AA:
Even with frequency responses that look very similar - there are differences that show up in the CSD. Look at the latest CSD plots here
http://www.diyaudio.com/forums/full...3in-5in-drivers-round-2-a-26.html#post4330883
Even with frequency responses that look very similar - there are differences that show up in the CSD. Look at the latest CSD plots here
http://www.diyaudio.com/forums/full...3in-5in-drivers-round-2-a-26.html#post4330883
CSD says it all
Hi xrk,
I think the CSD plots are the most informative measurement as they reveal energy decay over time, plus frequency response, plus resonance / cone break up modes.
Some care is required when reading the CSD graphs as the non linear energy spread in musical content Vs the totally " flat response" of a 20Hz to 20KHz wall of sound on a CSD plot needs to be allowed for.
In the real world with real music the long decays and extended ridges below
1 KHz are way more damaging & audible than the same size of ridge above
1 KHz.
Put another way, take two separate ridges both 2 octaves wide and both taking 10ms to decay.....
First a 2 octave wide ridge covering 200Hz to 800Hz and
second a 2 octave wide ridge covering 1,600 Hz to 6,400 Hz.
The lower frequency ridge is a massive problem which involves over 10 times the energy that the higher frequency ridge, despite the fact that the low frequency ridge covers just 600Hz Vs the 4,800 Hz bandwidth of the high frequency ridge.
Hope this helps and all the best
Derek.
Hi xrk,
I think the CSD plots are the most informative measurement as they reveal energy decay over time, plus frequency response, plus resonance / cone break up modes.
Some care is required when reading the CSD graphs as the non linear energy spread in musical content Vs the totally " flat response" of a 20Hz to 20KHz wall of sound on a CSD plot needs to be allowed for.
In the real world with real music the long decays and extended ridges below
1 KHz are way more damaging & audible than the same size of ridge above
1 KHz.
Put another way, take two separate ridges both 2 octaves wide and both taking 10ms to decay.....
First a 2 octave wide ridge covering 200Hz to 800Hz and
second a 2 octave wide ridge covering 1,600 Hz to 6,400 Hz.
The lower frequency ridge is a massive problem which involves over 10 times the energy that the higher frequency ridge, despite the fact that the low frequency ridge covers just 600Hz Vs the 4,800 Hz bandwidth of the high frequency ridge.
Hope this helps and all the best
Derek.
Off with the X-max, on with the X-mech...
Next Up...
I really do not understand Fostex's X-max listed on their spec sheets. It is very conservative. I have a Fostex FF225k (old version, white cone) - its listed X-max was 0.3 mm. I can guarantee You, it moved ALOT more then this, at least 6 millimeters, I saw it with my own eye and I am being conservative. And it "clipped" very gracefully. X-max specs on Fostex spec sheets are meaningless as Scott and Dave have mentioned:
Having said that, it would be nice if You guys would share what the X-mech of the FF85wk is, even if a little on the conservative side. In fact, we would be better off comparing the X-mech of all the drivers I am considering, as that is a finite, physical limitation of the driver that can not be exceeded. This will help me determine the cut-off frequency of the array.
Fostex FF85wk - 32/channel @ $1280/channel or $2568 for the pair.
2 Mms = 64 Kg , total Sd: 896 cm sq
10F/8424G00 - 25/channel @ $2145 or $4289 for the pair
3.24 Mms = 81 Kg, total Sd 925 cm sq (based on Klang & Ton specs
Truncating the bezels, I get 30/channel with a total Sd of 1110.
So if X-mech is the same, I get a little more with the 10F array, but I am expecting to run these lower anyway. Ideally, I would want to get down to 80 Hz. Based on the distortion plots on the Klang & Ton spec sheets, the 10F is cleaner lower, so I would expect to run this lower then the FF85wk. But if I can not get it down to 80 Hz, I will still need a bass array.
According to the CSD of both these drivers, the Fostex is cleaner between 300 and 400 Hz. But I do agree with X, that the on-axis peak at 10 kHz needs to be addressed, maybe physically with some kind high frequency absorber to mount in front of the driver to even out the power response. If Dave's EnAbling process tames that peak, I may go that way also...
What ever driver I choose, there is no doubt that this array is going to sound pretty amazing! 😀
That is all for me today, more to come tomorrow.
Allen
Next Up...
I really do not understand Fostex's X-max listed on their spec sheets. It is very conservative. I have a Fostex FF225k (old version, white cone) - its listed X-max was 0.3 mm. I can guarantee You, it moved ALOT more then this, at least 6 millimeters, I saw it with my own eye and I am being conservative. And it "clipped" very gracefully. X-max specs on Fostex spec sheets are meaningless as Scott and Dave have mentioned:
Having said that, it would be nice if You guys would share what the X-mech of the FF85wk is, even if a little on the conservative side. In fact, we would be better off comparing the X-mech of all the drivers I am considering, as that is a finite, physical limitation of the driver that can not be exceeded. This will help me determine the cut-off frequency of the array.
