Can we hear 1.5dB difference?

Back in the 1980s, many commercial speakers had attenuators for tweeters; some also had attenuators for midranges. There were many attenuation levels. One of the interesting points is +/- 1.5 dB. I’ve read somewhere that human ears begin to hear a difference in sound level at about +/- 3 dB. Yet, why do many manufacturers put +/- 1.5 dB on their attenuators? Or, actually, the +/-3 dB was discovered after the 1980s era? Anyone who has experience with the 80s speakers or has knowledge about this, could you please confirm the fact about +/- 1.5 dB or +/- 3 dB?
 
Member
Joined 2009
Paid Member
I’ve read somewhere that human ears begin to hear a difference in sound level at about +/- 3 dB. Yet, why do many manufacturers put +/- 1.5 dB on their attenuators?

There is two things mixed: absolute level ( spl) and modification of tonal behavior.

If you talk about absolute level ( and without reference you can quickly switch for comparison purpose) then it can be difficult to hear 3db difference for untrained listeners.

If trained (and with a reference) you can hear way less variation though. In fact in double blind test comparison of same sample you'll favour the louder one by 0,1db but without being able to tell why. It's just it's more a feeling than something you'll choose on something you thought about.
One of the issue which make audio comparison difficults to do. If interested i could give audio examples ( but headphone listening will be mandatory).

Tonal behavior: we are more sensible to modification of tonal range. See Fletcher and Munson curves and notice it is related to 'absolute' Spl level too.
1 db variation with a high shelf eq centered around 3khz will be very audible as it is centered on a peak of sensitivity of our hearing system.
The same level variation over a much smaller area of the audio band ( depend of frequency too) might be unoticable.

That said with low Q ( wide range affected) you are able to feel very low variation in level, 0,5db with low Q shelf can be easily identified if you are trained or sensitive.
 
  • Like
Reactions: 6 users
1694252842617.png
 
  • Thank You
  • Like
Reactions: 1 users
+/- 0.2 can be possible in the more sensitive parts of the spectrum, when applied to a balanced speaker. Unless I'm thinking of someone else, Siegfried Linkwitz has spoken about smaller variations than this.
That is true. I've noticed interesting perceived effect in practice: system balance might feel off, say there is feeling that bass is too low. Now boost bass for one decibel, and now it feels too loud. What the heck?

Most likely there is peak in frequency response that makes for example kick and lowest lows feel lacking, but if you boost bass the peak just makes itself more offensive wanting to turn it back down.

The peak in frequency response throws the perception off, untill you notice it and fix it, and now suddenly its quite easy to find suitable level for bass. Lowest lows and kick both come through no matter what the level is, and one just balances it to liking at it always sounds fine, here a 0.5db difference can be evaluated as you hear it clearly without anything grabbing attention elsewhere. At least this is my current practical thought on the subject.

ps.
It's fun to notice something sounds weird but cannot quite understand, or hear, what it actually is that makes the feeling. There is no conscious detection of a problem yet, only subconscious one, understanding is lacking. One can try to find the offender by looking for some secondary perspective on the sound, like a visual graph, or mute button or EQ knob, or changing listening position, taking the system outside, what ever it is. Now one can have two audible perspectives of the system, where an issue is audible or not, and can reason what made it, hopefully improving sound. Best part is still that you learned to listen, and hear it, got understanding for the perception.
 
  • Like
Reactions: 1 users
1.5dB where and what?

  • As a small (EQ) band peak?
  • As a small (EQ) band dip?
  • A larger frequency range?
  • full-range?
  • full-range relative compared to something else?
  • Just an acoustic noise source?
  • Uncorrelated sound/noise sources?

I think we can make this list even longer.

For some of those on this list, the answer is yes.
For some others, the answer is no.
For some the answer is, it depends.
 
  • Like
Reactions: 2 users
I've noticed interesting perceived effect in practice: system balance might feel off, say there is feeling that bass is too low. Now boost bass for one decibel, and now it feels too loud. What the heck?
So you automatically assume that the sensors you're using (ears and hearing system) plus the processing of this sound (brain and mood) are constant?

I very much doubt that.
 
1.5dB is the attenuation of the tweeter.
Right, because your initial first post is pretty vague in the sense that it's unclear if you mean this as a general idea, or very specific.

For example, the ±3dB comes from general acoustics.
Which is very true, especially when acoustic sources are not correlated (random noise), it becomes pretty hard for the average human being to notice any difference.

However, for small frequency bands, especially peaks around the 700-3000Hz region, this is a very different story.
Even more so for a trained ear.
For attenuating a tweeter just about 1.5dB will give a every so slight perception of balance of the whole sound stage.

What I am still trying to say (also with the list of my previous post), is that it depends.
Even sometimes for a tweeter it depends on its range, directivity of the system etc etc.

But in general (again), yes 1.5dB can give a different perception of the whole system.
Also that can differ from surroundings and average sound pressure levels (loudness).
 
So you automatically assume that the sensors you're using (ears and hearing system) plus the processing of this sound (brain and mood) are constant?

I very much doubt that.
No, i try to find another perspective to verify what is it. If you fidn a way to toggle the perception on or off you can reason what it is, or xonduct further testing. If not sure, leave it be.
 
I'd generalize and say you can't generalize here. Very person dependent. Some have perfect pitch, not me. Some are more attuned to slight changes. A friend's husband years ago could hear a slight change in speaker position. She would clean and wouldn't get the speakers in exactly the right spot. He was OCD for sure. Anyway she blindfolded him because she did not believe he could tell. Came down to he could tell if they were moved more than an inch. She was in disbelief. Way back in college we did a rough experiment to tell how good of a spectrum analyzer humans were. So we put a 1 KHz tone in pink noise. My lab partners said they could tell there was a tone down to -20db snr. I could tell down to -40. They did not believe me, so they turned me around and would randomly inject the tone and I'd raise my hand if I heard it. They thought it was trick I could tell. Different people, different sensor qualities.
 
@b_force beat me to it. It depends on the bandwidth of the SPL deviation.

I use DSP, and so I can make small adjustments in all sorts of ways. IMO if I make even a 0.25dB adjustment of, say, the tweeter level I can tell that the tonal balance of the presentation has shifted slightly. The SPL change in this case is for all frequencies above 2kHz. OTOH, sometimes I do not bother to EQ a dip or even a peak if it is narrow. The ear is just not all that sensitive to higher-Q/low-bandwidth deviations. Of course this is just my opinion and I am talking about doing sighted tests where I know I have made the change.
 
  • Like
Reactions: 1 user
[...] Way back in college we did a rough experiment to tell how good of a spectrum analyzer humans were. So we put a 1 KHz tone in pink noise. My lab partners said they could tell there was a tone down to -20db snr. I could tell down to -40. They did not believe me, so they turned me around and would randomly inject the tone and I'd raise my hand if I heard it. They thought it was trick I could tell. Different people, different sensor qualities.

Yup. I can tell phase rotation between 100 and 250 Hz no problem, untested outside those frequencies. Others mostly don't hear it, although a few do, mostly people who've listened to instruments acoustically.
 
+/- 0.2 can be possible in the more sensitive parts of the spectrum, when applied to a balanced speaker. Unless I'm thinking of someone else, Siegfried Linkwitz has spoken about smaller variations than this.

No prob for a skilled listener on a good system.

Reminds me of a situation back in the 90s when I was still working on my 4-ways. A friend stood in the door asking me if I would join him for a beer on the local pub. Not an audiophile this one. Then he asked me "hey, what have you done to your system now, it sounds more open?" I had just gone from bi- to tri-wiring, and he could easily hear the difference from the door on the sidewall. Wow.