I have Ubuntu 14.04.02 x64 and Audacity 2.0.5. My questions concern setting appropriate sound levels to maintain the greatest SNR possible. I have used, at one time or another, PulseAudio Volume Control, alsamixer, and the Sound applet in System Settings. It seems that these all can affect the sound levels of line or microphone input. Since these three applications can change settings in each other (at least in some cases), there must be some rules for their uses. One might possibly be, if I am setting sound levels, which should I use and in what order? Another could be, is there one of these three that can handle everything? I know that PulseAudio is the audio server for Ubuntu, but it seems to me that I cannot completely control input sound levels in Audacity from PulseAudio Volume Control alone. Am I wrong? Another question is, in PulseAudio Volume Control on the input devices tab, what is the reference point labeled Base? Finally, is there a tried and true method for setting the volumes so that one can obtain the best SNR possible?
The only real place to set signal to noise is before the digitizer. After that, the noise floor is whatever the digital standard is that you picked. If you’re in 16 bit, it’s -96dB.
So that takes you to the microphone and the microphone preamplifier. The MicPre. Microphone signal is insanely low and delicate. The MicPre is designed to boost it so it can be used in production, and do it without adding noise of its own. But it does. Most modern mixers push the type of MicPre they have because it’s so important. “Our mixers feature European Signal Spritzer MicPres!!” (we assume because they’re quiet and well-behaved).
That’s why. Once the MicPre adds its own noise to the voice or music, that’s the best the S/N will ever get, no matter what else you do. You can certainly make it worse, but not better.
For live recording, we picked -6dB, the yellow zone in the Audacity sound meters as a happy place. It gives you room for slightly louder than normal expression without overloading and a good distance louder than the noise.
This is the improved sound meter available in Audacity 2.1.1.
Perhaps I was not clear in my question of SNR. While I understand the hardware issues and the digital standard selection, there is another issue that can affect SNR. If one chooses a value that is too high on an input port, say 70%, one may well get clipping or “flat waves” in Audacity. If one lowers the level to 50%, then the noise is a larger component of the sound wave and therefor the SNR is lower. Since input levels are not controlled in Audacity, one must use another software tool to set this level. As I mentioned, there are a number of options in Ubuntu, and use of one of these options affects the settings in the others, but the effect is not identical, i.e., if I set the line input level to 70% using alsamixer, the line input level in PulseAudio Volume Control (pavcontrol) will show something entirely different (in my case 24% or -37.5 db). Now if I try to record something with a large dynamic range, I will get flat wave forms in Audacity. In this case, I must lower the alsamixer setting to 18% which will lower the pvacontrol setting to 10% (-60db). At that low of a level, the noise component of the wave is much greater, so the SNR is lower. What I am trying to do is maximize the input sound level so that the noise component is less (thereby getting a greater SNR) without flattening the sound wave or clipping. I would like to do this by working with only one software tool in Ubuntu so that I can get a more precise setting (i.e., the “best” setting) for SNR as described above.
In that case, bypass PulseAudio.
Select one of the hardware devices (name ends with “hw:0” or “hw:1” or similar) for recording and playback devices in the device toolbar. Then set the levels using Alsamixer.
Note that if you do this, Audacity must (usually) be the only application accessing your sound card.
Great! That’s what I needed. BTW, is the reason to use alsamixer is that it directly affects the sound card whereas pvacontrol doesn’t? Just for my information.
One of the most important things that ALSA does is that it communicates with the sound card hardware.
PulseAudio is a software layer between applications and ALSA. It provides all of the features that we have come to expect of a modern sound system. In particular it allows multiple applications to access the same sound card at the same time, each with independent level control and even independent sample rates.
When you use the “hw” devices, you are telling Audacity, to use the sound card hardware directly (via ALSA), bypassing the Pulse layer.
Thanks again. Very clear and understandable explanations. I appreciate your assistance.