Yes, I can hear smaller differences than 0.2dB. I don't often talk about it because it's situationally dependent and subjective. It's also frequency dependent.. I can't reasonably hear that small a difference down into the bass range.
I'm sure we have experienced the converse on the first day after hooking things up but before processing measurements. Guessing at a crossover so you have something to listen to while you work. Such big changes can be made with such little indication of a change.
I'm sure we have experienced the converse on the first day after hooking things up but before processing measurements. Guessing at a crossover so you have something to listen to while you work. Such big changes can be made with such little indication of a change.
Last edited:
A transparent EQ like CamillaDSP, your system and a listening session with some random trials with PEQs of Q=1 can be quite informative...
//
//
I think instead of making any black and white style absolute statements regarding what is or isn't possible, its better to look at how different we subjectively perceive tonality and balance in music. The psychology of hearing isn't as simple as people portray. It depends on mood, health, experience, personal taste, conditioning and even moreso if one of our other senses is compromised, like with blind people. Its amazing what blind people can hear and perceive. Of course we're talking about spacial cues and deciphering small sounds in very cluttered and noisy environments. Transients are definitely easier to differentiate than sustained sounds.
The overall playback volume has alot of impact on the FR balance as well. Most recordings are mixed on the louder side, which makes a recording played back at lower volume sound dull and undefined. That has alot to do with the (fletcher munson) loudness curve built into our hearing, partially because of the non linear LF sensitivity, making our ears perceive different FR balance at varying volume levels.
From a comparison POV, your ears can perceive very small changes in FR, often down to 0.1dB in the delicate 3 - 4k range. A wide range only boosted a few tenths in treble is very noticeable, but very narrow peaks are harder to identify with average music unless the amplitude is much higher. Speakers with lots of smaller peaks tend to sound more lively and interesting, whereas a clean CSD with smooth FR usually means it will sound boring and clinical.
Even and odd order HD mixed into the equation also messes with your perception and is much more complicated to quantify. Some people like Gedlee try to make sense of how specific harmonics of the entire distortion spectrum and various types of reflected artifacts affect our perception of a speaker's SQ. If a person doesn't fall into the norm of average, it doesn't mean they're wrong or can't perceive relatively subtle nuances. Again, not everything we hear can be measured in detail by common instrumentation despite statements made to the contrary by so called experts.
Things are much different in the lower frequencies due to decreased sensitivity of our ears and a much lower available dynamic range. Our physiology isn't designed to hear LF that well. The ratio of wavelength to physical spacing of our ears is too great to perceive the direction of LF sound. Thats why the mono bass mix used on most vinyl works well and stereo recorded bass doesnt make much sense in 2 channel recordings.
IMO the biggest problem ts putting what you're hearing into words. All the colorful adjectives used by reviewers can be very confusing and non definitive. Terms like "edgy", dark" or "boomy" can be understood, but terms like PRAT make zero sense.
The other confusing issue is predicting how a speaker will sound in a typical room. A speaker that measures flat in anechoic environments can sound bright in one room and dark in another. Placement matters alot too, as well as how loud we listen and the preferred type of music. The smallest changes in peripheral gear can be magnified by a speaker with uneven, peaky and/or bright FR, especially when its off axis performance is irregular.
The overall playback volume has alot of impact on the FR balance as well. Most recordings are mixed on the louder side, which makes a recording played back at lower volume sound dull and undefined. That has alot to do with the (fletcher munson) loudness curve built into our hearing, partially because of the non linear LF sensitivity, making our ears perceive different FR balance at varying volume levels.
From a comparison POV, your ears can perceive very small changes in FR, often down to 0.1dB in the delicate 3 - 4k range. A wide range only boosted a few tenths in treble is very noticeable, but very narrow peaks are harder to identify with average music unless the amplitude is much higher. Speakers with lots of smaller peaks tend to sound more lively and interesting, whereas a clean CSD with smooth FR usually means it will sound boring and clinical.
Even and odd order HD mixed into the equation also messes with your perception and is much more complicated to quantify. Some people like Gedlee try to make sense of how specific harmonics of the entire distortion spectrum and various types of reflected artifacts affect our perception of a speaker's SQ. If a person doesn't fall into the norm of average, it doesn't mean they're wrong or can't perceive relatively subtle nuances. Again, not everything we hear can be measured in detail by common instrumentation despite statements made to the contrary by so called experts.
