FL Studio Useless for Brickwall EQ?

[JC Biffro]

So, I've been trying out some EQ presets, and I remember a number of different people mentioning on various videos I watched that FL Studio was poor at blocking out certain frequencies. The advice given most of the time was to use 3 Parametric EQs on a single track, rather than just the one. To the naked ear, that seems to work.

Aside from the built-in FL Studio EQ, I've used 2 other VSTs as well, and I've replicated 3 EQs on a single track. Following on from this, I've used a frequency analyzer, and there are STILL frequencies slipping through, even though I've bricked walled them THREE times.

An example track at the moment is I have a brick-wall high-pass filter on a melody at 200HZ, yet I can see (and faintly hear) frequencies around the 50-200HZ mark. Does anyone else have the same problem with FL?

I'm seriously considering using a different DAW to mix in.

 

[moses]

This is what happens when you don't read the manual and don't look up what all these numbers really mean.

 

[JC Biffro]

This is what happens when you don't read the manual and don't look up what all these numbers really mean.

I've read the manuals, all I seem to do these days is read!... Could you please explain exactly where I've gone wrong? I want all the frequencies from 0-200HZ cut off, I've applied a brickwall HP, yet the frequencies are still slipping through, albeit at a lower velocity.

 

[moses]

"Brickwall" filters don't exist. No EQ/filter claims to remove everything above or below the cutoff frequency. Instead, the filter cutoff is defined as the point where the filter reaches 3dB of attenuation. The filter action then usually continues with a more or less fixed slope, in most cases 6dB per octave or multiples of it (depending on the filter order). Draw it on paper, this definition clearly assumes that no filter ever reaches 100% attenuation in practice (which is correct from the physical point of view).

the term "Brickwall" is nonsense when it comes to filters and only confuses people.

It is possible to create extremely steep filters, but:

- There is no musically useful task for them. A 24dB/oct filter effectively sounds like brick-wall.
- They are very expensive cpu wise and always come with negative side-effects.

Another issue with your observation is the fact that realtime spectrum analyzers (the things you look at when you assume to "see" audio) are extremely inaccurate by their own. They are pretty useless below ~400Hz. So, even if such a filter would exist, no realtime spectrum analyzer on earth would be ever able to measure such a filter, it would be incapable to draw the graphics you expect!

Just trust your ears, that's all you need. They are several thousand times more accurate than you eyes.

 

[JC Biffro]

"Brickwall" filters don't exist. No EQ/filter claims to remove everything above or below the cutoff frequency. Instead, the filter cutoff is defined as the point where the filter reaches 3dB of attenuation. The filter action then usually continues with a more or less fixed slope, in most cases 6dB per octave or multiples of it (depending on the filter order). Draw it on paper, this definition clearly assumes that no filter ever reaches 100% attenuation in practice (which is correct from the physical point of view).

the term "Brickwall" is nonsense when it comes to filters and only confuses people.

It is possible to create extremely steep filters, but:

- There is no musically useful task for them. A 24dB/oct filter effectively sounds like brick-wall.
- They are very expensive cpu wise and always come with negative side-effects.

Another issue with your observation is the assumption that spectrum analyzers (the things you look at when you assume to "see" audio) are extremely inaccurate by their own. They are pretty useless below ~400Hz. So, even if such a filter would exist, no realtime spectrum analyzer on earth would be ever able to measure such a filter, it would be incapable to draw the graphics you expect!

Just trust your ears, that's all you need. They are several thousand times more accurate than you eyes.

Very confusing indeed, especially when 'big players' like iZotope use the term 'Brickwall' (which equates to a 90 degree angle linear line) as a filter option.

Thanks for the detailed response though, it is much appreciated. I changed the type of slope to 'flat' instead of this so-called Brickwall, and must admit it does sound a little better.

 

[moses]

Yes, definitely confusing.

BTW, in case you're interested. The idea with the "90°" angle exists as a theoretical construct. Scientists like to call this a "rectangle in the frequency domain" and you can actually explore how such a filter looks like when you take the inverse fourier transform* of a "rectangle". The result is the so called "sinc" impulse, which describes the perfect brick-wall filter discussed. In theory of course. The practical issue is that the sinc impulse both extends into the infinite past and infinite future! This is the technical explanation behind all this. So, it's not the best thing you want to calculated when you're in a hurry.

*the fourier transform is a pretty cool mathematical function which allows us to transform a time domain signal (e.g. a waveform, with amplitude being the Y axis and time the X axis) into a frequency domain signal (e.g. a spectrum analyzer, with amplitude being the Y axis and frequency the x-axis). On a side-note, you might have noticed that the latter signal has no straight forward time information, and this is where the real-time spectrum analyzer problems arise. The inverse function of the FFT also exists, that is, you can take frequency domain data (e.g. an EQ shape) and transform it into a time-domain signal (e.g. a filter impulse).

Check out wikipedia for "sinc" and "fourier transform"..

 

[Grave Thought]

So Ozone's brick wall filter (or filters like it) is just the closest thing we can logically get to an actual brick wall filter?

 

[sleepy]

The point is really that if you lowpass at 200hz it's not going to kill you if a little bit of the signal goes over that. If you really need nothing over 200hz, then go down a little bit frequency-wise instead of stacking the same filter with the same settings.

There's usually little need to be so precise with filters. Maybe you could see something extra with an analyzer, but the real question is, can you actually hear it?

---------- Post added at 03:27 AM ---------- Previous post was at 03:27 AM ----------

The point is really that if you lowpass at 200hz it's not going to kill you if a little bit of the signal goes over that. If you really need nothing over 200hz, then go down a little bit frequency-wise instead of stacking the same filter with the same settings.

There's usually little need to be so precise with filters. Maybe you could see something extra with an analyzer, but the real question is, can you actually hear it?