Stupid question regarding decibels

[04blackmaxx]

Ill keep this short and to the point. Why do amps show the decibel level as negative when there is sound being made at said level? Wouldn't the db be positive whenever there is sound being made?

 

[JP]

Stands for dB of attenuation.

 

[robgmn]

It's showing the level that's below the reference level.
For example: at "zero" on the volume display, you get 75 dB at the point of measurement when playing pink noise.
When you go to "-2", it should be 2dB below that reference level.

It's rarely exact, though and doesn;t have much use unless you need to be able to easily set your volume to the reference level without having to memorize what number that is on the volume control.

 

[so_ein_pech]

The decibel scale is by definition a relative measurement between a reference level an whatever is being measured. In terms of power it is defined as P_dB = 10*log(P/P_ref). In terms of pressure it is defined as spl = 20*log(p/p_ref).

This is confusing because sound intensity is often measured in dB and treated as an absolute scale. (I.e. threshold of pain = 130dB, normal conversation = 60dB etc, etc.) This can only make sense if everyone has agreed on a reference level, and indeed they have. The threshold of human hearing is defined as the 0 point of this scale such that 0dB = 0.00002 Pa or 10^-12 W/m^2.

Decibels are also used to define the voltage gain of an amplifier stage where gain(dB) = 20*log(V_out/V_in).

The nice thing about this convention is that if you reduce the gain by, say, 3dB, you also reduce the sound pressure level by 3dB. From an engineer's perspective it makes perfect sense to specify the volume (gain) control in terms of dB. For everyone else it is confusing as hell.

 

[whoaru99]

Yeah, it's quite common in HT to indicate volume relative to the calibrated reference volume.

Typically, after you have done the calibration, 0dB volume setting should = 105dB SPL peaks for main and 115dB SPL peaks for LFE. So, if I have the volume control set at -15dB I'll have 90dB peak levels on the main content and 100dB peaks on the LFE content.

It's a lot more nebulous in non-HT setup like normal 2-ch system because 0dB setting on the volume control isn't necessarily calibrated to anything specific. Reducing the setting by 3dB should made a 3dB difference in SPL but it's only relative to itself, not relative to a standard point per se, like in HT.

 

[04blackmaxx]

The decibel scale is by definition a relative measurement between a reference level an whatever is being measured. In terms of power it is defined as P_dB = 10*log(P/P_ref). In terms of pressure it is defined as spl = 20*log(p/p_ref).

This is confusing because sound intensity is often measured in dB and treated as an absolute scale. (I.e. threshold of pain = 130dB, normal conversation = 60dB etc, etc.) This can only make sense if everyone has agreed on a reference level, and indeed they have. The threshold of human hearing is defined as the 0 point of this scale such that 0dB = 0.00002 Pa or 10^-12 W/m^2.

Decibels are also used to define the voltage gain of an amplifier stage where gain(dB) = 20*log(V_out/V_in).

The nice thing about this convention is that if you reduce the gain by, say, 3dB, you also reduce the sound pressure level by 3dB. From an engineer's perspective it makes perfect sense to specify the volume (gain) control in terms of dB. For everyone else it is confusing as hell.

Thank you, wow. So....if I set the volume to make sound around the level of human speech at 60db, and I have a 300 watts per channel amplifier, what will the meters be showing as dbs?

 

[gadget73]

The meters also can be set to some reference level of power, and the - numbers are how many db below that level you are. My Phase Linear for example has both a DB and a watt scale. 350 watts is 0 DB. -10db is 35 watts, -20db is 3.5 watts. I wouldn't begin to tell you it scales accurately, nor that its actually accurate but thats what the setup process is. It is accurate at 60 hz into a dummy load. Outside of those particular conditions, it's anyone's guess. Meters are more of a "hey, this looks cool" function than a really accurate measuring tool. Impedance swings in the speakers and general inaccuracies and inability to respond to musical peaks limits their usefulness.

 

[whoaru99]

The actual dB ratio would depend on a number of factors including (but not necessarily limited to) speaker sensitivity, listening distance from speakers, and position of speakers relative to boundary.

With a few assumptions-

Speakers = 87dB/W/m
Distance =10 ft
2 speakers positioned near a wall

I come up with -47.8dB (relative to 300W) for 60dB peak SPL.

 

[loonze]

YA KNOW....I also had often wondered that.....just never was motivated enough to find out.

 

[fred soop]

.....
Meters are more of a "hey, this looks cool" function than a really accurate measuring tool. Impedance swings in the speakers and general inaccuracies and inability to respond to musical peaks limits their usefulness.

There are a couple of situations where they are actually useful. Recording tapes or burning CDs. You want the highest level possible without going into distortion.

 

[gadget73]

for recording levels sure, but watt meters on amps I question the real value in. Also, generally speaking with recording level meters, they are running off a constant load circuit so they will tend to not have such variable accuracy with frequency as something wired to a pair of speakers will.

 

[so_ein_pech]

for recording levels sure, but watt meters on amps I question the real value in. Also, generally speaking with recording level meters, they are running off a constant load circuit so they will tend to not have such variable accuracy with frequency as something wired to a pair of speakers will.

Beyond looking cool, power meters give you a sense of whether or not you are redlining your amp. Also they are good indication for if you are getting subsonic feedback on a record.

But, yeah, they are not very accurate.