Things are much different in the lower frequencies due to decreased sensitivity of our ears and a much lower available dynamic range. Our physiology isn't designed to hear LF that well. The ratio of wavelength to physical spacing of our ears is too great to perceive the direction of LF sound. Thats why the mono bass mix used on most vinyl works well and stereo recorded bass doesnt make much sense in 2 channel recordings.
IMO the biggest problem ts putting what you're hearing into words. All the colorful adjectives used by reviewers can be very confusing and non definitive. Terms like "edgy", dark" or "boomy" can be understood, but terms like PRAT make zero sense.
The other confusing issue is predicting how a speaker will sound in a typical room. A speaker that measures flat in anechoic environments can sound bright in one room and dark in another. Placement matters alot too, as well as how loud we listen and the preferred type of music. The smallest changes in peripheral gear can be magnified by a speaker with uneven, peaky and/or bright FR, especially when its off axis performance is irregular.
Last edited:
With Behringer DEQ2496 I did equalization with MiC 8000 and 0,5 is possible with this EQ and sufficient.
With MiniDsp 2x4 it was possible to manipulate by 0.1 db and its subtle but you can hear it if you adjust for one or two octaves - for example putting down the middle tone region.
But for building loudspeakers its the composition of energy in the room between bass mids and highs and off axis radiated energy in general so already a correct adjustment up to 0,5 db is very very good.
Also its not good to correct for every little deviation as the use of dsp can bring ringing if you try to correct for narrow bands with high Q.
Its for this that every filter I applied to loudspeakers with DSP due to first measurements had to undergo listening test for:
1. is the filter less steep adjusted already good enough?
2. can it be corrected with less decibels or how much is really needed to make it sound right?
so I checked every parametric EQ if it really helps the sound - its not important how nice the response looks. Usually more linear sounds better but in the mid and highs where radiation off axis play an imortant role you have to listen to the speaker in order to do it right.
Therefore I am very sceptical with Digi EQs which work automatically and even more criticial with automatic room correction as here usually excessive DSP is used and you can introduce ringing. You worsen the sound by doing too much.
Mr Kirchner from Kirchner Audio, germany who sells ATB PC measurement system offered a PDF explaining this all.
https://kirchner-elektronik.de/wp-content/uploads/2019/04/15508585-Guete.pdf
1.
Last edited:
Its true what you say.
For knowing if you finished your loudspeaker DSP correction to the point or not I recorded instruments myself and played then these recordings while playing the instruments life at the same time in the listening room.
Then you experience:
Some days you think your speaker does not sound right but its subjective - you have a bad day.
Then I play my recording and the instruments - everything sounds correct - so its my subjectivity, not the loudspeaker.
This helped me a lot to know - when did a maker of hifi gear finish his job?
And it prevents to change an already perfect system and make it worse again.
I learned that already cheap recording equipment like a zoom h1 makes recordings that untrained listeners would have problems to hear differences between recording and instrument. This is due to technical evolution offering good quality for little money in our days.
However I let a friend make the "dry" sounding recordings of single simple played instruments outside in the free field and 50cm close to the instrument with expensive microphones for several thousand euros.
However I let a friend make the "dry" sounding recordings of single simple played instruments outside in the free field and 50cm close to the instrument with expensive microphones for several thousand euros.
Super link! And now ... what about an open, spontaneous, and rather practical than scientific survey here amongst the population of the posters and also for all the until-now silent readers? Only condition to join in is to remain honest: As described above, several factors may influence on the sensitivity of SPL changes. In case of an unexpected (and maybe even offending) come-out, then please publish your results anyway an try to comment on this surprize. And please also describe your test-setup briefly when posting your personal test outcome.
As for me ...
Setup I:
Listening at the max. possible level from the inbuilt loudspeakers of a vintage laptop (Thinkpad T500)
Setup II:
Same basic setup, but listening with AKG K702 headphones directly plugged into the Thinkpad's connector.
Result (both Setups I and II):
1.0dB: 99.9% confidence == YES
0.5dB: 55% confidence == NO
I was puzzled how sharp this threshold limit was between the two deltas.
So for my very private DIY audio gear needs, I would pragmatically implement steps of 0.5dB for a (digital) volume control.