Fostex FF85wk - 32/channel @ $1280/channel or $2568 for the pair.
2 Mms = 64 Kg , total Sd: 896 cm sq
10F/8424G00 - 25/channel @ $2145 or $4289 for the pair
3.24 Mms = 81 Kg, total Sd 925 cm sq (based on Klang & Ton specs
Truncating the bezels, I get 30/channel with a total Sd of 1110.
So if X-mech is the same, I get a little more with the 10F array, but I am expecting to run these lower anyway. Ideally, I would want to get down to 80 Hz. Based on the distortion plots on the Klang & Ton spec sheets, the 10F is cleaner lower, so I would expect to run this lower then the FF85wk. But if I can not get it down to 80 Hz, I will still need a bass array.
According to the CSD of both these drivers, the Fostex is cleaner between 300 and 400 Hz. But I do agree with X, that the on-axis peak at 10 kHz needs to be addressed, maybe physically with some kind high frequency absorber to mount in front of the driver to even out the power response. If Dave's EnAbling process tames that peak, I may go that way also...
What ever driver I choose, there is no doubt that this array is going to sound pretty amazing! 😀
That is all for me today, more to come tomorrow.
Allen
A CSD ismuch more useful and easier to interpret if the time axis is in periods as opposed to time (msec)
dave
Thoughts anyone?
I was wondering this myself. Derek has brought up an interesting observation. What have the rest of You experienced with this?
If this is true, I am definitely better off considering the line of mid-basses to compliment the full-range array. The FF85wk would just squeeze in there with a 300 Hz cross-over. Based on this parameter, the 10F 8424 would need to be crossed at 350 Hz and the 10F 8414 matches the Fostex at 300Hz. The 8414 version of the 10F seems to do a hare better on the lower mid-range on the CSD. Would be nice to find the CSD on the SB:
SB Acoustics SB65WBAC - 40/channel @ $1260 or $2520 for the pair.
2.9 Mms/driver = 116 Kg/Array Total Sd: 800 cm sq.
I was wondering this myself. Derek has brought up an interesting observation. What have the rest of You experienced with this?
If this is true, I am definitely better off considering the line of mid-basses to compliment the full-range array. The FF85wk would just squeeze in there with a 300 Hz cross-over. Based on this parameter, the 10F 8424 would need to be crossed at 350 Hz and the 10F 8414 matches the Fostex at 300Hz. The 8414 version of the 10F seems to do a hare better on the lower mid-range on the CSD. Would be nice to find the CSD on the SB:
SB Acoustics SB65WBAC - 40/channel @ $1260 or $2520 for the pair.
2.9 Mms/driver = 116 Kg/Array Total Sd: 800 cm sq.
Hi Dave,
How can undefined time periods be easier to interpret??
If everyone uses msec then there is no doubt when comparing one driver to another if they are measured using different software.
Also, I like to actually calculate the energy involved as accurately as possible and for this you need to know mass and time as well as distance / acceleration.....
But hey ho each to his own!
Cheers
Derek.
How can undefined time periods be easier to interpret??
If everyone uses msec then there is no doubt when comparing one driver to another if they are measured using different software.
Also, I like to actually calculate the energy involved as accurately as possible and for this you need to know mass and time as well as distance / acceleration.....
But hey ho each to his own!
Cheers
Derek.
Dave, do you have a visual example? Any driver. (better if it is one of the ones I am considering, but not necessary, I want to see what this looks like..)
matching array of bass / mids from around 200Hz to 400Hz.
I would recomend this approach.
I use it with my silk BMR's crossed at 200Hz if the bass drivers are close to the BMR's.
If there is distance between the BMR's and bass drivers I use 80Hz crossover (THX) as crossing over higher causes problems if the distance is more than a meter or so.
An array of six or eight 10 inch or 12 inch drivers will rock...You can use more smaller drivers of course.
I have ended up with modified Pro drivers with low Qts, large voice coils, fabric surrounds and powerful motors all tuned for a very particular loading to suit my BMRs.
But there are some good off the shelf drivers that will still work well and give you endless hours of fun on the simulation trail....All part of the fun....No victory without a fight and all that!!😀
Cheers
D.
I would recomend this approach.
I use it with my silk BMR's crossed at 200Hz if the bass drivers are close to the BMR's.
If there is distance between the BMR's and bass drivers I use 80Hz crossover (THX) as crossing over higher causes problems if the distance is more than a meter or so.
An array of six or eight 10 inch or 12 inch drivers will rock...You can use more smaller drivers of course.
