Originally Posted by koturban
The decibel, for the most part, is a ratio
Most part? No, it *is* a ratio. It always is.
Without listing the absolute voltages (V (x) )that make up your graphs, "0db" has no context.
Of course it does. It says that the filter provided no attenuation. If I put in 1 volt, I get out 1 volt. If I put in 10 volts, I get 10 volts. That is the reason db is used. It gives us information that is invariant of what the input or output voltages are.
The ML may be "great" but at what voltage is it?
We don't care about the voltage in this situation. We only care about attenuation. Maybe an analogy would help.
Let's look at the common crossover settings in your AVR for the mains and sub:
Look to the left. It is a db scale. Not voltage values. dB scale tells us everything we need to know. In this situation, where the graphs cross "-6 db" line, is the frequency that is stated in Hertz. So 80 Hz means the response is down -6 db at 80 Hz for both low and high pass filters.
In our situation, the DAC in your AVR *must* have a low-pass filter. Incoming signal no matter how clean will change in time. If we did not filter out at least some of that, we may fail to detect the correct bits and actually get data errors. The filtering in the extracted timing allows us to ignore a lot of the variations so that we can correctly extract the right bits. The filter is a simple low pass filter just like your sub crossover filter in the AVR. It also has a -3 db point at some frequency. The test that I showed you samples that filter starting at 20 Hz and ending at 10 Khz. If the input and output are the same, it means that the -3 db point (so called "corner frequency") is well above the audio band. Or else we would see some attenuation on the way down to -3 db point. For the sake of discussion, the -3 db point may be 50 Khz. That would still let the DAC extract the correct bits, but will allow that "jitter" to get through the DAC. When that happens, we get sidebands that are in audio band.
The question you are asking with respect to what their voltage is, was addressed in the other graphs I showed where their amplitude relative to full scale can be read. It is not applicable to this measurement.
Since you like analogies, it's like filtering out a mouse fart in a hurricane.
. Our challenge then is to prove it is a mouse fart and that we are in a hurricane. Measurements are the first critical step in that direction.