AudioCircle

Industry Circles => Salk Signature Sound => Topic started by: audiocrazy on 9 Jun 2009, 03:41 pm

Title: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 9 Jun 2009, 03:41 pm
I took the measurements of my SongTowers using RoomEQ wizard. I have SVS sub and the xover is set to 80hz. I have SVS AS-EQ1 to equalize the sub & Integra DTC-9.8 with Audyssey.

I'm concerned about the FR from 1k-20Khz as its a down hill. I was expecting a very small variance about +/-3db but its more than 8db difference from rest of the range.
I also turned off all EQ but still the graph is same for 1k-20khz

I've GIK Panels(room kit 2) in my room 16x12. I've a 3 pairs of GIK 242 on the side walls to catch the reflection and 244 bass trap on all the corners and monster bass trap on the rear wall.

Has somebody taken any measurements with REW?

(http://farm4.static.flickr.com/3605/3611216886_1af09bb23b_o.jpg)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: fsimms on 9 Jun 2009, 04:32 pm
If you notice that people who post measurments usualy say "near field".  You see now why the don't post the measurements at the listening position.   This is a normal room effect of treble drop off.  Notice how your ears tell you that the treble is natural.  All flat speakers have that drop off.

Bob
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 9 Jun 2009, 04:40 pm
If you notice that people who post measurments usualy say "near field".  You see now why the don't post the measurements at the listening position.   This is a normal room effect of treble drop off.  Notice how your ears tell you that the treble is natural.  All flat speakers have that drop off.

Bob

There will generally be a roll-off due to high frequency absorption by various boundaries, but it should not be as extreme as this, and the cavity at 2 khz is hard to explain.  If the ST really sounded like that at your listening position, I think you would have returned them.  Does your equipment allow you to do any other type of measurement, say a closer field (1-2 meters) with a fast FFT signal that doesn't include room reflections, cancellations? 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TF1216 on 9 Jun 2009, 04:44 pm
Is it worth removing the panels to take another measurement?  You could then place them back in a couple at a time and witness the effects.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 9 Jun 2009, 05:24 pm
There will generally be a roll-off due to high frequency absorption by various boundaries, but it should not be as extreme as this, and the cavity at 2 khz is hard to explain.  If the ST really sounded like that at your listening position, I think you would have returned them.  Does your equipment allow you to do any other type of measurement, say a closer field (1-2 meters) with a fast FFT signal that doesn't include room reflections, cancellations?
Dennis thanks for your inputs. Actually when I got my ST I had the tweeter problem in my left speaker and Jim had sent me another pair of OW2 that I had to solder it. Do you think that might have messed up the tweeters?

I use Behringer ECM8000 microphone with EMU 0404 USB sound card to take measurment. Well I could put the mic stand closer to take near field measurment. I'll also try to take measurment without any side acoustic panels

Here is the FR response of my center speaker
(http://farm3.static.flickr.com/2442/3610698789_1610cebcbf_o.jpg)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 9 Jun 2009, 06:02 pm
Very interesting.   Now I'm worried that you didn't connect the new tweeter with the right polarity. That would cause the suck-out in the 2 khz area.  Are you sure you know which of the wires coming from the crossover is positive, and that you connected that to the positive tweeter terminal?  Generally Jim uses a blue wire for positive, and clear for negative.  The 0W2 voice coil is very robust and should not have been damaged by soldering unless you just let it cook and cook.   And then there should have been a total failure. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 9 Jun 2009, 06:39 pm
Very interesting.   Now I'm worried that you didn't connect the new tweeter with the right polarity. That would cause the suck-out in the 2 khz area.  Are you sure you know which of the wires coming from the crossover is positive, and that you connected that to the positive tweeter terminal?  Generally Jim uses a blue wire for positive, and clear for negative.  The 0W2 voice coil is very robust and should not have been damaged by soldering unless you just let it cook and cook.   And then there should have been a total failure.
I'm very positive that the tweeters were hooked up with correct polarity (but that was about 3 months ago). When I removed the bad tweeter I took a note of how the tweeters were hooked up & on the back of 0W2 there is + and - sign.
When I soldered the new tweeters i made sure they were the correct polarity. But when i go home tonite I will re-check it again.
Apart from physically checking the connection what's the best way to find out if my tweeters are either bad or hookup incorrectly.
Has somebody else on the forum taken any REW measuremnts of the ST's?
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 9 Jun 2009, 06:58 pm
Dennis how much variance in db is acceptable for on axis measurement at the primary listening position?
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Russtafarian on 9 Jun 2009, 07:06 pm
For perspective, look at the measured FR curves for headphones at headphone.com.  None of them measure flat, they'd be unlistenable if they did. If your looking for a target curve as a point of reference, look at one for a HD600 or K701 headphone.  You'll never replicate that with your speakers in your room but it will show the kind of treble rolloff you can expect.

Russ

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 9 Jun 2009, 07:33 pm
Dennis how much variance in db is acceptable for on axis measurement at the primary listening position?

Well, if you're going to make any sense out of a room measurement, you have to invoke smoothing, and maybe as much as 1/3rd octave.  If the anechoic response of the speaker is smooth within 1-2 dB, and you've smoothed enough to get rid of the meaningless dips that change with your position, the room response should be almost as flat except for a little more than the top octave, which will slope down.  Stereophile publishes room response curves regularly, and they often look smoother than the 1-meter anechoic response.   So I have my suspicions about the mic calibration and the basic measuring technique that the program is using for the ST's.   
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 9 Jun 2009, 07:34 pm
Very interesting.   Now I'm worried that you didn't connect the new tweeter with the right polarity. That would cause the suck-out in the 2 khz area.  Are you sure you know which of the wires coming from the crossover is positive, and that you connected that to the positive tweeter terminal?  Generally Jim uses a blue wire for positive, and clear for negative.  The 0W2 voice coil is very robust and should not have been damaged by soldering unless you just let it cook and cook.   And then there should have been a total failure.
I'm very positive that the tweeters were hooked up with correct polarity (but that was about 3 months ago). When I removed the bad tweeter I took a note of how the tweeters were hooked up & on the back of 0W2 there is + and - sign.
When I soldered the new tweeters i made sure they were the correct polarity. But when i go home tonite I will re-check it again.
Apart from physically checking the connection what's the best way to find out if my tweeters are either bad or hookup incorrectly.
Has somebody else on the forum taken any REW measuremnts of the ST's?

Why don't you just measure one ST at a time in the same position and see whether the curves agree?
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 9 Jun 2009, 09:21 pm
A couple of comments. 

First, I emailed audiocrazy with some test ideas he could try.  I won't go into them in detail here, but one of them was the same comment Dennis posted.  By measuring each speaker separately, you can easily see if the measurements track one another.  If they do, the tweeters are wired correctly.

Second, the wires we use in any given crossover can vary in color.  Most often, we use a white wire for the positive tweeter terminal.  But it all depends on what colors we can get.  In all cases, the darker wire is negative and the lighter positive.

Finally, I do have some comments about programs like this and auto-EQ functions on receivers.  But I don't have time to collect my thoughts right now.  I'll try to do that and post a few thoughts this evening.

- Jim
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: s_cervin on 9 Jun 2009, 11:23 pm
I'll try to do that and post a few thoughts this evening.

- Jim

Only between periods I hope!!

Audiocrazy,
If you are confident in your measurements and speaker setup (placement, wiring, etc), I would begin by removing various traps (maintaining symmetry) and overlay the plots as part of the clue generation.


GO WINGS!
SC
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 12:09 am
Hey all,

I think audiocrazy's mic calibration file is wrong.  I too use an ECM8000 mic with my REW setup and look at my "correct" file (the black line).  Please note this is a friend's in-room measurement prior to getting his setup better placed (not Salk speakers either).