I have ended up with modified Pro drivers with low Qts, large voice coils, fabric surrounds and powerful motors all tuned for a very particular loading to suit my BMRs.
But there are some good off the shelf drivers that will still work well and give you endless hours of fun on the simulation trail....All part of the fun....No victory without a fight and all that!!😀
Cheers
D.
The periods are not undefined.
When done this way it is easy to directly read off the ring down at any frequency, and the Q of any resonance looks the same at any frequencies.
Changing to periods in the time axis makes the display of the HFs equivalent to the display of the LFs. With straight time the HFs are much more compressed than the lows and cannot be directly compared.
Look at any ATB CSD or ARTA burst decay mode (they don't quite get the scaling right).
IIRC Toole covers this subject off in his book (i was long converted way before the book)
dave
"Its the only way to be sure"
Thanks Dave,
But I think I will stick with msec and calculator, and like you say they dont get the scaling right....
Manual calculation works every time for me...."Its the only way to be sure"
Free bass mid driver for anyone that can name the film that comes from in the next 10 mins....!
Cheers
D.
Thanks Dave,
But I think I will stick with msec and calculator, and like you say they dont get the scaling right....
Manual calculation works every time for me...."Its the only way to be sure"
Free bass mid driver for anyone that can name the film that comes from in the next 10 mins....!
Cheers
D.
From ATB:
Snapped from this article (a good read) http://www.kirchner-elektronik.de/files/LoadspeakerMeasurementTechnology.pdf
This too: http://www.kirchner-elektronik.de/files/Photo-Story-OK_eng.pdf
dave
Attachments
ARTA doesn't get the vertical scaling quite right (actually the extents not the scaling -- i had to go back and edit my earlier post). mSec scaling is just totally wrong.
dave
About x-max and x-mech.... It may be true that some companies do their measurement different than others.
x-max should be able to be considered as the maximum the cone moves in a pistonic way. I have no doubt this would be true for Scan Speak.
A test of the 4 ohm version of the 10F is done by an independent enthusiast (formerly known as bikinpunk on diymobileaudio.com, now posting under his true name Erin)
http://medleysmusings.com/scan-speak-10f4424g00-discovery-4-midrange-4-ohm-version/
The test is done on a klippel (https://www.klippel.de/)
Compare that to the sheet from Scan Speak and You'll know you can trust their specs. You might want to read up on the klippel site as it is loaded with explanations what to look for.
From the other full range makers I cannot tell you anything useful. But so far I'd say I trust x-max on the 10F to mean what it should be (in my eyes).
On CSD: I've mentioned it to you before: in a straight line array you already have a time difference in arrival times from the nearest speaker compared to the top one.
Any ringing a driver has, would be exaggerated by the time difference created by that difference in distance to your ear. Draw it out on a paper and see.
Wait, I've got it for you:
Might not fit perfect but gives you a clue... center to center here is 84mm. My actual Array end result is 85mm and I have it on file but had this uploaded already.
Granted, Derek is right about audibility, but you do get the speaker you put in that line back on steroids. No way to EQ things like that out, not even FIR filters can help you there.
That's why I'd still pick the driver with a clean upper frequency CSD.
x-max should be able to be considered as the maximum the cone moves in a pistonic way. I have no doubt this would be true for Scan Speak.
A test of the 4 ohm version of the 10F is done by an independent enthusiast (formerly known as bikinpunk on diymobileaudio.com, now posting under his true name Erin)
http://medleysmusings.com/scan-speak-10f4424g00-discovery-4-midrange-4-ohm-version/
The test is done on a klippel (https://www.klippel.de/)
Compare that to the sheet from Scan Speak and You'll know you can trust their specs. You might want to read up on the klippel site as it is loaded with explanations what to look for.
From the other full range makers I cannot tell you anything useful. But so far I'd say I trust x-max on the 10F to mean what it should be (in my eyes).
On CSD: I've mentioned it to you before: in a straight line array you already have a time difference in arrival times from the nearest speaker compared to the top one.
Any ringing a driver has, would be exaggerated by the time difference created by that difference in distance to your ear. Draw it out on a paper and see.
Wait, I've got it for you:
Might not fit perfect but gives you a clue... center to center here is 84mm. My actual Array end result is 85mm and I have it on file but had this uploaded already.
Granted, Derek is right about audibility, but you do get the speaker you put in that line back on steroids. No way to EQ things like that out, not even FIR filters can help you there.
That's why I'd still pick the driver with a clean upper frequency CSD.
An externally hosted image should be here but it was not working when we last tested it.
Aliens, do I win?
We have a winner...!
Well done my fellow Alien fan!!
Now you can choose from a bunch of drivers I have laying around....
A Dynaudio tweeter, removed from a nice £2,000 plus pair of Dynaudio floorstanders....
