Comb filtering? It's dreadful, can't live with it 😱... euhhh...
I can't say that it bothers me enough to look for another way to get what I want.
In fact, I'm not bothered at all. After careful considerations of other means I decided to take the plunge and just try. I don't regret that decision.
I won't (can't) deny the existence of comb filtering in a system like this. But so far the other benefits I get far outweigh the disadvantage of comb filtering for my goals.
I can't say that it bothers me enough to look for another way to get what I want.
In fact, I'm not bothered at all. After careful considerations of other means I decided to take the plunge and just try. I don't regret that decision.
I won't (can't) deny the existence of comb filtering in a system like this. But so far the other benefits I get far outweigh the disadvantage of comb filtering for my goals.
Last edited:
Because it's 8" in diameter and if you're going to use an 8" as a midrange you either need to cross over low so as to avoid the wide-narrow-wide radiation pattern, or cross slightly higher and directivity match it to a wave guide.
Just because the driver has been designed to have a wide bandwidth doesn't mean using it in that way is particularly sensible.
Yes it should starting at around 3khz in that particular case. Probably visible on a 1/24 octave scale.
Excellent paper on that subject:
http://www.audioroundtable.com/misc/nflawp.pdf
Last edited:
As Toole says, much more of an issue to the eyes looking at a graph than it is to the ear/brain.
dave
As Toole says, much more of an issue to the eyes looking at a graph than it is to the ear/brain.
dave
There is more going on than meets the eye...
You manage to have a flat response down to 20Hz with a Fs of 125Hz ( TC9FD18-08 ) and no visible comb filtering effects. That is beyond me!!
Why amplifier ABX can be difficult (or not)
I know that 🙂 But I will bet if the procedure is set based on my own rule...
Look at the 2 files in PMA's listening thread (1_inv.wav and 1.wav). They are different in absolute polarity. I'm confident that I can (because I have) get 8/8 on Foobar ABX (it's the default number of trials in one round) after less than 10 rounds. So that means I should be allowed to make 80 picks within an hour! This is the hardest ABX that I have tried, and I think this is already my hearing limit.
Now with amplifiers, depends on the difficulty (which I think is not going to be difficult), I might need around 1 or 2 days for making 1 pick! So, you see the difference? Why is it so long? It will be long if the difference is only exist in the form of non-linear distortion, or fatigue (It's not easy to recognize small difference in amplifier distortion). But remember that we will live with our amplifiers for a long time! So sooner or later it (the effect) will come to us!
From my experience, amplifier ABX is very easy when it is done through software because we can easily go back and forth to find the difference. This process of finding the difference may take time (extreme situation is to listen to the distortion difference), but once the difference is understood, blind test is very easy.
Without software, finding the difference will take time. First, I will listen to only one amplifier playing music that I know well at least with 10 repetitions. Then after the sound has been registered into my brain/memory, I will switch to the next amplifier. It will be easier (but not fair) if we choose the right music or speaker or cable that will put any of the amplifiers into problem (That's why software based is more convenient/fair).
But from my experience, I don't need to make too much effort because amplifiers usually differ a lot in sibilance. For reference, in XRK's listening test, I can hear any file between TG9 or 10F in a few second and I will know which file it is only from listening to the sibilance (it's a tip for the incoming listening test). I don't know how hard it is for others, but for me it is too easy.
And last but not least, picking differences is much easier than knowing which one is better than the other, because we have to relate the sound we perceived with our knowledge or previous experiences. For example, "Oh this is second order distortion, from my experience it is not too harmful", or "Oh this is non-linear distortion, from my experience I can still live with this level of distortion", or "Oh this details is not real, usually I will have fatigue after 4 hours, so let's wait for 4 hours...", or the harder part: "Oh, what I heard is amplifier inability to drive the speaker, but I think I will like this speaker once I change the amplifier to a push-pull one".
I know that 🙂 But I will bet if the procedure is set based on my own rule...
Look at the 2 files in PMA's listening thread (1_inv.wav and 1.wav). They are different in absolute polarity. I'm confident that I can (because I have) get 8/8 on Foobar ABX (it's the default number of trials in one round) after less than 10 rounds. So that means I should be allowed to make 80 picks within an hour! This is the hardest ABX that I have tried, and I think this is already my hearing limit.
Now with amplifiers, depends on the difficulty (which I think is not going to be difficult), I might need around 1 or 2 days for making 1 pick! So, you see the difference? Why is it so long? It will be long if the difference is only exist in the form of non-linear distortion, or fatigue (It's not easy to recognize small difference in amplifier distortion). But remember that we will live with our amplifiers for a long time! So sooner or later it (the effect) will come to us!
From my experience, amplifier ABX is very easy when it is done through software because we can easily go back and forth to find the difference. This process of finding the difference may take time (extreme situation is to listen to the distortion difference), but once the difference is understood, blind test is very easy.
Without software, finding the difference will take time. First, I will listen to only one amplifier playing music that I know well at least with 10 repetitions. Then after the sound has been registered into my brain/memory, I will switch to the next amplifier. It will be easier (but not fair) if we choose the right music or speaker or cable that will put any of the amplifiers into problem (That's why software based is more convenient/fair).
But from my experience, I don't need to make too much effort because amplifiers usually differ a lot in sibilance. For reference, in XRK's listening test, I can hear any file between TG9 or 10F in a few second and I will know which file it is only from listening to the sibilance (it's a tip for the incoming listening test). I don't know how hard it is for others, but for me it is too easy.
And last but not least, picking differences is much easier than knowing which one is better than the other, because we have to relate the sound we perceived with our knowledge or previous experiences. For example, "Oh this is second order distortion, from my experience it is not too harmful", or "Oh this is non-linear distortion, from my experience I can still live with this level of distortion", or "Oh this details is not real, usually I will have fatigue after 4 hours, so let's wait for 4 hours...", or the harder part: "Oh, what I heard is amplifier inability to drive the speaker, but I think I will like this speaker once I change the amplifier to a push-pull one".