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/augermerrymaid.jpg)

For the record, I have never seen a measurement roll off like audiocrazy's does.  He absolutely has his REW set up incorrectly. 

Don't panic guys, his Salk speakers are just fine.  :wink:
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 10 Jun 2009, 01:20 am
OK...here are a few randomly related thoughts.

Imagine this scenario...

We work very hard to turn out an absolutely gorgeous pair of speakers.  The owner, wanting to get the best performance possible out of them, rushes out and purchases a Radio Shack SPL meter.  He locates a table of calibration corrections on the internet and carefully makes measurements, applies the corrections and plots a graph.  It looks terrible.  He thought the speakers sounded great when they arrived, but the measurements indicate something is terribly amiss.  He contacts us in panic.  His speakers are certainly defective and he has the graphs to prove it. 

"How do they sound?", I ask?  "They sound fine," he responds.

If he had just avoided that Radio Shack purchase, he could blissfully enjoy his speakers.  The problem isn't with his speakers.  The problem is with his measurements.  But at least on a temporary basis, it ruins his listening experience.

Unfortunately, this is a common occurrence and one we have had to deal with far more often than we would like.   

Here's another...

A person uses a speaker measuring system to measure his speakers.  The bass response is terrible! It falls off early and there are ugly peaks and valleys in the response.  There must be something dramatically wrong.

So I ask how the measurement was done.  He indicates it was done at the listening position.  He goes on to say that no matter where in the room he takes the measurements, the graphs vary somewhat, but not that much.  There is clearly a problem.

I explain to him that he was measuring the room.  Unless you take near field measurements at the driver, there is no way to obtain accurate measurements below about 200Hz.  If the measurement times are long enough to measure the bass response from the listening position, you are also measuring the room. 

Room measurement tools like REW (and even a Radio Shack SPL meter) can be very useful in terms of optimizing your room.  But they will not provide an accurate indication of the performance of a speaker.  For that you need a quality speaker measurement system and a well-calibrated measurement microphone.  And you also need a good bit of time to learn proper techniques for using these tools. 

Too often, these tools do more harm than good.  Improper measurement technique and a failure to understand how these tools work can cause you to needlessly lose confidence in your audio equipment.

As long as I am on a roll, I might mention reservations I have with all the new room EQ functions in receivers these days.  I can't tell you how many times I have asked customers to turn off the EQ function and just use the speakers as they are.  In almost every case, they reported a dramatic improvement in performance.

This would not seem to make sense.  Here we have a magic electronic gadget that can correct for any flaw in the speakers/room and do it at the listening position.  But all too often, it simply doesn't work.

On quite a few occasions, I have used DEQX (a very powerful and capable system) to apply a simple correction to a single loudspeaker.  In other words, I was not using it as a crossover, just using it to measure and correct for anomalies in the speaker performance. In almost every case, the speaker performed far better without the corrections.

IMO, the only thing an automated receiver EQ system should be used for is to deal with room modes.  It should not touch the signal above, say, 100Hz.  The only thing it can do at higher frequencies is screw things up.  I am very confident that the speakers we produce have absolutely no need for correction.  There simply isn't anything to correct.

Any issue that does require correction is room-related.  Changing speaker performance in order to deal with poor room acoustics is, plain and simple, a very bad idea.

The right technology, applied correctly, can be extremely beneficial in achieving great system performance.  But sometimes technology can be a curse.  It seems like every major receiver manufacturer today is offering an auto-EQ function. This may be great marketing, but in many cases, it is a terrible use of technology. Others may disagree with this advice, but if you have a function such as this on your receiver, please turn it off.  In most cases, you will be a lot happier.

End of diatribe.

- Jim



Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 01:34 am
I'm am certainly no professional at audio stuff, but I can tell you that REW has to be used properly.  You absolutely cannot place too much value on it's measurements.  Most times, you would be hard pressed to even hear what it's showing you.  Also, movements of as little as 6" in any direction at the seating position will change the readings significantly. 

I use REW mostly for tuning subs.  For this it is priceless.  It is also perfect for finding any suck-outs you may have in-room.  I do use it to verify that my speaker placement is as good as it gets for the locations I am willing to live with (not much I can assure you).  Other than this, it's not much good.



Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: rahimlee54 on 10 Jun 2009, 01:48 am
I really like my EQ when I am watching a movie, because it does its job, which is balance all my speaker volumes.  For 2 channel I had to read the manual and figure out how to by pass anything the AVR will do, after that it was smooth sailing.

Personally I wouldn't want to measure a speaker unless I have a capable room and calibrated equipment, anything other than that and accuracy is out the window.  That could just be me though.  Unless you are EQing your sub.

I could be wrong though.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 10 Jun 2009, 02:33 am
Frankly, if I were Audiocrazy, I would feel a little ganged up on at this point.  I'm sure he's following the measurement instructions as best he can, and he's getting results that look problematic.  So I really want to be as constructive as possible.  I would be very interested to see separate measurements of each ST in the same position.   That should at least cast light on the tweeter polarity issue.   That leaves the droop in the response at higher frequencies.   Audiocrazy--where are you located?  I would be happy to send you a small, light 2-way that you could measure to see whether you get results anything like mine.  It's difficult to make any progress unless there is a known reference point   
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 03:25 am
Frankly, if I were Audiocrazy, I would feel a little ganged up on at this point.  I'm sure he's following the measurement instructions as best he can, and he's getting results that look problematic.  So I really want to be as constructive as possible.  I would be very interested to see separate measurements of each ST in the same position.   That should at least cast light on the tweeter polarity issue.   That leaves the droop in the response at higher frequencies.   Audiocrazy--where are you located?  I would be happy to send you a small, light 2-way that you could measure to see whether you get results anything like mine.  It's difficult to make any progress unless there is a known reference point

Ok Dennis, time to "gang up" on you.  :P

I may be wrong, but I don't think it's possible to reverse the polarity of one tweeter and have it cancel out the other tweeter above say...3kHz?  Audiocrazy's measurement should have nothing to do with tweeters out of phase.

I certainly wasn't trying to "gang up" on him, but I'm telling you without any doubt that his REW setup is not correct.  I'd suggest he post that graph pic on hometheatershack and let the pros (Bruce and Wayne) get audiocrazy fixed up.  I can assure you that his measurement will look as it should and all the speculation contained in this thread can be concluded.  :wink:

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 10 Jun 2009, 04:40 am


Any issue that does require correction is room-related.  Changing speaker performance in order to deal with poor room acoustics is, plain and simple, a very bad idea.

The right technology, applied correctly, can be extremely beneficial in achieving great system performance.  But sometimes technology can be a curse.  It seems like every major receiver manufacturer today is offering an auto-EQ function. This may be great marketing, but in many cases, it is a terrible use of technology. Others may disagree with this advice, but if you have a function such as this on your receiver, please turn it off.  In most cases, you will be a lot happier.
Bingo.  That sums up everything I was going to say, and I do use REW.  I only use it to EQ below 100Hz, though. 

Audiocrazy, I mean no harm in this question: did you follow the Home Theater Shack REW FAQ to a "T," especially the part about sound card calibration? 

What you're likely dealing with is room issues.  As suggested, measure each speaker individually at the listening position, then measure each speaker individually at 1-2 meters or so.  I bet those anomalies go away with nearfield measurements, which prove the issue is the room.  Speaking of which, start with zero room treatment panels, measure, place where necessary (if necessary) and move forward from there.  You may just have to find a more suitable position for the speakers, or perhaps better positions for the acoustic panels. 

