Thanks Peter. I guess I meant the guitar rifts in the opening of the Heart song. I found that the most accurate representation of the guitar strings was from the red speaker. To me it was natural, almost a lean sound compared with the other speakers which more heavy, bloated, mushy. Hard to find the words to express the sound.
Anyways, need more votes from folks!!!!!
Myles
Anyways, need more votes from folks!!!!!
Myles
Which red are we talking about? The one on top or bottom...
In my screen both has same colour
Peter
Based on overall number of views for the clip downloads about 2-3 votes are still out, barring last day holdouts who have not downloaded clips. I do not believe the established vote pattern will change much.
Driver B has no votes. Why is this?
Based on Round-5-Meas-HR.pdf with measurements offset by 10dB intervals for ease of viewing, the assumption can be made that if traces are moved by offsets in multiples of 10dB, a plot of relative differences is obtained that correlates to the levels used for recording the clips.
Here Photoshop is used to merge the FR traces in the above manner:
The upcoming reveal will show if the above is correct. If multiples of 10dB were used for generating original upload:
then the Photoshop merging should be correct. Based on the merging, one driver clearly rides higher than all the others across the 400Hz-20kHz plot range. This driver is highly likely to be the vote leading driver.
In general, louder speaker will be preferred in a demonstration.
Is the zero vote driver one of the drivers in merged plot with lowest levels? Or perhaps the BMR driver with flat, lowest directivity diaphragm has lowest direct/reflected ratio that may make it sound lifeless? Which ever driver is B, it was early reject in my listening.
Based on merged plot, aside from leading driver, it seems that no driver should have 0 votes, or even only 1 vote. My guess is that voters are biased by comments, and by looking at poll results prior to casting votes.
From X's description of XO alignment, I expected merge to look more like this:
Or like this:
Even with plots amplitude aligned at 600Hz, the apparent vote leader still sticks out with highest amplitudes overall above 3kHz.
In rounds 3 and 4, drivers may be close enough response wise to get away without BSC and no additional EQ below 1kHz, in this round the vote spread indicates otherwise.
With traces isolated in Photoshop, I also created plots for level matching at 1kHz, 2kHz, 3kHz, 4kHz, and 6kHz, and assembled a colage:
Driver B has no votes. Why is this?
Based on Round-5-Meas-HR.pdf with measurements offset by 10dB intervals for ease of viewing, the assumption can be made that if traces are moved by offsets in multiples of 10dB, a plot of relative differences is obtained that correlates to the levels used for recording the clips.
Here Photoshop is used to merge the FR traces in the above manner:
The upcoming reveal will show if the above is correct. If multiples of 10dB were used for generating original upload:
then the Photoshop merging should be correct. Based on the merging, one driver clearly rides higher than all the others across the 400Hz-20kHz plot range. This driver is highly likely to be the vote leading driver.
In general, louder speaker will be preferred in a demonstration.
Is the zero vote driver one of the drivers in merged plot with lowest levels? Or perhaps the BMR driver with flat, lowest directivity diaphragm has lowest direct/reflected ratio that may make it sound lifeless? Which ever driver is B, it was early reject in my listening.
Based on merged plot, aside from leading driver, it seems that no driver should have 0 votes, or even only 1 vote. My guess is that voters are biased by comments, and by looking at poll results prior to casting votes.
From X's description of XO alignment, I expected merge to look more like this:
Or like this:
Even with plots amplitude aligned at 600Hz, the apparent vote leader still sticks out with highest amplitudes overall above 3kHz.
In rounds 3 and 4, drivers may be close enough response wise to get away without BSC and no additional EQ below 1kHz, in this round the vote spread indicates otherwise.
With traces isolated in Photoshop, I also created plots for level matching at 1kHz, 2kHz, 3kHz, 4kHz, and 6kHz, and assembled a colage:
Welcome! It's very fun.
I like these tests and appreciate the value of them, but trying to apply any sort of statistical analysis, ESPECIALLY without any attempt at controlling the innumerable degrees of freedom, is an exercise in futility.
+1 Barleywater.
This little anecdote is in line with almost everything ive ever read on subjective preferences (as long as responses aren't vastly different. And in these tests although there are FR differences which would seem large, the bass and lower mids are constant, or thereabouts.)
Were these drivers demoed alone full range, I suspect the preference would come down to which driver sounded fuller, as well as which is louder.
The brain seems very good at ignoring the warts and focussing on the beauty spots.
(I wouldn't be surprised if the BMR got few or no votes....ive only ever heard one, and it was uninspiring, flat, and too diffuse, indistinct, distant, locationless sounding.)
Last edited:
I saw that the overall level of A was on average way high and I thought about adjusting its level so that on average, its SPL level would be similar to the other drivers. However, what I did was to consistently adjust levels of the driver under test to match well to the woofer at the XO frequency from 350Hz out to about 500Hz. Right or wrong, I was trying to be consistent for all drivers. Sure, the whole concept of louder wins, is true to a certain extent, but I thought we were beyond being fooled by being just louder around here. Does that mean all we need to do is turn the volume up and our speakers sound better?
Voted
Even very interested in this stuff and hobby had difficult this round to find the time needed because of job and other stuff, therefor this time just voted for my preferred sound clip which should not be hard to figure out was the last one by looking at horse or car race above. For me had a local finale between lightest blue and last one where last one won because at my setup it sounded so close to reference track, lighest blue was very audiophile sounding maybe the bigger one of the participating Fountek's : )
Regarding the leading driver is it alright ask if voters intend to play it as response is out of box or they intend some and/or electric/DSP/acoustic correction, no problem people have different preference/taste but ask because it really sticks out under listening and on measurement having its own resonance signature which some even called is sort of broken and should be sent back for repair.
Barleywater thanks the analyze work with photoshop merging, something to think about with the louder one.
xrk971 great thanks sharing thread and hard work, bass for recorded drivers is much better and close to reference track this time, only complain : ) track A and F had not same sampling rate as REFERENCE/B/C/D/E/G.
Even very interested in this stuff and hobby had difficult this round to find the time needed because of job and other stuff, therefor this time just voted for my preferred sound clip which should not be hard to figure out was the last one by looking at horse or car race above. For me had a local finale between lightest blue and last one where last one won because at my setup it sounded so close to reference track, lighest blue was very audiophile sounding maybe the bigger one of the participating Fountek's : )
Regarding the leading driver is it alright ask if voters intend to play it as response is out of box or they intend some and/or electric/DSP/acoustic correction, no problem people have different preference/taste but ask because it really sticks out under listening and on measurement having its own resonance signature which some even called is sort of broken and should be sent back for repair.
Barleywater thanks the analyze work with photoshop merging, something to think about with the louder one.
xrk971 great thanks sharing thread and hard work, bass for recorded drivers is much better and close to reference track this time, only complain : ) track A and F had not same sampling rate as REFERENCE/B/C/D/E/G.
I think you approached the whole level matching problem well, and there isn't much you can do better here - its just one of the many 'degrees of freedom' (Derno) mentions/ed that are pitfalls it these tests which for me, preclude any real scientific 'best driver' out comes.
Besides average spl....well its almost meaningless. Sound power MAY be a better normalisation method vis a vis loudness perception (just riffing tbh, but it is widely used for noise analysis)
WRT the louder sounds better thing...subconscious behaviour isn't consciously avoided by critical listening.
Well every one of my speakers always sounded better louder ( to a point )
E.g. >90dB they seem to get moving. >100 and everything else starts moving
Perhaps another aspect of the ear brain mech? Fletcher Munchson curves?
Last edited:
You realised that B was early reject (don't you think that this also happened to everyone else?).
In term of dB level, B is average. Lowest dB level is I believe the most expensive FR88 (E). Don't forget that high distortion tends to promote higher dB level!!!
And frequency response also didn't tell why B is early reject.
I don't think this is bias because I believe in the power of human mind/brain in perceiving sound, even if they have terrible ears.