Here's the 1/48 oct. smoothed result, does that help? 😉
Must all be EQed.
There is no way you can get same output at 20hz than 120hz ''naturally'' with such drivers, unless of course the measurement tools/methods are flawed.
Well, I did say:
But I think it's still pretty darn good in a "live" living room at the listening position, 3 meter away from the speakers...
What you say is very interesting, Jay.
I'm strongly under the impression that human auditory memory is very short and comparison test accuracy depends of the A/B switching speed.
But maybe you're right. Maybe an individual with enough ''practice'' will be able to nail easily a blind test while others can't. But i never witness that....
The biggest blind test i ever organized in our lab was about MP3/AAC v.s. CD (16/44/48) v.s. HD 24/96... Over 350 participations spreaded on many days...
And first thing we had to do was to downgrade even more the MP3's ''very poor quality reference'' from 128kbps to 64kbps. Because the 128kbps was -very surprisingly- extremely difficult to identify my most testees (the AAC version was not identified at all, not even once).
Kind of stimulated by that, a small group decided to ''practice'' for many hours and by using other audio reproduction equipment; planar speakers, headphones, monitors, other DAC's, etc...
No one succeeded to identify anything at 192kbps or above. And i'm not even talking CD v.s. HD, which were completely unidentifiable of course.
Last edited:
Appreciation-tests have next to no scientific value. On the other hand it has Statistic value.
We can end up with interesting data such as ''587 people out of 1,000 prefered Driver A while Driver F was disliked by 912 people out of 1,000''
It doesnt really prove anything, but it has statistic value if the methodology is correct and applied the same to every testees.
But then again, we must be aware of the placebo effect! That is why identification test should always be made before a subjective/appreciation test.
'' CAN YOU identify differences ? ''
if yes, then you're allowed to express your appreciation. 😉
The problem with those blind tests are, they pick "any man on the street" and make a conclusion on audibility based on that. What important is quality, not quantity. I believe you can get at least 5 from DIYA members who can pick the difference between MP3 (128k) and CD.
I knew that my ears are sensitive, but my hypothesis has been that even if you cannot hear the difference/distortion it doesn't mean that you are not affected by it.
So if you cannot do an ABX between $3 amplifier and $3000 amplifier, it doesn't mean that you are okay choosing the $3 one. The amplifiers are NOT the same to you, you just cannot do it in an ABX! (of course in a real life, there are situations where you have to run the $3 amp into clipping)
Try to measure the level of enjoyment you have with $3 amp after 3 months and compare with the $3000 one. If it is still difficult, there is an easier way: calculate how many hours you listen to your system (with different amps) in a day or week or month. Usually it is so obvious.
So don't be happy if you want to change your system every week. It is usually a sign that it is not worth it 😉 Many truly good systems end up become a classic and for whatever reason you want to keep it.
Last edited:
Another issue is how first impression is the opposite of long listening experience. Details, non transient perfect low bass, second harmonics, etc.
Spot the line array
Hi Jon,
Please post post a photo of your line arrays.....Not the 3 way WMTMW (or 4 way ? WMMTMMW) speakers you have suspended from the ceiling!
Why curve a WMTMW design?
This will worsen the already poor (unless carefully managed with powerful DSP and the Smaart Live or similar software ) time domain errors in this design.
By definition the height of a line array must be equal to or greater than 75% of the floor to ceiling height of the room it is in.
Your WMTMW design is in a very tall room, a line array to cover that room is going to be a challenge!
Pro sound line arrays (usually curved and flown) are only possible with the combination of very powerful DSP speaker management systems, high power amps and very high power handling and efficient drivers.
Re comb filtering in true line arrays, as Wesayso (and Toole!) says, its a near field theoretical issue that only exists in anechoic chambers for one point in space....
In real world rooms / venues, at any kind practical listening distance, room reflections and head movement eliminate the problem.
Power response of driver also plays a major role hear....The better the power response the less of an issue it becomes.
Hope this helps
Derek.
Hi Jon,
Please post post a photo of your line arrays.....Not the 3 way WMTMW (or 4 way ? WMMTMMW) speakers you have suspended from the ceiling!
Why curve a WMTMW design?
This will worsen the already poor (unless carefully managed with powerful DSP and the Smaart Live or similar software ) time domain errors in this design.
By definition the height of a line array must be equal to or greater than 75% of the floor to ceiling height of the room it is in.
Your WMTMW design is in a very tall room, a line array to cover that room is going to be a challenge!
Pro sound line arrays (usually curved and flown) are only possible with the combination of very powerful DSP speaker management systems, high power amps and very high power handling and efficient drivers.
Re comb filtering in true line arrays, as Wesayso (and Toole!) says, its a near field theoretical issue that only exists in anechoic chambers for one point in space....
In real world rooms / venues, at any kind practical listening distance, room reflections and head movement eliminate the problem.
Power response of driver also plays a major role hear....The better the power response the less of an issue it becomes.
Hope this helps
Derek.
But you're assuming that.
On the contrary; besides a lot of audiophiles, musicians, music engineers, even music critic and luthier did the test.
I think you're a bit drowned into the Golden ears theory. 🙄
i have no line arrays, like i said it's array-like.
Tomahawk, who helps me build these, was very concerned as well about comb-filtering. But these speakers were not at all intended for any audiophile/hifi purposes, it's only used for ambiant music in a showroom (non-audio products) and they do that job extremely well.
If i were to experiment with line arrays i would -of course- follow the basic rules of the concept. But for my home system, i'm now looking for a WMMTMW project.
Last edited: