Anti-Loudnesswar: I just discovered how to de-brickwall music

ahjteam

Anssi Tenhunen
This is not bullshit and it actually works. And its actually super fucking simple to do.

Step 1: Open the song in a wave-editor
Step 2: Drop the volume of the audiofile by 6dB
Step 3: Put a highpass filter at 30hz (I used the Waves Q10, I think it is 12dB/oct)
Step 4: Normalize to -0.3dB
Step 5: Save the file

sikth_debrickwall.jpg

deathmagnetic_debrickwall.jpg


As you can see, both are still pretty loud, but the coolest thing is that the difference is so minimal, the only thing you will propably notice is that it is only slightly quieter, but if you listen at the same volume as the brickwalled version, the loudest parts will sound a bit louder.

But what about if you have something under 30hz? Boo fucking hoo. Usually the mastering engineers highpass rock music at 35hz anyway and 99% of the speakers don't even play frequencies that low and you can't even hear them.

But does it remove the distortion on Death Magnetic? No, but it atleast sounds tolerateable after that
 
i dunno, man...seems to me like it's just a convoluted way of turning down the level of the track

of course it shows a lower RMS value now, along with more headroom - but you still haven't re-created the peaks in the dynamics that got lopped off during mixing/mastering
 
99% of the speakers don't even play frequencies that low and you can't even hear them.

Sorry, but I call bullshit.

Even the smallest Genelec sub has a frequency range of 25-85Hz (+-3dB), and most decent subs go even lower. Granted, the human ear can't hear sounds below 20Hz, but the air movement can still be felt.

Not that it's of significant importance here, just felt like nitpicking ;)
 
I can see your files have more dynamics in the end, visually, but I don't get where it is made, because you basically just deal with gain that have no impact on dynamics ?
 
i dunno, man...seems to me like it's just a convoluted way of turning down the level of the track

of course it shows a lower RMS value now, along with more headroom - but you still haven't re-created the peaks in the dynamics that got lopped off during mixing/mastering

That is where you are wrong :) Here is a picture of 5 second excerpt from the beginning of Cyanide that is lowered by 6dB and is displayed as the dark gray blob. The red areas are the peaks that are added in the process

debrickwall_addedpeaks.jpg


edit: LeSedna, its the highpass filter
 
I think I got it.

If I understood well, the first thing that get beyong the brickwall when you boost are the bass frequencies.

If you get rid of most of them, you make naked the other frequencies that still have their dynamics, but that are lost beyond the bass that you brain still have to suffer from.

Am I ok ?
 
You can't "un-limit) something just by slapping an EQ on it. Limiting affects the level of the track, you are just removing some frequencies, the mids and highs have still been affected by the limiting. The reason you are seeing more peaks is because the low end takes up more energy, and they are hitting the limiter more, cut those out and you SEE more peaks, but i bet if you listen to it, there is still distortion and it's still smashed to hell.
 
Yeah and he may agree. The deal is to cut the most limited part, so that it sounds "for-your-ears-and-your-brain" a bit less limited, even if it is still.
 
I am not entirely sure about the science behind it because I just thought of it up myself about 17 minutes ago, but I think that eventho if you highpass your track once, the highpassfilter works in a musical way and only removes like 6-48dB per octave, so that means if you highpass it at 12dB/oct@30hz, you reduce the 30hz only by 3dB and the 12dB at 15hz, which means that you can have massive amount like 30dB of excess mud there that has summed up from bad mixes. Then interpolation or resonance or some sort of other thing I don't know brings up the dynamics back up

but also if you read my first post, I clearly stated that this method doesn't remove the distortion if it already is there.
 
dropping the gain -6dB seems to me like dropping the volume knob on your hifi by 6dB, and the hi pass filter is going to do nothing to the dynamics of the song. The RMS is lowering because you have lowered the tracks gain and are therefore increasing the headroom - the difference between the quiet and loud parts will remain the same however, so effectively you aren't achieving anything other than a volume drop.

sorry.
 
dropping the gain -6dB seems to me like dropping the volume knob on your hifi by 6dB, and the hi pass filter is going to do nothing to the dynamics of the song. The RMS is lowering because you have lowered the tracks gain and are therefore increasing the headroom - the difference between the quiet and loud parts will remain the same however, so effectively you aren't achieving anything other than a volume drop.

sorry.

The reason for the -6dB volumedrop is just a precaution because the track would clip in the highpassfiltering process otherwise. With the normalization process you bring the volume back up to maximum allowed.

And listen to this clip, can you hear the difference in the drums?

http://www.ahjteam.com/upload/debrickwall.mp3 (short 12 second before and after clip, RMS matched)
 
The sum of the amplitudes of all the frequencies has become more varied.............giving the visual appearance of (and technically it's true) greater dynamic range

However, all the frequencies above 30hz, and their respective amplitudes, have just got slightly louder.



No changes have occured to anything above 30hz, and therefore you wont hear any changes in anything above 30hz.

The stuff below 30hz has just got quieter
 
The reason for the -6dB volumedrop is just a precaution because the track would clip in the highpassfiltering process otherwise. With the normalization process you bring the volume back up to maximum allowed.

And listen to this clip, can you hear the difference in the drums?

http://www.ahjteam.com/upload/debrickwall.mp3 (short 12 second before and after clip, RMS matched)


i hear a difference and it's called a high pass filter with volume compensation. Cutting the bass and then turning the volume up
 
iZotope RX has a DeClipper which makes a noticeable difference (visually and acoustically) to brick walled tracks - but I'm pretty sure that lowers the volume and then actually attempts to 'fill in the gaps' where the signal is clipped (at least that's what they claim).

Whatever it does, I think I'll stick to that :p

Steve
 
The sum of the amplitudes of all the frequencies has become more varied.............giving the visual appearance of (and technically it's true) greater dynamic range

However, all the frequencies above 30hz, and their respective amplitudes, have just got slightly louder.



No changes have occured to anything above 30hz, and therefore you wont hear any changes in anything above 30hz.

The stuff below 30hz has just got quieter

I was thinking of something too : the visual representation of amplitude is not linear, it's logarithmic. So a track visually peaked like a brickwalled track, if put -6dB, would visually look more dynamic.

EDIT : no, its the contrary, since where the difference in term of amplitude is the most readable in the extrema high amplitude. So if visually it looks more dynamic, and sounds more dynamic, why do you still say it is not more dynamic ?
 
Remember that when you introduce a highpass filter to audio even at 30hz, its going to cause a phase shift that will change the envelope of the original signal, causing higher peak levels. I'm not sure if this is such a desirable quality to bring to the table, but hey, if it makes it sound more dynamic, go for it.