Keep us posted.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 10 Jun 2009, 05:28 am
Frankly, if I were Audiocrazy, I would feel a little ganged up on at this point.  I'm sure he's following the measurement instructions as best he can, and he's getting results that look problematic.  So I really want to be as constructive as possible.  I would be very interested to see separate measurements of each ST in the same position.   That should at least cast light on the tweeter polarity issue.   That leaves the droop in the response at higher frequencies.   Audiocrazy--where are you located?  I would be happy to send you a small, light 2-way that you could measure to see whether you get results anything like mine.  It's difficult to make any progress unless there is a known reference point
Dennis I'm located in Minneapolis, MN.
Sorry I did not mean to start a flame  :oops: but I'm trying to figure out if there is something wrong in my setup.
I took measurments of speakers individually and here is the graph.
I'll add more details tomorrow morning. Gotta go to work.

Red Line is Right Speaker
Green line is left speaker

(http://farm3.static.flickr.com/2474/3613307944_497461b67a_o.jpg)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: fsimms on 10 Jun 2009, 01:18 pm
I have trouble believing that chart.  It shows the SongTowers have a peak at 18 Hz!!!!  I didn't know they were that good!  :lol: :lol: :lol: 

A flat speaker would bedown about 10 db at the listening position at 20 kHz.  According to Jim's chart the SongTowers are down 5 db at 20 kHz.  The chart shows a drop of 20 db which is about 5 db worse than would expected.

Edit.  I expect the chart is shifted to the left due to the speed of your computer.  You are probably seeing the drop at the right at 30 or 40 kHz.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 10 Jun 2009, 01:28 pm
Frankly, if I were Audiocrazy, I would feel a little ganged up on at this point.  I'm sure he's following the measurement instructions as best he can, and he's getting results that look problematic.  So I really want to be as constructive as possible.  I would be very interested to see separate measurements of each ST in the same position.   That should at least cast light on the tweeter polarity issue.   That leaves the droop in the response at higher frequencies.   Audiocrazy--where are you located?  I would be happy to send you a small, light 2-way that you could measure to see whether you get results anything like mine.  It's difficult to make any progress unless there is a known reference point

Ok Dennis, time to "gang up" on you.  :P

I may be wrong, but I don't think it's possible to reverse the polarity of one tweeter and have it cancel out the other tweeter above say...3kHz?  Audiocrazy's measurement should have nothing to do with tweeters out of phase.

I certainly wasn't trying to "gang up" on him, but I'm telling you without any doubt that his REW setup is not correct.  I'd suggest he post that graph pic on hometheatershack and let the pros (Bruce and Wayne) get audiocrazy fixed up.  I can assure you that his measurement will look as it should and all the speculation contained in this thread can be concluded.  :wink:

Hoooookay--time to gang back.  I wasn't claiming one tweeter would cancel out another.  Reversing the polarity on one tweeter, or both, will cause a suckout in the crossover region, because the woofer and tweeter are supposed to be perfectly in phase at that point.  if you flip the polarity, they cancel in the region of 2-3 khz, which is why I suggested there might be a problem.  I wasn't referencing the droop further up.  That looks more like a mic calibration problem. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 10 Jun 2009, 01:36 pm
Dennis I'm located in Minneapolis, MN.
Sorry I did not mean to start a flame  :oops: but I'm trying to figure out if there is something wrong in my setup.
I took measurments of speakers individually and here is the graph.
I'll add more details tomorrow morning. Gotta go to work.

Red Line is Right Speaker
Green line is left speaker

(http://farm3.static.flickr.com/2474/3613307944_497461b67a_o.jpg)
[/quote]

Hmmmmmm.   Thanks for doing that, but it's difficult to believe you have the speakers in the same position.  There are just too many differences in the specific and overall frequency response to be explained by crossover issues or driver variation.  If you do have them in the same position, then I would have to agree with others that there is something off in the calibration or measuring technique.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 03:23 pm
Frankly, if I were Audiocrazy, I would feel a little ganged up on at this point.  I'm sure he's following the measurement instructions as best he can, and he's getting results that look problematic.  So I really want to be as constructive as possible.  I would be very interested to see separate measurements of each ST in the same position.   That should at least cast light on the tweeter polarity issue.   That leaves the droop in the response at higher frequencies.   Audiocrazy--where are you located?  I would be happy to send you a small, light 2-way that you could measure to see whether you get results anything like mine.  It's difficult to make any progress unless there is a known reference point

Ok Dennis, time to "gang up" on you.  :P

I may be wrong, but I don't think it's possible to reverse the polarity of one tweeter and have it cancel out the other tweeter above say...3kHz?  Audiocrazy's measurement should have nothing to do with tweeters out of phase.

I certainly wasn't trying to "gang up" on him, but I'm telling you without any doubt that his REW setup is not correct.  I'd suggest he post that graph pic on hometheatershack and let the pros (Bruce and Wayne) get audiocrazy fixed up.  I can assure you that his measurement will look as it should and all the speculation contained in this thread can be concluded.  :wink:

Hoooookay--time to gang back.  I wasn't claiming one tweeter would cancel out another.  Reversing the polarity on one tweeter, or both, will cause a suckout in the crossover region, because the woofer and tweeter are supposed to be perfectly in phase at that point.  if you flip the polarity, they cancel in the region of 2-3 khz, which is why I suggested there might be a problem.  I wasn't referencing the droop further up.  That looks more like a mic calibration problem.

Ha ha!  :D  Thank goodness I was misunderstanding you because for a little while there, I was really disappointed in you. :P

Again, audiocrazy's REW setup is not correct.  The mic calibration is definitely wrong and the sound card calibration looks strange as well.  I wish I could help, but the bad thing about REW is that it is far from easy to setup your first time through.  I've set it up twice now and each time it's like starting all over again.  I want a new laptop, but the fact that I'd have to reinstall REW makes me live with what I've got.  Seriously!

Audiocrazy:

I think you should start by downloading the latest mic calibration file from hometheatershack for your ECM8000 and load it into your REW software.  That calibration file you have right now looks way off.  I wouldn't worry too much about the soundcard calibration for now.  It really doesn't affect the issue you've got going on right now anyway.   
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 10 Jun 2009, 03:57 pm

I think you should start by downloading the latest mic calibration file from hometheatershack for your ECM8000 and load it into your REW software. 

Keep in mind that the ECM8000 calibration file on that site is a generic one.  To be accurate, each individual ECM8000 requires a calibration file for that individual mic.  The generic file will get you close, but it will not be totally accurate.

- Jim
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 04:46 pm

I think you should start by downloading the latest mic calibration file from hometheatershack for your ECM8000 and load it into your REW software. 

Keep in mind that the ECM8000 calibration file on that site is a generic one.  To be accurate, each individual ECM8000 requires a calibration file for that individual mic.  The generic file will get you close, but it will not be totally accurate.

- Jim

That is true Jim, but there is no way the generic file is as far off as what audiocrazy's graph is showing.  I'm fairly certain all of the mics tested for the calibration file proved pretty close (or close enough for home use).  We're not discussing a few db's here, we're talking over 20db.  It is obvious to me what the issue is, apparently not so for the rest of you...

Good luck with findings audiocrazy, I'm fairly certain everything is just fine with your speakers and you should sit back and enjoy your Salk's by listening to music instead of frequency sweeps.

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 10 Jun 2009, 06:33 pm
Hmmmmmm.   Thanks for doing that, but it's difficult to believe you have the speakers in the same position.  There are just too many differences in the specific and overall frequency response to be explained by crossover issues or driver variation.  If you do have them in the same position, then I would have to agree with others that there is something off in the calibration or measuring technique.
Dennis you are right I took the measurement at listening position and did not move the speaker to same location.  I'll do that when I get a chance. I'm tied up at work and will be travelling.

Keep in mind that the ECM8000 calibration file on that site is a generic one.  To be accurate, each individual ECM8000 requires a calibration file for that individual mic.  The generic file will get you close, but it will not be totally accurate.
- Jim
Jim you are right I should get my mic calibrated to get accurate results.

