According to my second year Psychology class when we just listened to single tones - no.
The "Just Noticeable Difference" turned out to be 3dB; as I recall, although it's 40 something years ago, the test tones were played one at a time and we had to say when we could hear a difference. The tutor threw in some tones of the same volume to confuse us.
However, it wasn't music. These days my hearing stops at about 10kHz, but below that it's pretty good; when mucking about with crossover resistors, I can (I think) hear the difference between using a 1 ohm or 3 ohm resistor in the tweeter padding circuit.
Geoff
The "Just Noticeable Difference" turned out to be 3dB; as I recall, although it's 40 something years ago, the test tones were played one at a time and we had to say when we could hear a difference. The tutor threw in some tones of the same volume to confuse us.
However, it wasn't music. These days my hearing stops at about 10kHz, but below that it's pretty good; when mucking about with crossover resistors, I can (I think) hear the difference between using a 1 ohm or 3 ohm resistor in the tweeter padding circuit.
Geoff
With closed back headphones, yes I can perceive a 1.5db difference and less when I’m mixing. But in near/midfield space no unless there’s really good directivity from the system and it’s a distance from any significant boundaries. But I’m trained so YMMV.
And I’m not proud of those learned abilities….it tends to ruin the fun of just listening when your ear/brain is stuck on analyzing. If I owned a pizza place, I’d likely not be eating it anymore. Lol
And I’m not proud of those learned abilities….it tends to ruin the fun of just listening when your ear/brain is stuck on analyzing. If I owned a pizza place, I’d likely not be eating it anymore. Lol
Depends on where and the bandwidth.
I have my compression driver attenuation in 1db steps. Can hear that over a rather large 700Hz to 20Khz bandwidth.
Rob 🙂
I have my compression driver attenuation in 1db steps. Can hear that over a rather large 700Hz to 20Khz bandwidth.
Rob 🙂
That's a GREAT site btw! I could hear 1dB of the sine signals in my office over simple KEF LS50 without error, but that was already pretty hard.
But don't forget - that's the worst case signal, a naked sine. Nothing more narrow band as that. Wideband noise should be easier.
Also, with music there is so much masking effect going on from the brain.
So it will become harder and harder from there on.
Weird. I've run a pizza business for the past 8 years and i still eat some from time to time with great pleasure ( when they are good ones! Otherwise i prefer eat something else).
I don't get that technical/pleasure opposition. Sound like we have only one mindset availlable, which is obviously not the case. I struggled a lot with my close buddy with which i made music when i became pro engineer: they kept on telling me i wasn't able to listen to music for what it was ( i have a technical listening)... until i discovered they just had no words to define some things as i was able to do ( technical) and they were the one complaining about recording quality of old time recording performance of great piece of music... to the point to skip the art which i don't.
Iow for me it was just jealousy and assumptions. As i don't think you can jealous yourself and make self assumptions about yourself... i find it weird! 😉
@b_force, if i understood correctly what you said you think it's harder to hear difference with music source? It's not my case, it's easier for me but i think it is something learned by training. When mastering i often do shelf of +/- 0,xxdb and it wasn't obvious to hear the difference at first, now i know what to listen for it's easier to spot the difference.
Last edited:
That is also a mindset.
I used to be also hyper focused on everything many many years ago.
Like always critical listening to everything, comparing every little difference etc.
One day I just realized that is just not a healthy way of enjoying music.
Ever since, it's enjoying first, all other things second.
These days I just kinda switch it on or off if I want to.
I personally believe that it all has to do with what I just call "what if anxiety".
Which than can be categorized in sub-sets like "distortion anxiety", "directivity anxiety" etc etc.
People seem to forget that all these things are just tools and the result is nothing more than just the outcome of certain things within a certain time frame (and budget).
It's all about just letting things go at a certain stage.
I think to strive for perfection is a good thing and motivator, but it shouldn't take the upper hand.
Because I believe that at that point you're not being involved in music anymore, but are becoming just some kind of measuring tool.
I know a few people that are involved in audio, where I know for sure that they just don't enjoy music anymore.