A Peerless 8 inch mid bass ( got about 12 of these!)
A something "Gold" full range rectangular driver, I will need to check in the garage the full name.
A Manger....No only joking!!
There are more, I will check tomorrow if you want one, I would imagine you have plenty of spare drivers anyway!
.
Cheers
D.
Well done my fellow Alien fan!!
Now you can choose from a bunch of drivers I have laying around....
A Dynaudio tweeter, removed from a nice £2,000 plus pair of Dynaudio floorstanders....
A Peerless 8 inch mid bass ( got about 12 of these!)
A something "Gold" full range rectangular driver, I will need to check in the garage the full name.
A Manger....No only joking!!
There are more, I will check tomorrow if you want one, I would imagine you have plenty of spare drivers anyway!
.
Cheers
D.
Not that many actually... just two spare TC9's and some car woofers 🙂.
One tweeter wouldn't do me much good even if I generally like Dynaudio. But I'd take the Peerless 8"... 12 is good too (lol)
One tweeter wouldn't do me much good even if I generally like Dynaudio. But I'd take the Peerless 8"... 12 is good too (lol)
The peerless is actually an interesting driver, I will dig out some pics tomorrow.
It was designed for small sealed box, has good spec and sounded sweet from memory.
The voice coil is in front of the dust cap...Sounds weird I know.... I orederd them years ago when I was experimenting with an on wall Manger array....££££ for the Mangers killed the idea and I never found another use for the drivers.
Off to bed now but I'll get the driver info tomorrow.
Cheers
D.
It was designed for small sealed box, has good spec and sounded sweet from memory.
The voice coil is in front of the dust cap...Sounds weird I know.... I orederd them years ago when I was experimenting with an on wall Manger array....££££ for the Mangers killed the idea and I never found another use for the drivers.
Off to bed now but I'll get the driver info tomorrow.
Cheers
D.
yup.
you can see the arrival times on a straight array in this measurement of a 32 x 2" flat array. You see the closest, then the next, etc, etc......
you do hear it as a "tall" image.
But many feel the dynamics and coherency outweigh this.
Ideally you don't make the driver work hard either.
I didn't figure out that my array of 9 x 4" rolled off below 200hz until it was too late.
No wonder it was unsatisfying crossing it at 80hz (for a bass head like me).
Oh well, live and learn.
Norman
you can see the arrival times on a straight array in this measurement of a 32 x 2" flat array. You see the closest, then the next, etc, etc......
you do hear it as a "tall" image.
But many feel the dynamics and coherency outweigh this.
Ideally you don't make the driver work hard either.
I didn't figure out that my array of 9 x 4" rolled off below 200hz until it was too late.
No wonder it was unsatisfying crossing it at 80hz (for a bass head like me).
Oh well, live and learn.
Norman
Last edited:
"Kill zone" for arrival times
Hi Norman,
This is an interesting subject....I've been experimenting a fair bit to try and figure out the thresholds for audibility in real rooms (ie with reflections) compared to the theoretical thresholds.
I got quite close to getting some answers from the master, Tom Danley on another thread a few months ago.....He was talking about this issue in general but when I asked the specific questions he went silent and stopped posting ....
I cant blame him as its very useful information and has commercial value, especially in his business!
There appears to be a band of time delayed reflections that most listeners object to ie below a certain time delay our ears ignore / cant detect any issue, and above a certain delay we identify the sound as a room echo and treat it accordingly.
Its the time delays in between that "kill" the sound....the " kill zone" needs to identified and avoided if possible.
Food for thought????
Cheers
Derek.
Hi Norman,
This is an interesting subject....I've been experimenting a fair bit to try and figure out the thresholds for audibility in real rooms (ie with reflections) compared to the theoretical thresholds.
I got quite close to getting some answers from the master, Tom Danley on another thread a few months ago.....He was talking about this issue in general but when I asked the specific questions he went silent and stopped posting ....
I cant blame him as its very useful information and has commercial value, especially in his business!
There appears to be a band of time delayed reflections that most listeners object to ie below a certain time delay our ears ignore / cant detect any issue, and above a certain delay we identify the sound as a room echo and treat it accordingly.
Its the time delays in between that "kill" the sound....the " kill zone" needs to identified and avoided if possible.
Food for thought????
Cheers
Derek.
8 inch small sealed box woofer
Here is a photo of the 8 inch Peerless drivers, I have 10 of them, all boxed and unused.
You can have one for free (you pay the shipping!) and if it tests well I will sell the rest for say £20. per driver?
Cheers
D.
Here is a photo of the 8 inch Peerless drivers, I have 10 of them, all boxed and unused.
You can have one for free (you pay the shipping!) and if it tests well I will sell the rest for say £20. per driver?
Cheers
D.
- Status
- Not open for further replies.