Audibility, Measurements & Test Tones


#1

As there is broader understanding here of how to interpret various audio measurements, and the real significance (or lack thereof) of each, it becomes more important to have a practical example of the actual audibility of a signal at a given level.

I say this because most modern components (DACs and amplifiers in especially), and certainly all decent modern test gear, can show differences that are well below the threshold of audibility even for a single tone - i.e. without the masking effects that are present in all actual recorded music.

What is masking? The simple explanation is that if you have two tones close together in frequency, then the louder one will generally render the quieter one inaudible. The difference in level for this to occur varies with the frequency distance between the two tones. And since music generally contains many parallel tones, masking is a big factor in terms of what is audible or not. In fact this is the primary principle employed by perceptual encoding schemes (lossy encoding) such as MP3, AAC or Vorbis - tones that are likely to be inaudible due to masking are omitted and not-encoded.

When dealing with a single, pure tone, say a typical 1 kHz sine wave, which is a mainstay of basic audio testing (a bit odd, since human hearing sensitivity is higher around 3500 Hz), no masking is in effect. So if you cannot hear THAT tone at a given level, measurements showing differences below that level are going to be inaudible, as well as being musically irrelevant.

While there is certainly technical merit to things like lower noise floors, or distortion products that are -120 dB (or more) from the fundamental, it’s useful to know how audible that really is before getting too hung up on simple, single, numbers (being able to interpret and overall set of graphed measurements is much more useful).

To that end, I’ve created a series of test files that’ll let you see how audible a single, pure, 1 kHz sine wave is at different signal levels. The basic idea here is to see just how far you have to crank up the volume on your rig to be able to hear them at all.

Each file is 60 seconds long. All are a pure, single, 1 kHz sine wave. Anything that is at -90 dB or higher is a 16-bit/44.1 kHz file. Below that they’re 24-bit/44.1 kHz, simply because a 16-bit file can only directly represent signals down to -96 dB (dithering and noise shaping can yield an “effective” or “perceived” -120 dB range).

How to get the Files

WARNING:

The risk of INSTANT and PERMANENT hearing damage from misusing these, or similar, test tones/files is ENTIRELY REAL.

If any other sounds play (system sounds, an email or text message arriving etc.) while you have a -90 dB (or lower) level tone playing at audible levels, you’re likely to experience pain and possibly instant, and permanent, hearing damage.

Additionally, some players/systems/DACs generate audible pops/clicks - often at 0 dB (i.e. max level) when starting/stopping replay, and if this occurs when you have the volume cranked you are similarly likely to incur permanent hearing damage or at least significant discomfort.

CAUTION:

While these files are all very low in level, you should always begin with the volume on your amplifier turned all the way down. Once the file is playing you can gradually increase the volume until you can hear the tone - you have a full minute to do it, so there’s no rush.

Make SURE you turn the volume all the way down again before the file reaches its end.

How to really get the Files

If you want access to these test files, shoot me a PM with your Google account address (e.g. a G-Mail address) and I will make them available to you.

In requesting access to these files you acknowledge that you understand the risks involved in using them and undertake said risks freely and purely at your own will.

Additional Disclaimers:

  • The risk of INSTANT and PERMANENT hearing damage from misusing these files is very real.
  • In the event I don’t trust you not to screw this up, for any reason, I reserve the right to ignore the request.
  • Brand new members may have to wait and demonstrate some common sense per the above.
  • If you’re the type that puts milk in your Earl Grey, these files are not for you.

#2

Awesome stuff! Noise floor becomes an important consideration w.r.t. dynamic range. Amir from Audio Science Review has a great article discussing noise floor.


#3

Thanks. I will check them out. I’ve used the interactive tool below for several years, but it’s not meant for comparisons:

https://www.audionotch.com/app/tune/


#4

Just read the article and it’s very informative. It really is fascinating just how complex sound is. Sound theory is heavy going but I am trying to read up a little to pick up some basic stuff.


#5

Yeah, there’s a lot to it, and there’s a lot that we don’t yet understand. I think what consistently trips me up the most is the impact of environmental and psychological factors. I attend a fitness class with music that plays way too loud (I actually wear ear protection for it). The other day, after taking such a class in the late afternoon, I spent some time EQ’ing my IEMs and ended up thinking “this sounds pretty great!”. The next morning, I listened to some music with the new settings and they sounded awful, with way too much bass. I think my ears were probably fatigued from the fitness class the night before and I overcompensated on the bass. Too many variables!


#6

Fatigue, and caffeine, and lack of sleep, and in my case diet too (e.g., tea, aged cheese, or grapes). This is a big reason why I cap my headphones in the $500-$1,000 range. I’m simply too inconsistent to accurately and consistently judge beyond that. Some stuff sounds fantastic one day, and mediocre the next.


#7

It’s something we probably don’t take into account but your experience of what happens with environmental factors shows us that a whole plethora of things need to be considered when looking for great audio. I also think that psychological factors play a significant part of audio too.