For them it has become some kind of addiction to always be critical (quite literally).
Which I personally think is a very sad thing.
Because let's be honest, people who are totally unaware, probably enjoy their time more at that "crapping sounding concert" 😉
I don't think you can call this 'audiophile': Mayhem is sound engineer as i'am.
I won't talk in Mayhem's name but i can garantee you i don't feel i am an audiophile.
It's an issue about being focus on a task and the needs of it. What i find weird is the impossibility to switch on/off as B_force explained.
I won't talk in Mayhem's name but i can garantee you i don't feel i am an audiophile.
It's an issue about being focus on a task and the needs of it. What i find weird is the impossibility to switch on/off as B_force explained.
On average most people would hear a variation of +/- 3dB whether before or after 1980. Musicians and those in the audio industry inevitably become trained to pick out finer detail. It also depends whether you are actually listening for a variation. It is unlikely someone will walk into a room and say "Oh you've turned that speaker up by 1.5dB" that said when you have layers of tone as you do from a multi driver speaker, very small changes do make an audible difference. The same is true with an audio mix.
I did a study related to this while at Bose. The context was to decide what appropriate response tolerances should be for production. A reference was compared to a second system (actually one system and 2 EQs) where the second system had random response errors added in the form of a single1/3rd or 1 octave peak or dip. I thought long about what the test questions posed should be and came up with 3 answers:
1) I can hear no difference between A and B.
2) I can hear a difference but both sound acceptable.
3) I hear an unaceptable difference. (Presumably the B unit sounds bad)
This got to the question of whether a product should be shipable or not as there could be samples where a small difference was audible but deemed acceptable. (Thus preventing a war between engineering and production).
I was unable to take my data with me as I left the company but the results were about what you expect. The border for unacceptable response was in the 2, 3, 4 dB range for errors in the middle and low treble region. At the frequency extremes the acceptable error became pretty broad. Dips in bass response were just about inaudible and almost never objectionable. Octave wide peaks were distinctly worse than 1/3rd Octave peaks.
Because of the number of frequencies involved and the multiple steps of up and down error this took a lot of trials, over 10,000 individual sets, but at least it quantified subjective limits. Note that we aren't thinking in terms of JND (a just noticeable difference) and jokingly started to refer to the threshold of JTDB (just too damn bad!).
I still think this is the way to think about the subject, at least with regard to loudspeaker manufacturing.
David S.
1) I can hear no difference between A and B.
2) I can hear a difference but both sound acceptable.
3) I hear an unaceptable difference. (Presumably the B unit sounds bad)
This got to the question of whether a product should be shipable or not as there could be samples where a small difference was audible but deemed acceptable. (Thus preventing a war between engineering and production).
I was unable to take my data with me as I left the company but the results were about what you expect. The border for unacceptable response was in the 2, 3, 4 dB range for errors in the middle and low treble region. At the frequency extremes the acceptable error became pretty broad. Dips in bass response were just about inaudible and almost never objectionable. Octave wide peaks were distinctly worse than 1/3rd Octave peaks.
Because of the number of frequencies involved and the multiple steps of up and down error this took a lot of trials, over 10,000 individual sets, but at least it quantified subjective limits. Note that we aren't thinking in terms of JND (a just noticeable difference) and jokingly started to refer to the threshold of JTDB (just too damn bad!).
I still think this is the way to think about the subject, at least with regard to loudspeaker manufacturing.
David S.
I don't believe that being able to hear "finer detail" is a function of time and teaching, but just a physical property of the human ear (and brain). Obviously with a certain deviation.
And obviously you can improve (slightly) with training.
We are talking statistically here.
Or in other words, human beings were able to hear the same amount of detail 3000 years ago as we do know.
Probably even better, since we totally butcher our hearing these days with everything that is around us.
Incl listening to loudspeakers for too long and too loud. 😉
I was able to detect this myself using the stereo L-R balance control. 1.0dB shifts the center image ever slightly left or right and 0.5dB was not detectable at least on my system in a home setting.
