Hi-pass filter on the master fader?

[ENW]

High Pass Filters?

I was looking at a frequency analyzer while playing a few commercially released tunes. I noticed a stark drop-off around 20 kHz.

Is anyone out there using a high pass filter at 20 k? If so, why?

Thanks,

ENW

[scribe]

Hi there,

Maybe because there is a bit of inaudible low end energy down there that is making the speakers move for no reason which could could mess up the bottom end clarity a bit. Just a guess.

Cheers.

[ENW]

Thanks Scribe:

I mistakenly said "hi-pass" while meaning lo-pass. I've been aware of the possible need to cut off the lowest frequencies to prevent rumble. I'd never considered any need to cut off above 20kHz where the sound is supposed to be inaudible.

Here is the explanation I received...

"Commercial tunes off a standard CD - there's an LPF in place prior to the Nyquist freq. (Nyquist = 22.05K (half the sampling rate @ 44.1k)."

"In short, yes - we are using an LPF around 20k."

Here is what I found corroborating that statement...

"In order to recover all Fourier components of a periodic waveform, it is necessary to use a sampling rate at least twice the highest waveform frequency. The Nyquist frequency, also called the Nyquist limit, is the highest frequency that can be coded at a given sampling rate in order to be able to fully reconstruct the signal (Wolfram Math World)."

Apparently (according to the Nyquist Theorem) it is only possible to produce frequencies up to half the sampling rate. I guess, to keep the data neat, the unusable higher frequencies are just eliminated.

Sometimes this stuff makes my head hurt

ENW

[StephenMarshMastering]

point of order - 20 Hertz low end - 20 Kilohertz - high end.

Now - I think were mixing terms here - the FREQUENCY of the sampling in an A to D conversion is 44100 - it doesn't sample up to 44.1Khz - it samples 44100 times per second. This ALLOWS it to capture audio frequencies up to 22.050.

An analogy: think of a wave form of the screen - you need as much space above the center line as below to get a full waveform - you need exactly DOUBLE the amount of space.

Make sense?--Steph

[the.blode]

I don't know that this is to do with sampling theorems.. All A/D converters will have aliasing filters which cut off everything above the nyquist frequency to eliminate aliasing side-bands. If this didn't happen then a 30khz signal would be a major hassle going into an a/d converter, even if it's outside our frequency range.. this is because The sampling process creates sideband frequencies in the same way as sum-and-difference sidebands created by multiplying sinewaves. So a 30Khz frequency will create a difference sideband at around 11khz of half it's amplitude.. that's not going to be friendly.

I think this lack of energy at 20Khz is probably just the Eq of the song - but I could be wrong.. there's not a lot happening at 20khz with most audio, anyway.

[brandondrury]

Some claim that harmonics going on in the inaudibly high regions is a great thing. That's the argument for high sample rates. I personally had way too many intense experiences on plain old 44.1K cds to care about higher sample rates.

I've done analysis of some very bright recordings and there still isn't much going on at 20K. That's just the way "good sounding" music sounds to us. 20Khz isn't of that much use to us really, but I certain wouldn't bother rolling it off.

It also needs to be said that guys in their 40s and up can't hear 20k. When I was at the Wagener Workshop, we fired up one of the guitar amps. The amp was hissing. I had to run out of the room with my hands over my ears. The other guys looked at me like I was stupid. (I was 25 at the time). Then they checked Wagener's frequency analysis thingy and it turned out that there was a HUGE spike @ 16k. It was killing me and they were oblivious to it.

Brandon