I agree that REW graphs does not tell the full story and I've confirmed that when I added SVS AS-EQ1. The REW graph looked almost similar to what I had using Audyssey EQ. But it sounded much different. Nevertheless I've followed the REW instructions to the T and have consulted BruceK and Wayne several times when I was learning to use REW.

Jim and Dennis I really apologize if I have offended you in anyway  :oops:. My intent was trying to figure out if there was any issue in my setup.
All in all I am very happy with my ST's :inlove:. I value the efforts and dedication Jim and Dennis :thankyou: takes in building such nice speakers.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 10 Jun 2009, 07:37 pm
Hmmmmmm.   Thanks for doing that, but it's difficult to believe you have the speakers in the same position.  There are just too many differences in the specific and overall frequency response to be explained by crossover issues or driver variation.  If you do have them in the same position, then I would have to agree with others that there is something off in the calibration or measuring technique.
Dennis you are right I took the measurement at listening position and did not move the speaker to same location.  I'll do that when I get a chance. I'm tied up at work and will be travelling.

Keep in mind that the ECM8000 calibration file on that site is a generic one.  To be accurate, each individual ECM8000 requires a calibration file for that individual mic.  The generic file will get you close, but it will not be totally accurate.
- Jim
Jim you are right I should get my mic calibrated to get accurate results.

I agree that REW graphs does not tell the full story and I've confirmed that when I added SVS AS-EQ1. The REW graph looked almost similar to what I had using Audyssey EQ. But it sounded much different. Nevertheless I've followed the REW instructions to the T and have consulted BruceK and Wayne several times when I was learning to use REW.

Jim and Dennis I really apologize if I have offended you in anyway  :oops:. My intent was trying to figure out if there was any issue in my setup.
All in all I am very happy with my ST's :inlove:. I value the efforts and dedication Jim and Dennis :thankyou: takes in building such nice speakers.

I'm certainly not offended.  I'm just a little frustrated to see results that bad and not have a better grasp of what might be going on.  I'm still a little concerned about the 2-2.5 k suckout on one of the speakers.  That wouldn't be a calibration issue.  But I'll wait until I see the results when you do both from the same position.   
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: BradJudy on 10 Jun 2009, 08:52 pm
That is true Jim, but there is no way the generic file is as far off as what audiocrazy's graph is showing.  I'm fairly certain all of the mics tested for the calibration file proved pretty close (or close enough for home use). 

Yes, at one point I compared the calibration files for multiple ECM8000 mics purchased at different times, by different people.  While each was unique, the variations were very small (+/- 1db range) and within the bulk of the range, they were within a fraction of a db of each other. 

For home use, a calibration file shouldn't be needed with the ECM8000 and the mics are not consistent enough to make using a generic file worthwhile IMO.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 10 Jun 2009, 09:13 pm
Audiocrazy - can you measure the speakers individually, but when placed in the exact same listening position?  This will rule out a crossover or driver issue.

I assume the two measurements above vary because they were measured individually from where they reside rather than when placed in the exact same spot.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Jeff B. on 10 Jun 2009, 09:30 pm
It actually seems fairly obvious to me also, but unfortunately, I don't think my appraisal is going to agree entirely with TJHUB's conclusion, so you guys can do whatever you like with my comments here.

After reading through everything in the thread and looking all of the plots over I am seeing essentially what I would expect to see based on what I have read (with one exception, which I will get to in a minute).

First, regarding the mic calibration file, the Behringer ECM8000 is a fairly flat mic to begin with. Even to use it without a calibration file will not result in these apparent discrepancies, in fact I doubt you would notice the difference. Rather, I think the explanation is much simpler -

From what I gather, these are in-room response measurements measured at the listening location ( I did not see the actual distance specified, unless I missed it somewhere). If this distance is 8-10 feet then I would expect to see:

....room gain in the lowest octaves
....boundary cancellation and peaks due to the room response
....and a falling top two octaves due to directivity and absorption

And this appears to be exactly what we see. There is a difference between the two speaker, but he commented that he did not move them to the same location, so that can easily explain these differences too (at least partly).

Now, about the falling top-end response - I do not know how well the mic is set-up on the tweeter axis. If it is off-axis much it will contribute to an even greater drop, so this is something to consider as well.

I always check my speakers in my room using an RTA with pink noise to see how the balance looks in-room. When the balance is very flat at one meter it will still show a significant loss in top octave energy at 8-10 feet, but it doesn't sound that much different, because at that distance we are hearing a greater proportion of the speaker's power response, which the mic does not pic up. Again, what I see in these graphs correlates well with what I see in listening position measurements in my own room, with one exception.....

The dip at 2kHz in the one speaker should probably not be present in one and not the other. If his measurments were vertically off-axis (from the tweeter axis) then I could see picking up the beginning of a null in the lobing along this axis, but the top-end should be more rolled off if this were the case. Based on this, I tend to agree with Dennis, I think the tweeter in that speaker may be connected with the wrong polarity compared to the other speaker. Otherwise, I don't have much of any issue with what I am seeing here.

Jeff Bagby
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 10 Jun 2009, 11:21 pm
Seriously guys, please look at this:

Here is the graph I posted earlier in this thread.  The black line in this graph is the Behringer ECM8000 mic's calibration line.  In the blue circle, note which way the line goes and at what frequency.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/sierramerrymaid2.jpg)

Now look at audiocrazy's graph.  Again, the black line in this graph is the mic calibration line.  In the blue circle, note that the line is WAY off from where it should be.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/audiocrazyrew2.jpg)

I could be wrong, but I'm almost certain that the mic calibration file that audiocrazy is using is subtracting from the actual mic readings from 3kHz and up when it should be adding from 10kHz and up.  This in turn would cause his upper frequency measurement to be skewed way down.  Hmm...

There is no doubt that the 2kHz dip Dennis has noted shouldn't be affected by the mic calibration and there certainly could be an issue there, but it doesn't change what I'm trying to get you guys to see.

For grins, here is the in-room response of my HT2-TL's at my seating position.  The speakers are running full-range and are about 11.5' apart and just under 14' from my listening position.  The blue line is the left speaker and the red line is the right.  The low end peak is from my room.  Every speaker and subwoofer I've had in my room has that 38Hz peak and weird behavior around 60Hz.  I have no room treatments whatsoever. 

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/ht2tlleftbluevsrightred.jpg)

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/DSC03524Medium.jpg)



Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 10 Jun 2009, 11:57 pm
Seriously guys, please look at this:

Here is the graph I posted earlier in this thread.  The black line in this graph is the Behringer ECM8000 mic's calibration line.  In the blue circle, note which way the line goes and at what frequency.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/sierramerrymaid2.jpg)

Now look at audiocrazy's graph.  Again, the black line in this graph is the mic calibration line.  In the blue circle, note that the line is WAY off from where it should be.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/audiocrazyrew2.jpg)

I could be wrong, but I'm almost certain that the mic calibration file that audiocrazy is using is subtracting from the actual mic readings from 3kHz and up when it should be adding from 10kHz and up.  This in turn would cause his upper frequency measurement to be skewed way down.  Hmm...

There is no doubt that the 2kHz dip Dennis has noted shouldn't be affected by the mic calibration and there certainly could be an issue there, but it doesn't change what I'm trying to get you guys to see.

For grins, here is the in-room response of my HT2-TL's at my seating position.  The speakers are running full-range and are about 11.5' apart and just under 14' from my listening position.  The blue line is the left speaker and the red line is the right.  The low end peak is from my room.  Every speaker and subwoofer I've had in my room has that 38Hz peak and weird behavior around 60Hz.  I have no room treatments whatsoever. 

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/ht2tlleftbluevsrightred.jpg)

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/DSC03524Medium.jpg)


That's very interesting.  Thanks for the comparisons.  I'm having trouble getting my mind around the mic calibration plots you're showing.  Where do they come from?  Are you showing the generic correction file for Audiocrazy's set up?  If so, I would interpret it the way you do--the correction should add high treble. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Jeff B. on 11 Jun 2009, 12:24 am
Hmmmmm.....well, your graphs do make a good point worthy of consideration. We will see how this resolves itself.

