#Dbfs decibels full scale pro

(Most converters are calibrated so that -18dBFS digital = +4dBu (standard line level for pro equipment for a lot irrelevant historical reasons).

Where digital audio converters get really confusing is that you're arbitrarily assigning a voltage to a number. I'd feel better saying 0 is the smallest value you can measure accurately. The lowest value (0 bytes) would be when humans can't hear anything below it, so don't bother storing the data and set it to zero. Then based on the precision of the data, we calculate a negative scale. Given that, at some point in the past it was decided that if we flip that around we can just always refer to 0dB as the ceiling for max loudness/power for whatever size of data we have for a given audio sample. To rephrase yet another way, since audio can be digitally stored in any size (8 bit, 16 bit, 16384 bit, etc) there's no way to universally have a max data size for a segment of audio information. Or, we could also say the measurement system is -128dB to 0dB (which is how I'm understanding this to work.) So with those two values, we could make an arbitrary measurement system that goes from 0dB to 128dB. Suppose the largest dBFS value that can be stored is 128 bytes.

#Dbfs decibels full scale software

(I'm a software engineer and this makes the most sense to me.) Let me try to restate your example with one of my own to see if am getting the underlying concept. I think this is the closest statement to what I'm trying to understand. Twice the quantity would be +3 dB, 10 times the quantity would be 10dB, etc, 1/10th the quantity would be -10 dB. Then anything that's above or below that will be a positive or negative number. The underlying concept is that you just arbitrarily pick 0 dB to be some useful quantity. You never see positive values in digital audio because you can't go higher than the reference value. That maximum is chosen to be 0dBFS, and you measure how far below that is, using negative dB. Digital audio is stored as numbers, and there is an absolute maximum number that a computer can store. Lastly, digital audio is usually measured in dBFS (decibels below full scale). You usually don't see negative dB SPL because you can't hear it you usually use dB SPL to describe how loud something is relative to the quietest sound you can hear. Sound pressure (what you actually hear) is usually dB SPL, with a reference of 0 being the quietest noise that the average human can hear. More than that would be some positive dB, less than that would be a negative dB value, because it's relative. 0 dBu, for example, is ~0.775 millivolts RMS.

The reference level chosen usually depends on what makes sense for the particular application.įor example, output levels of analog gear are measured relative to a standardized quantity. If our average level is -18 dBFS, that means our average level is 18dB less than full scale, and we can have peaks that go 18dB higher than average before getting in trouble.ĭecibels are a relative unit it's always a ratio between two absolute quantities, so you need to know what kind of decibel you're talking about. So levels for digital audio tend to be in relation to 'full scale' and called dBFS.

The point where we 'run out of bits' at the top end and will distort if the signal goes any louder is called 'full scale' –– all the lights on our equipment's level meter are lit. In the digital age, the loudest thing we can get without digital distortion is a pretty important thing to know - we want to maximize our signal above the quantization noise, but not distort. so 85dB SPL is 85dB louder than the threshold of hearing. 'SPL' means 'Sound Pressure Level', and convention is that we measure SPL relative to the threshold of hearing. Hence all those letter abbreviations after it that we find. The something else is often not mentioned explicitly, but is always there. DB is always a comparison of something with something else.