Quite another and a bit off-topic, but possibly related subject is the the inter-stereo (R<->L) channels SPL balance. This balance does influence on the location of the stereo phantom image: A 1.2dB interchannel intensity difference is said to shift the stereo phantom by some 5° to the louder side. In other terms, within a standard 60° aequiangled, 2m sided stereo triangle, this 1.2dB SPL delta makes the stereo phantom shift away by some 8.5cm from the basis midpoint. It would be interesting to test also the individual thresholds of this specific phenomen when talking about SPL differences.
Last edited:
^ Yes it's interesting link but... with a sine. It changes things wrt a 'normal' musical signal imo. Worth doing it though. With headphone i was ok to 0,2db variation.
0,1... i didn't pass.
About stereo you omit one thing: we use delta level AND delta phase to define localisation so it's not as easy as a simple distance offset. That said, we are really good at localisation on lateral scene. At least i'm constantly amazed by how acurate we are.
Wesayso and i are working on a series of files to test our setup and acuity to it in relation to Griesinger researchs. I'm finalizing it atm, files should be availlable at the end of this week if not postponed as it have been during this summer. And it's exactly what you said: based on delta level only ( it's done using a pan pot function).
Wesayso gave link to preliminary set and explanation about why and what here:
Post in thread 'The making of: The Two Towers (a 25 driver Full Range line array)' https://www.diyaudio.com/community/...ver-full-range-line-array.242171/post-7397190
0,1... i didn't pass.
About stereo you omit one thing: we use delta level AND delta phase to define localisation so it's not as easy as a simple distance offset. That said, we are really good at localisation on lateral scene. At least i'm constantly amazed by how acurate we are.
Wesayso and i are working on a series of files to test our setup and acuity to it in relation to Griesinger researchs. I'm finalizing it atm, files should be availlable at the end of this week if not postponed as it have been during this summer. And it's exactly what you said: based on delta level only ( it's done using a pan pot function).
Wesayso gave link to preliminary set and explanation about why and what here:
Post in thread 'The making of: The Two Towers (a 25 driver Full Range line array)' https://www.diyaudio.com/community/...ver-full-range-line-array.242171/post-7397190
Depends. Mostly yes.
In a mix if you are changing only small range of freqs by +-1.5db, may be not.
In a mix if you are changing only small range of freqs by +-1.5db, may be not.
I know from adjusting freqs with DSP that 2db too much in the highs will drive you on the long run mad.
In the bass after finding the linear EQ position +2db makes people say that the speaker sounds fat or mellow. - 2db makes it sound fast and dry in tendency. For a given loudspeaker position.
Usually I let the mids after tuning all to the point 0.5 to max 1db louder as I like a lively middle tone.
In the bass after finding the linear EQ position +2db makes people say that the speaker sounds fat or mellow. - 2db makes it sound fast and dry in tendency. For a given loudspeaker position.
Usually I let the mids after tuning all to the point 0.5 to max 1db louder as I like a lively middle tone.
Agree - really quite problematic but perhaps a start... Sines are not music for sure.
//
im one of them, i can definitely hear 0.2db changes in eq, specially wider band eq, tho i took some training/messing around with eq to actually realize that 0,2db can make a difference
i also read that +/-3db was said to the audible limit for "non-trained" listeners but trained listeners can far go beyond that, some might even hear 0,1db but i think the last time i did this test i also checked out at 0,2db https://www.audiocheck.net/blindtests_level.php?lvl=6
100% agree
as eq im actually appliying just house/target curves because it just sounds smoother then equalizing just certain bands
if you are trained to hear these kind of changes 1db feels "plenty", too much sometimes, i found that i usually tweak eq`s in 0,5db steps and after i find the right spot i fine-tweak in 0,1db steps
funny story is that the changes cables can make can be compensated for with lets say a +0,3db high treble boost, its not like the change is completely masked then (so it still makes a difference) but its way closer to before that way
Yes, you can hear that.
In mastering you somtimes use EQs in the 0,5 to 1dB area which changes more the "feel" as is clearly audible as big change. But it is audible in 1:1 comparison.
For broadband changes surprisingly low amounts are audible. Narrow band surprisingly big changes are sometimes unnoticed.
A untreated room changes low frequencies in the area of +-15dB. And it changes every step you go through your room. So i nreal living room situations you have plenty of problems to solve before talking about these nuances 😉