Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[negopus]

Since I installed MC 11.1 I have been interested in the new real-time spectrum analyzer.

I have played with it a bit using Fred Nachbaur's test signals, and I figured out better how MC's spectrum analyzer works.

The frequency scale of the analyzer is not linear, as I first guessed. It behaves more or less like a third-octave analyzer, but it is not really a standard third-octave analyzer, which has only 31 bands from 20 Hz to 20 kHz. There are 32 bars in MC's analyzer, and the upper frequency is limited to 16 kHz. The number 32 was clearly chosen because it is programmer-friendly (it is a power of 2), and to double the 16 bars in Windows Media Player. But the 32 bars in the analyzer do not exactly match with the bands in a standard third-octave analyzer. They are wider in the low part of the spectrum (20 Hz and 25 Hz show up in the same bar) while the are narrower in the high part of the spectrum, so that 4 kHz and 5 kHz are have a bar in between, and are not adjacent as in a standard third-octave analyzer). That gave me the (wrong) impression that the scale was more linear than logarithmic.

I also pointed out that it seemed that the analyzer introduced a +3dB/oct boost, because a pink noise would show a flat spectrum. The remark would have been correct for a linear frequency scale. But for a third-octave analyzer (with constant Q bands) it is correct that a pink noise exhibits a flat response. Pink noise is defined as a random signal with the same power content in each octave. So as we increase the center frequency of the band, the width of the band increases too. For every octave, the width of the band doubles. The power content, on the other hand, decreases by 3 dB (halves) per octave. So the power contained in each band remains the same. The spectrum analyzer is thus behaving correctly.

Audio-conscious users (audio geeks) would prefer to have each bar in the analyzer to correspond to a standard band in a third-octave analyzer (there are 31 bands from 20 Hz to 20 kHz, but since MC's analyzer limits its frequency range to 16 kHz,  only 30 bars remain).

A global level meter should go in the Action Windows instead, as I pointed out in another thread.

Any thoughts?

[marko]

it's just eye-candy to me. sorry.

[negopus]

It could become a serious tool only with some minor tweaks.

[jgreen]

I like to think of it as eye-parsley.  Left-handed/backwards eye-parsley.

Although I am grateful to Negopus for taking the time to write in depth about this stuff--I'm learning a lot from it. 

[hit_ny]

I think nego's input on this matter is valuable.

why not have an analyser that's tone perfect ?

MC is already sample perfect, this would give it just that little bit more.

...keep bumping this topic till it gets more traction, keep up the good fight.

If you are going to examine JRiver's work with a magnifying glass, i think they will be happy to oblige.

[Matt]

The spectrum analyzers in the player area are tuned by hand for looks. 

They're not intended to be mathematical measuring equipment.

You can create 2D visualizations of spectrum analyzers in MC and lots of variables are user selectable using Visualization Studio.

[negopus]

Thanks for the tip. There are parameters in Visualization Studio (whose exact meaning I don't know) that make me believe that, when tweaked, can make it act like a standard third-octave band analyzer. You should include such settings for a standard third-octave analyzer by default, however.

I think that Media Center has thousand of features buried under its hood. Probably I am using only 1% of Media Center (and I am already satisfied with it...). It deserves some study of the documentation, as well as some experimenting.

I am going to take a look at Visualization Studio. Documentation is not always easy to find for all areas in Media Center. I don't know where to find documentation for Visualization Studio, for example