You make a good point. In my early career I was testing and installing audio consoles I found I was able to detect small differences in both level and tone which saved a lot of time testing and ultimately led to finding an issue. Not sure if it was learning or I was just lucky. Many years on and I am still asked to "listen"when something is not quite right.
Very interesting.
I suspect some additional factors come in to play for someone who can design their speaker and room as a pair and who is able to tune them together?
My experience with small dB changes.....my 2c iow.
At the macro level, using full spectrum, my perceptual judgement is that+6dB is a doubling of sound volume.
I know that flies against all conventional literature, that says a +10dB is statistical perceptual doubling.
(Just trying to say maybe I'm more sensitive than stats)
I use real-time level controls on my 6-way system, like shown in this remote control. (HC buttons are house curve presets that set various fader level combos)
A change in any of the 6-ways sections of 1.5dB is very audible. Less so in the lower freq passbands that the higher, of course.
perhaps better said, a 1.5dB change in any of the passbands ranges from blatantly audible to mildly audible, with any particular music track.
With pink noise, a 1.5dB change is very audible on any of the 6 pass-bands altered individually.
These pass-bands are all only1-2.5 oct wide.
1.5db is pretty big deal, imho.
At the macro level, using full spectrum, my perceptual judgement is that+6dB is a doubling of sound volume.
I know that flies against all conventional literature, that says a +10dB is statistical perceptual doubling.
(Just trying to say maybe I'm more sensitive than stats)
I use real-time level controls on my 6-way system, like shown in this remote control. (HC buttons are house curve presets that set various fader level combos)
A change in any of the 6-ways sections of 1.5dB is very audible. Less so in the lower freq passbands that the higher, of course.
perhaps better said, a 1.5dB change in any of the passbands ranges from blatantly audible to mildly audible, with any particular music track.
With pink noise, a 1.5dB change is very audible on any of the 6 pass-bands altered individually.
These pass-bands are all only1-2.5 oct wide.
1.5db is pretty big deal, imho.
Perhaps it doesn't. 10dB is nominal at best in practice. The curves of equal loudness show that for us it varies with both frequency and level.
Context and specifics are very important here.
Can we hear the difference of 1.5 dB in broad band volume level? Definitely. All tests that I am aware of find the JND for volume change to be below 1.5 dB.
Can we hear 1.5dB as a difference in frequency response? This depends on the shape of the difference. Are we shelving up treble by 1.5 dB? Or a sub octave band in the lower bass? The first is probably audible, the second is questionable?
Are we listening to pink noise or busy music? Or sparse music? Each can give a different result and sparse music may make it hard to reliably hear a difference.
Are we doing a quick AB with the ability to alternate between two EQ cases (alternate over and over)? This will help you discern a 1.5dB difference. If on the other hand the changeover is delayed (long pause between the two instances) then you may not be able to discern the difference.
For example, your nefarious neighbor sneaks into your home one night and turns your tweeters down by 1.5 dB. Will you instantly notice it the next morning? You think so but really?
I did another study where a system had two equalizers in circuit. The first put random response errors into the chain and the second was adjusted by the person under test to get back to a desired response. Picture it as a speaker with 2 graphic equalizers. I use one to mess up the system response and you have to blindly adjust the second one to fix the mess that I made. By doing this a large number of times (across a large group of test subjects) we would define the perfect target response. This is similar to some work by Harman but extended to a 5 band equalizer case.
Results were terrible. A few people returned to a consistent desired target. Others were totally inconsistent. Even the consistent subjects did not converge on a common response target. Variations were huge. A 1.5dB error added to EQ 1 would not have been detected or corrected by this group.
This was very sobering to me as I strongly believed that a group should converge on a consistent system target curve. It might be that a better trained group was needed but these were all audio professionals, albeit with a range of ear training.
An interesting subject that needs more research.
Can we hear the difference of 1.5 dB in broad band volume level? Definitely. All tests that I am aware of find the JND for volume change to be below 1.5 dB.
Can we hear 1.5dB as a difference in frequency response? This depends on the shape of the difference. Are we shelving up treble by 1.5 dB? Or a sub octave band in the lower bass? The first is probably audible, the second is questionable?