Thanks, Jeff



Seriously guys, please look at this:

Here is the graph I posted earlier in this thread.  The black line in this graph is the Behringer ECM8000 mic's calibration line.  In the blue circle, note which way the line goes and at what frequency.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/sierramerrymaid2.jpg)

Now look at audiocrazy's graph.  Again, the black line in this graph is the mic calibration line.  In the blue circle, note that the line is WAY off from where it should be.

(http://i9.photobucket.com/albums/a90/TJHUB/REW%20Measurements/audiocrazyrew2.jpg)

I could be wrong, but I'm almost certain that the mic calibration file that audiocrazy is using is subtracting from the actual mic readings from 3kHz and up when it should be adding from 10kHz and up.  This in turn would cause his upper frequency measurement to be skewed way down.  Hmm...

There is no doubt that the 2kHz dip Dennis has noted shouldn't be affected by the mic calibration and there certainly could be an issue there, but it doesn't change what I'm trying to get you guys to see.

For grins, here is the in-room response of my HT2-TL's at my seating position.  The speakers are running full-range and are about 11.5' apart and just under 14' from my listening position.  The blue line is the left speaker and the red line is the right.  The low end peak is from my room.  Every speaker and subwoofer I've had in my room has that 38Hz peak and weird behavior around 60Hz.  I have no room treatments whatsoever. 

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/ht2tlleftbluevsrightred.jpg)

(http://i9.photobucket.com/albums/a90/TJHUB/Salk%20Speakers/DSC03524Medium.jpg)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 11 Jun 2009, 12:38 am
Jeff:

Thanks.  I have a hard time getting my point across sometimes.  I'm glad you are now seeing what I'm trying to communicate.  Until audiocrazy gets his REW running correctly, I wouldn't worry too much about anything else.  There may be other issues, but you cannot draw any solid conclusions from his graphs.

The first graph is mine to show the correct ECM8000 mic calibration plot (the black line).  I can only assume that the mic reads low above 10kHz and the calibration file adds to the actual measurement.

The second graph is audiocrazy's to show that his mic calibration file is the opposite direction.  I can only assume that his calibration file is subtracting from the actual measurements thus causing the poor looking graph.

I'm at work right now and can't play with things to completely understand what's going on here.  If I wasn't working 13-15 hours days, I could be of more help.  I haven't even turned on my setup to listen to my Salks now since weekend before last.  :(



I could graph his data for him if he send me the data file.  I've done that with other forum members in the past to show them things.   

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: idiotec on 11 Jun 2009, 02:54 am
Interesting thread.  I must say, it makes me very happy to have neither measuring equipment nor an EQ.  :P
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 11 Jun 2009, 05:13 am
So the Behringer doesn't measure flat up to 20KHz?  Hmm...that stinks.  That was the mic I was going to purchase.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: BradJudy on 11 Jun 2009, 10:49 am
So the Behringer doesn't measure flat up to 20KHz?  Hmm...that stinks.  That was the mic I was going to purchase.

I just double-checked the cal file for mine and it measured +/- 1.1db from 20-20kHz.  As I said, the cal files I have compared were all pretty flat and not consistent enough to warrant the use of someone else's cal file. 

I hadn't noticed the extreme calibration used in these examples - definitely stop using the cal file and then see how it graphs. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: turkey on 11 Jun 2009, 01:49 pm
Now, about the falling top-end response - I do not know how well the mic is set-up on the tweeter axis. If it is off-axis much it will contribute to an even greater drop, so this is something to consider as well.

I always check my speakers in my room using an RTA with pink noise to see how the balance looks in-room. When the balance is very flat at one meter it will still show a significant loss in top octave energy at 8-10 feet, but it doesn't sound that much different, because at that distance we are hearing a greater proportion of the speaker's power response, which the mic does not pic up.

With most speakers, people don't normally listen on-axis, so the on-axis response doesn't mean much. (It's a very small part of the overall output of the speaker.)

What is far more important is the off-axis response and power response. What are these like in the various Salk models? I've seen some of the graphs on the web site, but they were only at one angle and weren't anything special anyway.




Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: srb on 11 Jun 2009, 02:17 pm
Speaker drivers are electro-mechanical transducers.

So are microphones, and quality among them varies greatly.  The average studio microphone costs between $1000 and $5000 (and more).

So I would be suspect of test results anyway without using the very best a very high quality microphone and A/D converter in the test rig.

Steve
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Art_Chicago on 11 Jun 2009, 02:40 pm
Now, about the falling top-end response - I do not know how well the mic is set-up on the tweeter axis. If it is off-axis much it will contribute to an even greater drop, so this is something to consider as well.

I always check my speakers in my room using an RTA with pink noise to see how the balance looks in-room. When the balance is very flat at one meter it will still show a significant loss in top octave energy at 8-10 feet, but it doesn't sound that much different, because at that distance we are hearing a greater proportion of the speaker's power response, which the mic does not pic up.

With most speakers, people don't normally listen on-axis, so the on-axis response doesn't mean much. (It's a very small part of the overall output of the speaker.)

What is far more important is the off-axis response and power response. What are these like in the various Salk models? I've seen some of the graphs on the web site, but they were only at one angle and weren't anything special anyway.






turkey;

Here is a cut&paste from Nels Ferre's review on ST dome tweeter speakers which can be found elsewhere.


Even more impressive, in my opinion, is the frequency response plot is still pretty smooth all the way to 15kHz, where it drops off rapidly. The high frequency extension at 60 degrees (!) off axis is phenomenal ? most manufacturers do not disclose this information, and those who do tend to disclose a more forgiving angle. All of the measurements of other loudspeakers that I have seen performed by the National Research Council of Canada, an independent testing facility, "only" go to 30 degrees off axis. Not only does this indicate an excellent design, but also that more people in the listening area will be able to hear most of the music, rather than those seated in the "sweet spot." The off axis response also indicates that the SongTower QWT should image quite well.

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: turkey on 11 Jun 2009, 03:06 pm


turkey;

Here is a cut&paste from Nels Ferre's review on ST dome tweeter speakers which can be found elsewhere.

Yes, the graph shown there at 60 degrees is nice-looking. I wonder what it looks like at other angles?

I wish that more speaker manufacturers and more reviewers would show detailed info on the polar response of speakers.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: turkey on 11 Jun 2009, 03:23 pm
Speaker drivers are electro-mechanical transducers.

So are microphones, and quality among them varies greatly.  The average studio microphone costs between $1000 and $5000 (and more).

Studio microphones are designed differently than instrumentation microphones, and are for different purposes. I don't know that it's useful to compare prices in this way.

Quote
So I would be suspect of test results anyway without using the very best a very high quality microphone and A/D converter in the test rig.

Steve

Good repeatable FR in an instrumentation microphone isn't that difficult to achieve. You can get mass-produced condensor mics that are like peas in a pod and are suitable for this use.

Flat FR in an A/D converter is quite easy to achieve too. We're not dealing with RIAA curve phono preamps here...

We're also not seeking a really high degree of precision in most cases. Differences between mics of say 0.5 dB or even 1 dB aren't going to be that noticeable since the response of most speakers will vary at least that much or more in production due to variances in drivers and other components. In other words, there's no point in designing the FR to 0.00001% tolerances if the drivers are going to vary by 10% or more in production anyway. Horses for courses.



Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 11 Jun 2009, 07:59 pm

I could graph his data for him if he send me the data file.  I've done that with other forum members in the past to show them things.   

I've send you a PM and if you send me your email address I can email my files to you.

