For this newbie, please. 2 questions about interpreting the audacity waveform.
What is the difference between the colors of the waveform, i.e. what do the light blue and dark blue portion represent?
I ripped a track off a CD w/EAC, semi-simultaneously cutting the .wav and the .mp3 (w/lame). The .wav is 66,803,900 big and the .mp3 is 15,155,439 (what I would have expected). However, when I load the .wav and the .mp3 into Audacity, the waveforms look identical. Not what I would have expected, as there was - naturally - some loss incurred during the creation of the .mp3. I would have expected there to be a difference in the .mp3. Is it because the loss is not discernabIe to the human eye? Are they really the same regardless of the input format? I would appreciate it if someone could clear this up for me.
So, is there any way within a waveform to discern a .wav from .mp3 of the same track. As you say “… overcompression in MP3 …” is “… (intentional damage) …”. Don’t all of the “delicate overtones” and “complex bowing, strike, tail, and emphasis” show up as subtle peaks and valleys in the .wav waveform?
I’m also intriqued by another statement in your answer; “… playing it in Windows Media (not Audacity) to make sure it sounds OK …” Why? What’s the difference in the playback quality? I prefer Winamp anyway, since it does not make as many presumptions as WM, but can I safely presume that the playback quality is the same 'twxit WM and Winamp?
Yes they do, but you are unlikely to be able to identify them just from looking.
When you compress audio a great deal as an mp3, the bandwidth will be reduced (the upper frequencies will be cut off) and you can see this if you look at a “spectrum view”.
Assuming that you are using 128 kbps mp3 or above, this does not happen, and so the main difference between the mp3 and the original wav will be very small differences where the mp3 compression algorithm has managed to save some data bits. These differences will not usually be noticeable in a wave display. With certain test signals, you may see a difference if you compare compressed/uncompressed waveforms side by side, for example, you may see signs of “ringing” around short audio pulses, but these would be difficult to spot in complex waveforms of music. The most obvious visual difference would be that the mp3 has a few milliseconds of silence added at the beginning, but this would only be noticeable if you have the original to compare with, or if you know that the recording does not start as silence.
Audacity supports some formats that are not supported by all media players, so it’s worth testing the exported audio in the target software.
Anyone out there want to tackle the different color blues in the waveform question … what is represented by the light blue and what is represented by the dark blue.
It used to be one was peak value and the other was Something Else. Average sound level? RMS values? I gotta look it up.
One of the reasons we urge people to immediately convert the timeline to dB rather than percent, is how the ear works. (Black down arrow on the left > Waveform dB.)
0.5 is not half loudness to your ear. 0.125 is. Now look at where 0.125 is on the percent scale, and you’re only down to half volume. You can go half volume another two or three times before the performance vanishes. The dB range reflects that oddity. 0.5 percent is only -6 dB. Half volume is -18 dB. Pull the bottom of your timeline down to see better values on the left.
Notice where -18 is on that waveform view (not the meters). Half way to silence.
The meters show you what’s really happening. You can see those change all the way down to complete silence (-60s).
So that gives you an idea of the subtle changes that sound compressors make to the show. Some changes are in the -30 dB range. That kind of change is literally impossible to see on a waveform view I don’t care which scale you use. I will say that there are carefully created test signals that engineers use to manage all this stuff, but even they can’t actually see it. They depend on instruments to tell them what happened.
I can answer the color question. Dark blue represents the highest peak within that small section of the waveform (remember that each pixel might be representing hundreds of samples, depending on how far you’re zoomed out). Light blue represents the local RMS value (which is better at visually representing the Loudness of a signal than the peaks are).
Also, there is an easy way of figuring out what the difference between an mp3 file and a WAV file. Just subtract one from the other like so:
Import copies of both the mp3 and Wav files into the same Audacity project.
Amplify both to the peak volume, use the amplify effect’s default value on each signal separately.
Zoom way in on a distinct part of the waveform. Zooming on something percussive will make this part much easier, it’s easier to see.
Use the Time-Shift Tool to line up the waveforms exactly. Keep zooming in and adjusting until you’re lined up as accurately as possible, make sure you can see each sample by the time you’re done.
Now invert one of the signals, either will work.
Highlight both signals and select Tracks → Mix and Render or Project → Quick Mix.
The signal you have left will be the difference between the two original files. Everything here is what you lost when you went to the mp3 format, it’s mostly high frequencies and quick changes in dynamics (such as percussion).
You probably won’t find the final product nearly as exciting as I do. Fair warning.
I still don’t understand where you get those figures from.
A doubling of power is equivalent to a 6dB change.
0.5 percent of what? You must mean +/- 0.5 indicates -6dB.
It’s not a % scale though - but even so, if you compare a single frequency that measures +/- 1 to the same frequency at an amplitude measuring +/- 0.125, then the smaller amplitude (-18dB) sounds much lower than half volume, at least it does to my ears.
“Volume” and “Loudness” have no direct relationship with power (intensity).
Loudness is the subjective experience of how loud something is, and is dependant on frequency among other things.
My preference for the vertical scale would be a linear amplitude (voltage), numbered in dB or bits (which is the way most other programs show it - preferably an option for either) which creates a display like Audacities scale of -1 to +1 (but with different numbers). I’m not at all keen on the Waveform dB scale. I prefer sine waves to look like sine waves, and linear fades to look like linear fades.
A doubling of power is equivalent to a 6dB change.
Actually, a doubling of amplitude is equal to a +6dB change. Power is a product of both amplitude and frequency, so your statement is only true at 1Hz. For instance, a 100Hz tone has half as much power as a 200Hz tone at the same amplitude, but there’s no 6dB difference.
You’re right about what Koz wrote, he shouldn’t have used the word “percent.” Half amplitude is -6dB, I think we can all totally agree on that.
As for the “half loudness” argument, it’s too subjective. I’ve seen -18dB as a general ball-park measure of half loudness that many people agree on, but I’ve also seen -10dB for the same thing. It really depends on the person and the environment.
Interestingly, our ears are more sensitive to changes in volume at the extremes of our dynamic range. A +1dB change is more noticeable if the starting sound is either very quiet or very loud, but if we’re starting from a moderate volume level, a +1dB change might go unnoticed if you’re not listening carefully. So even if there were a generally agreed-upon “half loudness” factor, it would shift depending on the starting loudness.
Finally, I also prefer the linear display scale. It’s easier to see how close to clipping you are.