Are we listening to pink noise or busy music? Or sparse music? Each can give a different result and sparse music may make it hard to reliably hear a difference.
Are we doing a quick AB with the ability to alternate between two EQ cases (alternate over and over)? This will help you discern a 1.5dB difference. If on the other hand the changeover is delayed (long pause between the two instances) then you may not be able to discern the difference.
For example, your nefarious neighbor sneaks into your home one night and turns your tweeters down by 1.5 dB. Will you instantly notice it the next morning? You think so but really?
I did another study where a system had two equalizers in circuit. The first put random response errors into the chain and the second was adjusted by the person under test to get back to a desired response. Picture it as a speaker with 2 graphic equalizers. I use one to mess up the system response and you have to blindly adjust the second one to fix the mess that I made. By doing this a large number of times (across a large group of test subjects) we would define the perfect target response. This is similar to some work by Harman but extended to a 5 band equalizer case.
Results were terrible. A few people returned to a consistent desired target. Others were totally inconsistent. Even the consistent subjects did not converge on a common response target. Variations were huge. A 1.5dB error added to EQ 1 would not have been detected or corrected by this group.
This was very sobering to me as I strongly believed that a group should converge on a consistent system target curve. It might be that a better trained group was needed but these were all audio professionals, albeit with a range of ear training.
An interesting subject that needs more research.
I mainly did live engineering and am trained to hear "problems". An upcomming feedback but also krackling, noises etc.
When these occur I still startle up - and most of the time it's just a cooler or door or a noise baked in the recording which I can hear in my silent room and they didn't care about in the production. (I don't remember the song but there was a door opening noise exactly as it sounds with my 2nd outer door. That was puzzling as nobobody was entering. Played that 3 times to be sure 😀 ).
Live recordings ... I still wince when something is at the border of feedback and all the others look puzzled 🤓 And when you have actual feedback in productions like MTV unplugged it still hurts my soul a little.
There are a lot of recording mistakes you hear with the proper setup and knowledge but I learned to accept these - and sometimes these are also artistic choices (like Tori Amos has a lot of distortion when her voice get's loud. As I found this on different albums it's probably on purpose to give these passages extra "drive"). I also use plenty of distortion on bass and some drumset signals - that's the genre and helps with a loud and distinctive sound.
There is always a reason why something sounds not as we like it (no time for sound check or wrong setup or horrible location or ...). When you are not part of the production you don't know what happened.
Yes artistic choices can sometimes bother people.
Like you i accept them but it's easier as i've been involved into production stage and know time spent on mixes ( i whitnessed a guy spend three days on a snare which didn't glue in a mix! Extreme for something nobody'll care anyway...).
Distortion on voice can be pleasing. And not only on industrial tracks! 🙂 I've got Otis Redding's recording where we hear the preamps distorting when he pushed level, i find it suit well the style and his voice and it's the 'sound' of that era. I think i would find it boring if someone cleaned this parts now.
'When you are not part of the production you don't know what happened': so true! I often smile when i see musicians disgusted because there was an issue during a set, except for very obvious mistake ( track structure related most often) most of the time it goes unnoticed or best a cocky move or experiment! Why when it happen you should always keep an attitude letting other think it was under control! Lol.
Like you i accept them but it's easier as i've been involved into production stage and know time spent on mixes ( i whitnessed a guy spend three days on a snare which didn't glue in a mix! Extreme for something nobody'll care anyway...).
Distortion on voice can be pleasing. And not only on industrial tracks! 🙂 I've got Otis Redding's recording where we hear the preamps distorting when he pushed level, i find it suit well the style and his voice and it's the 'sound' of that era. I think i would find it boring if someone cleaned this parts now.
'When you are not part of the production you don't know what happened': so true! I often smile when i see musicians disgusted because there was an issue during a set, except for very obvious mistake ( track structure related most often) most of the time it goes unnoticed or best a cocky move or experiment! Why when it happen you should always keep an attitude letting other think it was under control! Lol.