I downloaded the ECM8000 cal file from Hometheatershack rew forum. and here is the link
http://www.hometheatershack.com/forums/downloads-area/19-downloads-page.html

Here is the part of data what i have in my cal file.
1000.0   0.00
1120.0   0.10
1250.0   0.12
1400.0   0.02
1600.0   -0.28
1800.0   0.11
2000.0   0.19
2240.0   -0.15
2500.0   0.21
2800.0   0.14
3150.0   0.08
3550.0   0.44
4000.0   0.67
4500.0   0.78
5000.0   0.89
5600.0   1.28
6300.0   1.78
7100.0   1.85
8000.0   2.55
9000.0   3.93
10000.0   4.46
11200.0   4.29
12500.0   4.73
14000.0   4.99
16000.0   5.35
18000.0   5.70
20000.0   5.37
22400.0   5.09
25000.0   4.48
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: TJHUB on 11 Jun 2009, 08:25 pm

I could graph his data for him if he send me the data file.  I've done that with other forum members in the past to show them things.   

I've send you a PM and if you send me your email address I can email my files to you.

I downloaded the ECM8000 cal file from Hometheatershack rew forum. and here is the link
http://www.hometheatershack.com/forums/downloads-area/19-downloads-page.html

Here is the part of data what i have in my cal file.
1000.0   0.00
1120.0   0.10
1250.0   0.12
1400.0   0.02
1600.0   -0.28
1800.0   0.11
2000.0   0.19
2240.0   -0.15
2500.0   0.21
2800.0   0.14
3150.0   0.08
3550.0   0.44
4000.0   0.67
4500.0   0.78
5000.0   0.89
5600.0   1.28
6300.0   1.78
7100.0   1.85
8000.0   2.55
9000.0   3.93
10000.0   4.46
11200.0   4.29
12500.0   4.73
14000.0   4.99
16000.0   5.35
18000.0   5.70
20000.0   5.37
22400.0   5.09
25000.0   4.48

PM'd you back.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: face on 14 Jun 2009, 06:14 am
It actually seems fairly obvious to me also, but unfortunately, I don't think my appraisal is going to agree entirely with TJHUB's conclusion, so you guys can do whatever you like with my comments here.

After reading through everything in the thread and looking all of the plots over I am seeing essentially what I would expect to see based on what I have read (with one exception, which I will get to in a minute).

First, regarding the mic calibration file, the Behringer ECM8000 is a fairly flat mic to begin with. Even to use it without a calibration file will not result in these apparent discrepancies, in fact I doubt you would notice the difference. Rather, I think the explanation is much simpler -

From what I gather, these are in-room response measurements measured at the listening location ( I did not see the actual distance specified, unless I missed it somewhere). If this distance is 8-10 feet then I would expect to see:

....room gain in the lowest octaves
....boundary cancellation and peaks due to the room response
....and a falling top two octaves due to directivity and absorption

And this appears to be exactly what we see. There is a difference between the two speaker, but he commented that he did not move them to the same location, so that can easily explain these differences too (at least partly).

Now, about the falling top-end response - I do not know how well the mic is set-up on the tweeter axis. If it is off-axis much it will contribute to an even greater drop, so this is something to consider as well.

I always check my speakers in my room using an RTA with pink noise to see how the balance looks in-room. When the balance is very flat at one meter it will still show a significant loss in top octave energy at 8-10 feet, but it doesn't sound that much different, because at that distance we are hearing a greater proportion of the speaker's power response, which the mic does not pic up. Again, what I see in these graphs correlates well with what I see in listening position measurements in my own room, with one exception.....

The dip at 2kHz in the one speaker should probably not be present in one and not the other. If his measurments were vertically off-axis (from the tweeter axis) then I could see picking up the beginning of a null in the lobing along this axis, but the top-end should be more rolled off if this were the case. Based on this, I tend to agree with Dennis, I think the tweeter in that speaker may be connected with the wrong polarity compared to the other speaker. Otherwise, I don't have much of any issue with what I am seeing here.

Jeff Bagby
Jeff hit the nail on the head! 

To the OP, dump Room EQ Wizard and download ARTA: http://www.fesb.hr/~mateljan/arta/
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Rocket on 14 Jun 2009, 06:55 am
Hi Audiocrazy,

Is there a problem with the sound of the songtowers in your system?  Is this the reason why you have taken measurements of the speakers?  Or were you trying to confirm that the speakers are okay after you replaced the tweeters?

Regards

Rod
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Scott F. on 14 Jun 2009, 01:59 pm
I didn't realize audiocrazy had two identical threads running at the same time.  :roll:

Here's what I posted in the other thread. It may be of help explaining why the speakers measure as they do.


<snip>....I have ....Integra DTC-9.8 with Audyssey.

Bingo. That right there is your problem. I have the 9.8 also. I tried Audyssey calibration with the supplied mic and what I heard coming out of my system sounded exactly like how your graph reads. Go back into the 9.8 settings and defeat the EQing completely. Set everything to flat. When you do that your speakers will come back to life, mine did.


...some additional commentary for this thread;

After hearing how bad the Audyssey "calibration" screwed up the sound of my HT system, I'll never use it again...but that's just my opinion.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 14 Jun 2009, 02:20 pm
After hearing how bad the Audyssey "calibration" screwed up the sound of my HT system, I'll never use it again...but that's just my opinion.

I have never played with Audyssey and do not know if there is a learning curve involved.  I did not want to mention any particular system by name, but I did allude to this in a post earlier in this thread.  I have run into this a number of times now.  In each case, I asked the owner to please turn this off and check the speakers with Audyssey out of the system.  In every single case, the "problem" of poor sound quality was solved.

Like I said, I have never worked with Audyssey before and it may indeed be a wonderful system.  But when on numerous occasions turning it off solved the problem, I do have my reservations.  It is obviously not a fool-proof solution and is certainly capable of doing more harm than good.

- Jim
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 14 Jun 2009, 11:52 pm
Turn all room EQ off, then re-measure and used the proper setup, calibration, speaker placement and room treatments to get the FR flat from 200Hz to 20KHz.  ;)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 24 Jun 2009, 07:24 pm
Turn all room EQ off, then re-measure and used the proper setup, calibration, speaker placement and room treatments to get the FR flat from 200Hz to 20KHz.  ;)
Sorry for the absence as I was caught up with work
Brandon i did all this but the same result and I suspect my mic rolls off quickly after 10KHz. I checked few threads over HT shack and there are people who has similar problem like mine. I'm not loosing sleep over it.

I've sent my mic for full professional calibration from 5Hz-25Khz. I'll re-measure after i get my mic calibrated.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 24 Jun 2009, 08:33 pm
Sounds good.  We look forward to seeing if that helps.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: fsimms on 25 Jun 2009, 02:14 am
Quote
I've sent my mic for full professional calibration from 5Hz-25Khz. I'll re-measure after i get my mic calibrated.

Remember to check on why it is showing a peak at 18HZ.  If you post that again Jim will have to raise the price of the SongTowers to over 10k then!  :lol: :lol: :lol:

Bob
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 25 Jun 2009, 04:56 pm
I honestly don't think calibrating the mic will make a significant difference.  You've got some crazy room node action going on there.  I'd guess your room is a little too dead in the HF's. 

Would it be possible to post pictures of your room?  Also, when time is available, experiment with speaker placement.  Start moving them a little and see what you come up with.  Remember, the MIC position is crucial.  It needs to be at ear level and unmoved during the entire measurement process. 

Also, do NOT use the HTS MIC calibration file.  It won't change much, so I'd try measuring without it (although, you said you sent it off for professional calibration, so I guess you can't do that.). 

Both tweeters roll off sharply at the same place, so it's either the MIC or the room in my opinion.  Just to be sure, though, have you tried swapping out cables (speaker wire), as a loss of resistance could be occurring?

