based on the sum of the selected tracks. All selected tracks must have the same sample rate.
More information is available in the manual: http://manual.audacityteam.org/manual/help/manual/man/analyze_menu.html#spectrum
It’s not unusual for people to want to take a spectrum at one point on a song. Waves have to change to get any reading at all and they have to repeat to get a spectrum. You can’t tell the arrival rate of waves at the beach unless you sit there and wait for at least two waves, and even that only gives you a bad sample. You need to wait for multiple waves to arrive before you can call your friends to announce the waves are 46 seconds at Zuma Beach. Then you have to average the sizes of multiple waves before you can tell people the surf is three feet at Playa del Rey.
So that’s the averaging and counting over time that the tool is doing to give you the graph. Given that the waves are different speed and size at Zuma and Playa, that would give you two peaks on the graph. One faster (left to right) and one larger (top to bottom).
All right…how about this attempt at getting it?! First, Steve, are you saying that the spectrum “curve” as it appears, is actually upside down, so to speak, so if I wanted to calculate the “area under the curve” I would determine the area of the white portion and not the blue portion?
I notice that if I export the Spectrum Plot into an excel spreadsheet, I get columns of data representing, I think, the curve, that is, a column of db vs a column of HZ…I haven’t yet plotted this data into an excel graph/plot to see what I get yet, but I could extract differences at 8000 and 16,000 Hz, one track vs another track, using Excel data and overlaying curves, I think?