Question about rockbox equalizer, or equalizers in general

[grzzlo]

I don't know if this question has been asked before, and if it has then sorry.

My rockboxed h140 clips when I use the eq, so I was wondering about any easy way to get around it. I think some people use replaygain, or whatever, to turn down the volume of all their tracks, so they can boost some bands, but I haven't used replaygain, so I don't know how much work that would be to update all my tags (I imagine not to hard).

Anyway, here's my question: is there a difference btw setting the eq at +2,+2,0,-1,-1 vs. 0,0,-2,-3,-3 with the volume turned up to compensate? It seems like there shouldn't be, but does anyone know for sure? Is there any reason the two should sound different?

 

[bigshot]

One is called additive equalization and the other is called subtractive. Subtractive equalization is usually much cleaner because you aren't pushing into overdriving the sound, and equalizer circuitry is usually cleaner in attentuating sound than it is in boosting it.

If you want to use additive equalization, you will need to normalize your tracks down to allow for headroom. I generally normalize at 85%.

See ya
Steve

 

[russdog]

Quote:

Originally Posted by bigshot If you want to use additive equalization, you will need to normalize your tracks down to allow for headroom. I generally normalize at 85%.

For reasons I am not familar with, the default on MediaMonkey is 89%. I use it since I have no reason to think a different number is better. It works fine for me.

 

[grzzlo]

if subtractive is cleaner, why do you use additive?

 

[donunus]

the curve wont look exactly the same with those values above unless there is a gain bar like foobars I dont think. The difference will be in the edges of the octave(in the case of an octave eq). The additive style will have more frequencies than the edges of the wave while the subtractive will have vise versa. Hope this makes sense

 

[bigshot]

Quote:

Originally Posted by donunus the curve wont look exactly the same with those values above unless there is a gain bar like foobars I dont think. The difference will be in the edges of the octave(in the case of an octave eq). The additive style will have more frequencies than the edges of the wave while the subtractive will have vise versa. Hope this makes sense

I think I understand, but I'm not familiar with Foobar. In theory, additive should have the same shape above 0dB as subtractive has below the line. I don't know about digital equalizers, but analogue graphic equalizers all have different spill to the sides of the target frequency. My Rane is very accurate, with almost no spill. The DOD I had before overlapped a little bit. With a parametric equalizer you can control both the height and width of the adjustment, allowing for any sort of shape to the correction.

See ya
Steve

 

[bigshot]

Quote:

Originally Posted by grzzlo if subtractive is cleaner, why do you use additive?

If you've already adjusted most of the frequency range, and you run across a dip in one frequency, it's easier to just boost it a littler rather than pushing everything else down and having to rebalance at a lower level. I guess the answer to your question is "laziness".

See ya
Steve

 

[grzzlo]

Well, I guess I need to learn some more about equalizers, because I don't know what an octave equalizer is. But I guess you're saying that the curve will be different w/ negative eq, since rockbox has no eq gain. So I guess I should use replaygain...

 

[gloco]

Quote:

Originally Posted by grzzlo I don't know if this question has been asked before, and if it has then sorry. My rockboxed h140 clips when I use the eq, so I was wondering about any easy way to get around it. I think some people use replaygain, or whatever, to turn down the volume of all their tracks, so they can boost some bands, but I haven't used replaygain, so I don't know how much work that would be to update all my tags (I imagine not to hard). Anyway, here's my question: is there a difference btw setting the eq at +2,+2,0,-1,-1 vs. 0,0,-2,-3,-3 with the volume turned up to compensate? It seems like there shouldn't be, but does anyone know for sure? Is there any reason the two should sound different?

odd...what eq settings are you using? I play with the eq on my H140 all the time and have never heard any of my cans distort. What cans are you using?

 

[grzzlo]

Quote:

Originally Posted by gloco odd...what eq settings are you using? I play with the eq on my H140 all the time and have never heard any of my cans distort. What cans are you using?

er-6i's. I don't recall what exactly the eq settings were, but I was boosting the lowest 2 frequencies a little. I think at one point they were up as little as +2 and I still heard clipping. I just stopped using the eq for a while. I'd have to mess around a little to figure out what settings make it clip.

 

[Zenja]

The only thing you need to know is to never put things above 0 in an equalizer. The exceptions are if you can turn down the overall gain to balance it out, or if the song is recorded quiet so you won't notice anyway.

If there's a bar to let you lower all the levels down, then you can turn that down and raise the individual points. If not, then you need to only subtract from the equalizer.

Think about it this way. Let's say that there's a 100Hz sine wave being played, and its peaks reach all the way to the maximum allowable value (0dB). If you add just 1dB to 100Hz on the equalizer, you'll now chop off the peaks and make them flat because you can't go any higher than the 0dB maximum in the digital signal. This is called clipping.

 

[grzzlo]

Quote:

Originally Posted by Zenja The only thing you need to know is to never put things above 0 in an equalizer. The exceptions are if you can turn down the overall gain to balance it out, or if the song is recorded quiet so you won't notice anyway. If there's a bar to let you lower all the levels down, then you can turn that down and raise the individual points. If not, then you need to only subtract from the equalizer. Think about it this way. Let's say that there's a 100Hz sine wave being played, and its peaks reach all the way to the maximum allowable value (0dB). If you add just 1dB to 100Hz on the equalizer, you'll now chop off the peaks and make them flat because you can't go any higher than the 0dB maximum in the digital signal. This is called clipping.

ok, that makes sense, but it is different than what some other people here have said. This negative eq stuff may give you a different sound than positive eq with negative gain. Or so, I believe, the kind person above says.

Does anyone know how exactly the rockbox eq works in this regard?

 

[Febs]

Quote:

Originally Posted by grzzlo My rockboxed h140 clips when I use the eq, so I was wondering about any easy way to get around it.

Find the "Pre-cut" setting in the EQ and turn it down.

 

[donunus]

Quote:

Originally Posted by bigshot In theory, additive should have the same shape above 0dB as subtractive has below the line.

yup thats right meaning if you subtract the same amount of db from any given frequency as the additive, one would be a vertical mirror image of the others curve.
The difference is for what we are talking about that we dont know how tight the curves are between the centered frequencies of a particular eq so the edge of the curve might remain unchanged. hmm maybe i should draw something on a paint program. Or ya this should make sense right

 

[donunus]

check out the beautiful attachment. the top eq represents the same db added or subtracted for each frequency

the middle one represents the OPs eq graph in a perfect world. In this type of curve his question would have an answer no, there shouldn't be a difference between those values except for the level causing distortion if passed a clipping point.

The one at the bottom represents an eq where the two +2dbs and the two -1dbs are not so tight beside each other so with this type of eq the answer to the OPs question is yes the sound would be different with those 2 settings. The letter e is what I have represented as the edge of my octave. That is the part that makes most of the difference. As you see the 0,0,-2,-3,-3 curve is flat at 0,0 because of the edges of the equalized frequencies