Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Equalizer distortion? (not the obvious kind) (Read 6051 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Equalizer distortion? (not the obvious kind)

So I've always known that software equalizers set to boost frequencies will inevitably boost.  I keep all my equalizer settings below the zero line and watch the peak meter on fb2k just in case.

But the music I'm listening to recently (old Jimi Hendrix albums on the one hand and J-Pop from this really great band Hokago Tea Time (  ) on the other hand are distorting so often that I'm having a hard time telling whether it's in my setup or just part of the music.  It's also disconcerting that these $10 earbuds are now picking up every bit of distortion that's usually covered up by phones at this price.

I'm using pretty radical EQ settings:

(if you're interested in how I arrived at these settings, click here
And I hear that narrow-Q filters on an EQ can cause distortion.  How narrow are we talking about?  The narrowest Q filter I have right now is 0.2, at ~11kHz -4.5dB.

BTW, is this the best subforum for talking about audio DSP processing for listening output?  I'd really love to find an active forum / subforum dedicated to this.

Equalizer distortion? (not the obvious kind)

Reply #1 –
I believe those $11 IEM are giving you $11 performance. EQ can help but, if possible, it is best to attack the problem at its source. Your results demonstrate why.

Equalizer distortion? (not the obvious kind)

Reply #2 –
Even when frequency response is completely below 0 dB, EQ may still increase peak levels of the recording, which may lead to clipping.

Equalizer distortion? (not the obvious kind)

Reply #3 –
I see no reliable evidence that confirms your perception that the equalizer itself is distorting.

There are many other sources of distortion, including the earphones, the headphone amplifier (whether built in or external) and your ears themselves.

Finding out what the exact problem is, could take time and effort. Are you up for it?

Equalizer distortion? (not the obvious kind)

Reply #4 –
Quote from: Alexey Lukin on
Even when frequency response is completely below 0 dB, EQ may still increase peak levels of the recording, which may lead to clipping.

Nah. Don't see how that would happen. But attenuation in the EQ will require you to add gain elsewhere in the signal path and you could get distortion there. Another distinct possibility is that the distortion is happening in the transducer itself.

Equalizer distortion? (not the obvious kind)

Reply #5 –
Are you saying that a FIR filter for a narrow band reject would have no coefficients greater than 1.0?

Equalizer distortion? (not the obvious kind)

Reply #6 –
Quote from: Notat on
Quote from: Alexey Lukin on
Even when frequency response is completely below 0 dB, EQ may still increase peak levels of the recording, which may lead to clipping.

Nah. Don't see how that would happen...
I'm not going to say that's the problem here, but I believe that can happen due to phase-shift.  If you slightly time-shift some frequency components, you can get higher than original peaks when these time-shifted frequencies are re-combined with non-shifted frequencies.  This is not a bump in the frequency response or a bump in the average level, it's just some peaks being increased and other peaks being reduced.

Equalizer distortion? (not the obvious kind)

Reply #7 –
Good (esoteric) point DVDdoug. This concept is often used in reverse to help advance the loudness war. EQ would undo some of it. Maybe there is something to it.

Equalizer distortion? (not the obvious kind)

Reply #8 –
Quote from: Notat on
Good (esoteric) point DVDdoug. This concept is often used in reverse to help advance the loudness war. EQ would undo some of it. Maybe there is something to it.

assuming that:
y[n] = sum(h[1:m]*x[n-(0:m-1)])
max(abs(x)) <= 1 for any i

We want:
max(abs(y)) <= 1 for any i

A sufficient requirement would be:
h[1:m]=h[1:m]/sum(abs(h[1:m]))

This takes into account the worst-case input where x[n-(0:m-1)] is a train of +/-1 pulses of the same polarity as the corresponding element in h.

Equalizer distortion? (not the obvious kind)

Reply #9 –
@Joe: Simply convert some of the tracks in question with fb2k to 32-bit wav with the equalizer enabled. Analyze resulting files with replaygain and check if there are any peaks > 1.0. It's not a guarantee that you'll never see clipping with your EQ preset though, on the other hand really tiny amounts / very rare occasion of clipping probably won't be noticable.

What mode (IIR, etc.) are you using with Electri-Q? Could you share that preset?
"I hear it when I see it."

Equalizer distortion? (not the obvious kind)

Reply #10 –
Quote from: Notat on
Nah. Don't see how that would happen. But attenuation in the EQ will require you to add gain elsewhere in the signal path and you could get distortion there. Another distinct possibility is that the distortion is happening in the transducer itself.
Try summing up a sine wave of frequency f and an attenuated (20%) sine wave at 3*f. This will somewhat flatten (lower) the peak of the base wave. If you filter out the 3*f component from such a signal, it will raise the amplitude back.
Or simply try low-passing a square wave.
 

Equalizer distortion? (not the obvious kind)

Reply #12 –
Does the distortion go with the EQ disabled?
daefeatures.co.uk