Best wishes to you!
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 25 Jun 2009, 05:12 pm
I honestly don't think calibrating the mic will make a significant difference.  You've got some crazy room node action going on there.  I'd guess your room is a little too dead in the HF's. 

Would it be possible to post pictures of your room?  Also, when time is available, experiment with speaker placement.  Start moving them a little and see what you come up with.  Remember, the MIC position is crucial.  It needs to be at ear level and unmoved during the entire measurement process. 

Also, do NOT use the HTS MIC calibration file.  It won't change much, so I'd try measuring without it (although, you said you sent it off for professional calibration, so I guess you can't do that.). 

Both tweeters roll off sharply at the same place, so it's either the MIC or the room in my opinion.  Just to be sure, though, have you tried swapping out cables (speaker wire), as a loss of resistance could be occurring?

Best wishes to you!
My room is heavily treated with GIK acoustic panels. I took measurments without the panels but the same result and yes I also took measurement without the mic cal file. Nothing helped.
That's why i suspect that my mic needs calibration.
I'll remeasure with various combination after i get my mic calibrated.


Brandon here is the link to pics of my HT room. Any suggestions.
Front
http://www.flickr.com/photos/arshadmalik/sets/72157616172223502/
Side walls
http://www.flickr.com/photos/arshadmalik/sets/72157616172168822/
Rear wall
http://www.flickr.com/photos/arshadmalik/sets/72157616172074934/
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 25 Jun 2009, 07:09 pm
Thanks for the info. 

Well, your speaker positioning looks good (not to close to the boundaries).  You're right, you DO have a lot of acoustic panels...maybe too many.  But you said removing them made no difference (which they certain should make a difference).  That right there tells me there is something wrong with the mic or the calibration of the sound card you're using to measure.  Which soundcard did you use, and are you absolutely positive the calibration is solid? 

Looking at Terry's calibration of his Behringer mic, that dropoff makes me wonder...  In fact, it's the same dropoff that your measurements are showing.  Terry's doesn't drop off like that, but it makes me wonder if your .cal file was working properly, thus the measured response followed that sharp downward slope.

Finally, have you experimented with more toe-in?  Since you aren't sitting immediately on-axis of both speakers, there should be some natural roll-off in the HF's (typical in many rooms at that listening distance), but not that sharply.  Just for kicks, try toeing them in so they point directly at your ears.  Of course, don't do that until you've re-measured with your newly calibrated mic, as the mic may have been the problem and everything else will then be moot.   
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Rocket on 26 Jun 2009, 11:50 am
Hi Audiocrazy,

Is there a problem with the sound quality of the speakers?

Regards

Rod
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 26 Jun 2009, 02:45 pm
Hi Audiocrazy,

Is there a problem with the sound quality of the speakers?

Regards

Rod

Rod my speakers seems to be fine :D.
I took REW graphs to check how my room FR looks like to get the best out of my Song Speakers. And to figure out if there are any EQ issues that need to be addressed either by changing speaker placement or applying any acoustic treatments.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: avahifi on 27 Jun 2009, 12:22 pm
Then there is the easy way to do all this.

Just play a track of white noise, set the speakers side by side, have a friend run the balance control back and forth, move head just enough to stay as close as possible to on axis with each while listening. Any differences between the speakers will be obvious.

With a good speaker and drive electronics in a reasonably damped room, you should hear the sound of a rather large waterfall in the near distance  - - - like it it is ahead just around the next corner of the hiking trail.

Regards,

Frank Van Alstine

P.S.  This method will also quickly point out any significant peaks or dips in the response of two different speaker brands or designs,  the best "waterfall" wins.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 27 Jun 2009, 03:02 pm
Yup--it's old school, but it works. Unfortunately, most tuners don't seem to have defeatable interstation hiss controls anymore.  If I want to spot a frequency response issue quickly, I just tune between stations on my trusty Marantz 2100 and compare. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: avahifi on 27 Jun 2009, 03:27 pm
Actually I use a test CD with tracks I have compiled from several sources.

Several are white noise and various square waves I normally use for DAC testing, along with an all bits turned on 20KHz sine wave and a couple of zero dB tracks too for noise testing.

I can make you a copy if you wish, Dennis.

You will be amazed to see how bad most CD players and DACs are on the square waves, clipping and tons of distortion is the norm.

Regards,

Frank Van Alstine
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 27 Jun 2009, 06:36 pm

You will be amazed to see how bad most CD players and DACs are on the square waves, clipping and tons of distortion is the norm.

Regards,

Frank Van Alstine
Really?  Yikes! 

Can you recommend a good DAC that doesn't perform poorly on the square wave test?  I'd like to get one for my squeezebox eventually.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: zybar on 27 Jun 2009, 06:39 pm

You will be amazed to see how bad most CD players and DACs are on the square waves, clipping and tons of distortion is the norm.

Regards,

Frank Van Alstine
Really?  Yikes! 

Can you recommend a good DAC that doesn't perform poorly on the square wave test?  I'd like to get one for my squeezebox eventually.

What do you think he is going to recommend??   :scratch:

Here are some measurements for a stock Transporter:

http://www.stereophile.com/computeraudio/207slim/index4.html

George
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: avahifi on 28 Jun 2009, 03:11 pm
Nuance, all you need is a copy of my test CD and a decent dual trace scope.  Play the square wave tracks on a variety of CD players and DACs and note that the problems you will see are really nasty on many units.

No opinions or debate required about this after you make the tests.

Regards,

Frank Van Alstine
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 28 Jun 2009, 04:52 pm

You will be amazed to see how bad most CD players and DACs are on the square waves, clipping and tons of distortion is the norm.

Regards,

Frank Van Alstine
Really?  Yikes! 

Can you recommend a good DAC that doesn't perform poorly on the square wave test?  I'd like to get one for my squeezebox eventually.

What do you think he is going to recommend??   :scratch:

Here are some measurements for a stock Transporter:

http://www.stereophile.com/computeraudio/207slim/index4.html

George
His own stuff, I'd imagine.  I forget he offered DAC's.  DOH!  I'll be getting one of his then, down the road.

I cannot afford the Transporter, but thanks for the links to the measurements.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 28 Jun 2009, 04:54 pm
Nuance, all you need is a copy of my test CD and a decent dual trace scope.  Play the square wave tracks on a variety of CD players and DACs and note that the problems you will see are really nasty on many units.

No opinions or debate required about this after you make the tests.

Regards,

Frank Van Alstine
True true.  Well, I assume you engineer your DAC's so they don't have this issue, which is why I'll buy one of yours when finances permit.  Speaking of which, what is the main difference between the Ultra and Insight DAC?  I'll probably go with the Insight unit, but I'd at least like to know.  :)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: rahimlee54 on 28 Jun 2009, 05:29 pm
Nuance, all you need is a copy of my test CD and a decent dual trace scope.  Play the square wave tracks on a variety of CD players and DACs and note that the problems you will see are really nasty on many units.

No opinions or debate required about this after you make the tests.

Regards,

Frank Van Alstine

Is there somewhere we could get a copy of this test cd?  I'd like to try it out around my house.

Thanks
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Art_Chicago on 28 Jun 2009, 06:27 pm
Nuance, all you need is a copy of my test CD and a decent dual trace scope.  Play the square wave tracks on a variety of CD players and DACs and note that the problems you will see are really nasty on many units.

No opinions or debate required about this after you make the tests.

Regards,

Frank Van Alstine
True true.  Well, I assume you engineer your DAC's so they don't have this issue, which is why I'll buy one of yours when finances permit.  Speaking of which, what is the main difference between the Ultra and Insight DAC?  I'll probably go with the Insight unit, but I'd at least like to know.  :)

there was an AVA Insight+ DAC on A'gon for 650 or so. Lasted less than a day, so I missed it. OITP got T-8 DAC for that cheap from ebay.
In case Frank does not see this post, Ultra uses tubes for voltage and mosfet for current if I am not mistaken.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: avahifi on 29 Jun 2009, 05:51 pm
Now, if somebody out there who knows a lot more about computers than I do will help me, here is what I would like to do.