Judging from the sound of B, there are several things IME that can make drivers to sound like that. Lack of amplifier "drive" in combination with too heavy moving mass (usually for big woofers). The box volume is too big than required by Low Vas and low Qts drivers (usually for small drivers).
Yes. A is too loud but this cannot be 100% avoided as the response isn't flat. The higher frequency will always be louder.
But I think preference for A is the most honest and natural. Use only ears and no "bias" especially visual bias, and no reference to compare, A is the most logical preference imo. Technical people are usually biased by FR curve, but do you know exactly how good they can judge FR flatness by ears alone??
Alpine driver is not popular in home audio, but they are from great company with great history/experience. I think even Jaguar has used one of their drivers...
I have several Alpine drivers and IMO, despite their weaknesses they are "special".
Well every one of my speakers always sounded better louder ( to a point ). E.g. >90dB they seem to get moving. >100 and everything else starts moving
Perhaps another aspect of the ear brain mech? Fletcher Munchson curves?
Yes, there is optimum volume for every driver/amp. Fletcher Munson is one reason, the others are mentioned in my previous post.
When comparing top drivers such as within round 4, dB level is very critical because other parameters (distortion, FR, etc) are equal or above minimum threshold.
When comparing cheap drivers like in this round 5, preference and taste plays a big role...
Strangely, I didn't champion G (I think its the cheapest driver), even tho it has the right "balance". I tend to prefer DEF. Why? Because my preference has always been for (1) "sonic" (which is dependent on electronics), (2) distortion and (3) musicality.
Sonic performance can be improved or perceived differently in real implementation. But low moving mass and/or sensitive drivers tend to bring more "sonic". I had a feeling that D is better than EF in this aspect.
Distortion is so obvious in listening, that's why I'm sure E is the most expensive driver. It is the best for long time headphone listening with my Samsung because the distortion is very very low.
Musicality, it's hard to explain. Correct tonality as usually displayed by speaker with tilting down frequency response. With my headphone I didn't find F special, but with my speaker it is, so I voted even tho the fatiguing level is too high for my taste (but these are cheap drivers that I will never used anyway )
When comparing cheap drivers like in this round 5, preference and taste plays a big role...
Strangely, I didn't champion G (I think its the cheapest driver), even tho it has the right "balance". I tend to prefer DEF. Why? Because my preference has always been for (1) "sonic" (which is dependent on electronics), (2) distortion and (3) musicality.
Sonic performance can be improved or perceived differently in real implementation. But low moving mass and/or sensitive drivers tend to bring more "sonic". I had a feeling that D is better than EF in this aspect.
Distortion is so obvious in listening, that's why I'm sure E is the most expensive driver. It is the best for long time headphone listening with my Samsung because the distortion is very very low.
Musicality, it's hard to explain. Correct tonality as usually displayed by speaker with tilting down frequency response. With my headphone I didn't find F special, but with my speaker it is, so I voted even tho the fatiguing level is too high for my taste (but these are cheap drivers that I will never used anyway
)
technical people? Lol is everyone else a underling? Such arrogance....
Technical people THINK they can diagnose issues, not based on some secret knowledge, but by experience, and all people are fooled by the same audio psychological tricks of the brain. To an extent it is true, maybe you'd be surprised how many techies are deaf.
As for loudness perception....its well established that full spectrum loudness is judged via sound intensity or power, not pressure; and also that octave to octave, sensitivities follow Fletcher Munsen.
Hence the use of noise power and noise intensity measurements, by industry, in order to put a value to the perception (read annoyance) of large machine noise emission (BS EN 1680 if I recall)
How one would normalise a recording to meet "average sound power by octave with A weighting" ....?
Well that is a tough one.
Last edited:
(but these are cheap drivers that I will never used anyway
This happens a lot on this site, disregard for drivers based on price. I fail to see what price has to do with it.
Sure there "could or should" be more potential for greatness in more expensive drivers. But that doesn't make every expensive driver perform great or better at all than a cheaper one. The one thing I am sure about is that price is no guarantee for quality in this world. It should be, but it isn't.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.