Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: dB Definition as it Relates to Peak Levels (Read 5384 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

dB Definition as it Relates to Peak Levels

0 dB is considered out of range of human hearing from what I've read.  However, there seems to be a different use as it applies to audio ripping.

For example, let's say, I don't like when there is 5 seconds of "silence" after a track.  This silence could actually have some data if you look at a spectral analysis, but unless you have the biggest system in the world and the volume cranked, you'll likely hear nothing.  So I don't want this present.

To solve this, I'll use dBpoweramp to transcode from a FLAC to an mp3 and I'll use one of the DSP effects to say "I want everything below the x% peak to be pruned" and it gives a corresponding value in dB.  0% seems to be equal to -96.3 dB, and seeing as I just read the "quietest place on earth" is ~-9.3 dB, I don't understand this.  Couple this with the fact that dB level changes as I turn my volume switch, and I still don't get this, so it seems to be some sort of "native" dB level, or whatever.

Basically, my questions are, what is this dB level?  And according to this dB level, what is out of range of human perception?

dB Definition as it Relates to Peak Levels

Reply #1 –
dB is always a relative measurement.

In the case of dBpoweramp the numbers are relative to full scale equaling 0dB.

In the case of the "quietest place on earth" you still need to consider the reference used.  If it's the threshold of human hearing and this is given a value of 0 dB, that's fine.

dB Definition as it Relates to Peak Levels

Reply #3 –
Quote
0 dB is considered out of range of human hearing from what I've read. However, there seems to be a different use as it applies to audio ripping.
0dB SPL (Sound Pressure Level) is the acoustic threshold of hearing. So, dB SPL is (almost) always a positive number. 


Most (home) audio software is calibrated to 0dBFS (Full Scale).  In this case, 0dB is the maximum level, and all other dB readings are negative...    A 16-bit wave file can hold numbers between  -32,768 and +32,767....  If you have a wave with a peak of 32,767, that wave has a peak of 0dBFS.  If you need to store a bigger value, say 40,000, you need more bits....  If you try to go above 0dBFS, you get clipping (distorted flat-topped "waves".)

A VU meter on an analog tape recorder is also calibrated for 0dB maximum.  But in this case, 0dB is where the tape just-starts to distort, and most recording engineers will allow the level to go "into the red" (positive dB) occasionally.


But professionals frequently use a different (lower) scale, so that they can go above 0dB.  And of course, the pros are using more than 16-bits.  Take a look at The K-System


There is no way to "convert" between the electronic level and the acoustic level, since as you said, the acoustic level is controlled by the volume control.  There is an exception for movie theaters, where the digital/electronic dB level is calibrated to a specific acoustic level. 

Quote
0% seems to be equal to -96.3 dB, and seeing as I just read the "quietest place on earth" is ~-9.3 dB, I don't understand this.
  Actually, "0%" is minus infinity dB!  It is "digital silence".    The  smallest "wave" you can store with 16-bits (other than silence) is somewhere around -90 dB.

P.S.
Although a digital file (or digital stream) can have pure, absolute, digital-dead-silence, your digital-to-analog converter (and subsequent amplifiers)  will always generate some analog noise.

dB Definition as it Relates to Peak Levels

Reply #4 –
Quote from: DVDdoug on
0dB SPL (Sound Pressure Level) is the acoustic threshold of hearing. So, dB SPL is (almost) always a positive number.


To be very accurate: 0 dB SPL (which equals to 20 microPascals sound pressure level or 10^-12 W/m^2 sound intensity [that is, 10^-12 W acoustic power flows through a square meter of surface perpendicular to sound travel direction]) is an average human threshold of hearing at 1 kHz.
Human ear is most sensitive to frequencies around 3-4 kHz, and at these frequencies threshold is minus a few dB SPL.
Ceterum censeo, there should be an "%is_stop_after_current%".

dB Definition as it Relates to Peak Levels

Reply #5 –
The OP, and perhaps a few other people, have a bit of confusion about what dB means. Without some additional letters after the B, it does not mean anything in particular, it is just a calculation comparing two numbers. Those numbers could be measurements of sound pressure in the atmosphere, X-ray density in the explosion of different sized nuclear bombs, or anything else in the universe for which one wished to compare two measurements.

Using the dB calculation is sometimes a convenient way of doing the comparison, most often when it relates in some way to human perception. This is because human response to most stimuli is logarithmic, or thereabouts, rather than linear. For many other measurements, a dB calculation would never be considered because people could not relate to it very readily, but it would indicate the same mathmetical relationship as where it is commonly used.

The dBfs for a soundcard, or any digital recording, is the comparison between maximum signal (before clipping) and anything less. Maximum is relative to the converters, which can vary quite a bit. The unit of measurement could instead be a direct one,  in volts, because that is what the ADC is measuring, but that means of expressing it would not have a high utility value. A give number could differ considerably in meaning from one situation to another.

The dBfs measurement does not have any relationship to sound pressure. For one soundcard, 0dB could equate to 1 volt, for another to 2.4 volts. Neither relate to the sound pressure level of the audio that impinged on the microphone. Both give the identical digital audio, as long as the recording levels are properly adjusted.

dB Definition as it Relates to Peak Levels

Reply #6 –
Quote from: AndyH-ha on
The dBfs measurement does not have any relationship to sound pressure.
  I almost said the same thing, but actually, there is a "relationship".  There is a direct correlation.  If we reduce the SPL at the microphone by 3dB, the level recorded in the digital file is reduced by 3dB, and the sound coming out of the monitor speakers will also decrease by 3dB SPL.  This assumes an approximately linear system (no clipping, no compression or expansion, and nobody adjusting the volume/gain controls, etc.).
 

dB Definition as it Relates to Peak Levels

Reply #7 –
Let me try to restate that, in hopes that some will not confuse a correlation with a relationship. Pressure is force per unit area. Sound pressure, which is relevant only to the field of a sound wave, is the difference between the actual pressure at any point and the average pressure at that point. Sound pressure variations on the diaphragm of a microphone are converted to a directly varying electrical signal at the output of the microphone.

Assuming the sound pressure is within the linear portion of the microphone’s operating range, and that the entire system of microphone preamp, soundcard, power amplifier and speaker are operating linearly, the above post says that an increase of 1.4X, or a decrease of 0.71X, in the sound pressure at the microphone’s diaphragm will result in the same factor of increase (or decrease) in the sound pressure produced by the speaker.

There will not be any particular relationship between the actual sound pressure measurement in dB at the microphone diaphragm and the sound pressure measurement in dB in front of the speaker. Change an output volume control setting, or a digital amplification factor, or change out the soundcard for another that is equally linear but has a different threshold, or change the microphone preamp gain setting, and the speaker output will change corresponding to that change without any difference in sound pressure level occurring at the microphone diaphragm.