Upload my test CD (uncompressed) up here somewhere where anyone who wants to download it and burn their own test CD from it can do that.

I use a Mac computer (OS10.4) and this stuff is a bit more complex than my pay grade allows.  Very clear step by step distructions are required.

Regards,

Frank Van Alstine
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 29 Jun 2009, 08:01 pm
Now, if somebody out there who knows a lot more about computers than I do will help me, here is what I would like to do.

Upload my test CD (uncompressed) up here somewhere where anyone who wants to download it and burn their own test CD from it can do that.

I use a Mac computer (OS10.4) and this stuff is a bit more complex than my pay grade allows.  Very clear step by step distructions are required.

Regards,

Frank Van Alstine

Frank -

If you want to send me a copy, I can create an .iso file that can be used to burn copies.  I can then post it for anyone to download.

- Jim
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 29 Jun 2009, 09:32 pm
^ That would be great! 

Thanks to Jim and Frank. :)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: avahifi on 3 Jul 2009, 08:04 pm
Jim and Mary Salk visited me a couple of days ago so we could play the HT3s with the little Ultravalve 60 prototype amplifier.  It was interesting and great to see them in person again.

I did give Jim a copy of my test CD with the white noise and square wave tracks, along with some music tracks I listen to.  Jim says he will be able to post it on the web for any to download and burn test CDs for themselves.

I have to write some notes as to how to best use the CD and I will do that early next week.

Happy 4th.

Frank Van Alstine
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 22 Jul 2009, 05:53 pm
Sorry it took long but I've taken few measurements with the calibrated mic. I got the mic calibrated from 5Hz-25Khz and got narrow band measurement file for both 0 degree and 90 degree mic positions.
The 1st measurement is 0 degree and 2nd with 90 degree using both speakers. As you can see there is a huge variation in higher frequency. I'll post individual speaker measurements later.
The solid black line is from mic calibration file.
(http://www.audiocircle.com/image.php?id=20590)
(http://www.audiocircle.com/image.php?id=20591)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 22 Jul 2009, 06:08 pm
You aren't still using a calibration file from Home Theater Shack, are you?  Since you had the mic professionally calibrated, I'd imagine you don't need to use any calibration file with REW.

Anyway, the second graph looks better through the HF's.  That's 90 degrees off-axis?
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 22 Jul 2009, 06:09 pm
I'm still not quite following this.  What does the 0 vs 90 mic setting refer to? 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 22 Jul 2009, 06:13 pm
I'm still not quite following this.  What does the 0 vs 90 mic setting refer to? 
Wondering the same thing.  I am guessing axis?

If the HF's are better "off-axis," toe your speakers in more toward you, audiocrazy.  Then measure again.  Also, like I said, don't use any mic calibration file through REW.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 22 Jul 2009, 06:22 pm
I'm still not quite following this.  What does the 0 vs 90 mic setting refer to?
0 degree means the mic is pointing towards ceiling when taking measurement
90 degree means the mic is pointing towards speaker.

Brandon when the mic is calibrated you get a calibration file that you use instead of generic calibration file. Each mic has different freq response and when the mic is calibrated you get an offset value to the freq where the mic has a varation from the normal response.
Here are the individual measurements of L & R speakers with 90 degree mic position.
Blue is Right and Purple is Left
(http://www.audiocircle.com/image.php?id=20592)

Here are the individual measurements with 0 degree mic position
(http://www.audiocircle.com/image.php?id=20593)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 22 Jul 2009, 07:43 pm
OK   So it looks like we're making progress.  I assume from the deviations between the two speakers that you haven't moved the speakers next to each other for the individual measurements?
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 22 Jul 2009, 08:02 pm

0 degree means the mic is pointing towards ceiling when taking measurement
90 degree means the mic is pointing towards speaker.

Brandon when the mic is calibrated you get a calibration file that you use instead of generic calibration file. Each mic has different freq response and when the mic is calibrated you get an offset value to the freq where the mic has a varation from the normal response.
Gotcha.  Good to know, thanks.  :)

Well, technically you are suppose to place the mic at the listening position pointed straight up at the ceiling (zero degrees, as you called it).  This will be the listening position measurement, while the other can be considered an on-axis measurement (90 degrees, as you called it).  Based on that logic I'd say you need to toe the speaker in more.  How much toe-in were you using when you took the above measurements?

Obviously the graph below is the better of the two concerning high frequencies.

(http://www.audiocircle.com/image.php?id=20592)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 22 Jul 2009, 08:06 pm
Oops...double post.  Sorry
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 22 Jul 2009, 08:17 pm
OK   So it looks like we're making progress.  I assume from the deviations between the two speakers that you haven't moved the speakers next to each other for the individual measurements?
Dennis you are right. I have not moved the speakers to take individual measurements.
Would it be better if I take each speaker and place it in same exact location and take measurements individually or placing next to each other will suffice??

I think they are toed in approx. 30 degrees but i'll experiment a bit more.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: DMurphy on 22 Jul 2009, 08:53 pm
Just so there aren't any questions, I would move one to a spot in the middle, measure, move it back, and then put the other in the same position.  I would measure on axis with the tweeter. 
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 22 Jul 2009, 09:48 pm
Dennis, are you just having him rule out the possibility of a defective driver?  I assume that's the case.

Based on his recent measurements I highly doubt he's got a defective SongTower, but anything is possible.
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: fsimms on 23 Jul 2009, 11:54 am
Quote
The 1st measurement is 0 degree and 2nd with 90 degree using both speakers. As you can see there is a huge variation in higher frequency. I'll post individual speaker measurements later.

I see you are still showing a peak at 18 Hz!   Obviously the frequency measurements are inacurate.  I suspect the computer program is running too fast.  I don't have a clue how to fix that.  You could try running a different  computer.   Do you still have your old one?

Bob


Bob
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 23 Jul 2009, 03:14 pm
It could just be room anomalies. 

Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: audiocrazy on 23 Jul 2009, 03:32 pm
I see you are still showing a peak at 18 Hz!   Obviously the frequency measurements are inacurate.  I suspect the computer program is running too fast.  I don't have a clue how to fix that.  You could try running a different  computer.   Do you still have your old one?
Bob
Good point Bob.
I'm using HP netbook in my theater room with E-mu 0404 usb sound card. Unfortunately I don't have a spare laptop and my desktop only has on board sound card.
I may have to borrow a better laptop from a friend and retry the measurement
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: Nuance on 23 Jul 2009, 03:41 pm
The computer won't make a difference so long as you're using the same external sound card and calibration file.  Unless your computer is from the early 90's, of course.  :)
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: rahimlee54 on 7 Aug 2009, 11:33 pm
Now, if somebody out there who knows a lot more about computers than I do will help me, here is what I would like to do.

Upload my test CD (uncompressed) up here somewhere where anyone who wants to download it and burn their own test CD from it can do that.

I use a Mac computer (OS10.4) and this stuff is a bit more complex than my pay grade allows.  Very clear step by step distructions are required.

Regards,

Frank Van Alstine

Frank -

If you want to send me a copy, I can create an .iso file that can be used to burn copies.  I can then post it for anyone to download.

- Jim

I was wondering if we were any closer to getting this up somewhere.

Thanks
Jared
Title: Re: SongTowers FR Question (not flat from 1k-20khz) with RoomEQ
Post by: jsalk on 8 Aug 2009, 03:21 am
Sorry.  Frank gave me the CD in early July.  I just haven't done anything with it yet.  I will try to get it posted this week-end.

Frank was also going to put some documentation together on it, but I don't think he has gotten around to doing that yet either.

But I'll email him and remind him.

Thanks for the reminder.

- Jim