Rational Acoustics



Riley Casey
April 27th, 2011, 07:48 PM
OK maybe not a treatise but it seems to me that at some point over the years and various Smaart purveyors and their sites that there was a fairly direct and succinct paper, virtually a 'how to' on creating crossover settings for your generic cone low / horn driver high speaker system. Now as I have just taken possession for the first time in years of an inventory of custom boxes in need of a retuning I can't find that document. Does anyone remember that of which I speak or have a suggestion on where else to look? This could easily have been on the EAW site or even at a Smaart class which would mean it should be on my drive but it isn't.

Thanks for any pointers

Dr. J
April 29th, 2011, 12:43 PM
Hey Riley -- I see that you haven't gotten a response yet so maybe this little thread I started a while back may touch on your question. It is chocked full of recommended articles by some of the instructors here and may just be what you are looking for.

http://www.rationalacoustics.com/forums/showthread.php?t=621&page=4

Kip Conner
May 5th, 2011, 10:35 AM
I'm no professor of audio engineering (and would love feedback on this article as for verbage errors), but here's a case study that I did for a Clair Brothers 12am. I wish that I had some video to go with the screen captures so you can watch the phase trace move with the application of delay.

http://www.athenssound.com/CBA12am

However, there's no substitute for signing up for a class through Rational. It's a large part of their income as a business to train end users.

Dr. J
May 6th, 2011, 12:51 AM
Kip -- Awesome case study. I love reading stuff like this. Does the Omnidrive have allpass filters?

I often wonder about crossover points on drivers and how manufacturers decide what to use. For example: Lets say a manufacturer uses the same horn driver in their 12" woofer box as they do their 15" box.

Are they going to use the same XO point or is it going to be different? I was wondering because you went with a 1K XO point. Acoustical I believe. Was this just a educational guess or starting point? What exactly made you settle on that? If you don't mind me asking....... I want to learn.

Since I am an amateur at this -- I run into what I call "Mutt" systems all the time where stuff just gets thrown together and when I am asked for advice it seems like it is so late in the game that everything needs revamped. Some of these boxes have no tunings available OR aren't programmed into a system controller.

A while back on my "Bi-amp or Not thread --- Arthur was telling me a good place to set your crossover point..... well, here let me quote him:

"When you're searching for that ideal crossover frequency, pick a frequency where the coverage angles match between drivers, for the smoothest transition between drivers. Paying careful attention to the spectrograph will help determine this rather quickly."

What Arthur stated here makes a lot of sense to me in choosing XO points IF you have no manufacturer data to work from. Is there any other criteria to work from when selecting XO points that you can think of?

Dr. J
May 6th, 2011, 11:23 AM
Kip -- one other thing if you don't mind. You mentioned (not in exact words) that just because there is a potential flaw in a driver outside of it's XO point that you still need to consider correcting it.

I have heard of this before BUT I haven't personally tried it. Since I am limited on Parametrics -- I am always concerned about goosing a frequency (post PEQ) only to cut it at the input (PEQ). I guess in fear of wasting a PEQ. Is this where close observation of the phase trace will indicate to you if you have made the right moves?

On my phase trace (of my TOPS) I have a smooth phase trace BUT right around 3K it sharply ramps up briefly and levels itself back out. I know there may be an issue right there but I haven't been able to measure in quite a while due to winter BS here.

I have a number of possibilities to experiment when the weather clears.

Kip Conner
May 27th, 2011, 12:38 AM
thanks guys for the kind words and I apologize for the tardy response. I have been out of for the last 6 weeks with one foot on the gas and one hand on the knob. I'll try to address my thoughts as best as I can... I'm still figuring out some things as I go myself and make my share of non-fatal errors. I spend a lot of time just experimenting and am still having many "wow, ok, cool" moments since I wrote that case study.

1. I chose the 1kHz electronic crossover setting with a little bit of thought and a little bit of random. What I did knew (and confirmed with analysis) is that they JBL 2450 had a usable frequency response from 500Hz to 16kHz and that the 2206 12" inch driver was solid from 50Hz to 3kHz. At this point I was looking at a range of 500 to 3000 cycles of overlap. It should also noted that the 2450 rms wattage increases from 100 watts to 150 watts at 1kHz. Up to this point I have never crossed a HF driver that low, but knowing that since it was going to be used for vocal reinforcement from a vocalist that needs high SPL and nothing above 8kHz, I went lower to direct the power to the area where I needed it.

I have tuned "mutt" systems before where I had to use smaart to find out what any given speaker in the system could actually do- exactly what Arthur was talking about. You have to be quick and slow at the same time. Slow and methodical in the thought process, but quick with the actual testing. I've taken crossovers and set them up for what we call known ranges for a speaker and slowly moved the filter lower or higher (depending on which side I'm measuring) to see where the driver tends to stop being functional. As an example, if I didn't know the range of a HF driver: I would probably start with a crossover range from 2kHz to 16kHz and at the lowest volume I could (and maintain high coherence) I would slowly lower the 2kHz and pay careful attention to the transfer function to see how the driver behaves. It's important to set some parameters to serve as safeties so that you don't 100Hz to your HF drivers and destroy them.

2. EQ'ing a speaker out of the crossover range... I like to think about a speaker system has being unrelated. The 12" driver doesn't know the HF driver actually exists. they are doing their own thing until they reach our ears and we perceive them as being conscious of each other. when one driver is producing a hump in the slope as it moves into another drivers range it is essentially boosting that frequency in the other drivers range. Your choice to deal with it or not would depend on if that small bump is hurting you. You might welcome that bump if it occurs in an area where it's safe. I choose to deal with these because I needed to try to create as perfect of a phase alignment as I could for the testing.

3. Reversing the HF driver instead of the LF driver. Honestly I have never done it. I always started with tuning the HF drivers and work my way down. I think that I started doing it this way because I was working with 3 way systems and it's definitely easier to find the delay time from the a narrow frequency. Also, HF drivers tend to display a linear phase response from the 2kHz to 5kHz range.

Also, after I had captured the HF trace and began working on the LF trace it was easier to not move backwards. I suppose that there are other things to consider such as the relationship to the monitor to the mains and that I would rather try to keep all of the intelligibility frequencies in phase. Going back to the drivers not knowing the other one even exists it really shouldn't matter which one you flip and the decision might need to be based on an assessment of how the box will be used. Flipping the HF driver may be beneficial to in high SPL situations of vocal reinforcement.

I wouldn't call this a major delay in the low frequency driver, albeit healthy. I admit that it made me a little uneasy at first, but once i heard it in action I was ok with the result. I think that I would rather have that delay there than in the intelligibility range.

4. the 3kHz phase shift on Dr J's tops... again, I can't even begin to claim to be an all knowing source since I, too, am figuring some things out as I go. If the same phase shift occurs in all the mid-high cabinets I wouldn't worry about it. If you had one box that was behaving differently than you might want to address it. As long as they are all the same, they should be playing well with each other.

5. no all pass filters in the omnidrives :(

I welcome all comments, especially- "you're doing it wrong... here's why"

Ferrit37
May 27th, 2011, 10:33 AM
Hey Kip,
One thing you don't mention when selecting your x-over points is component spacing and how this affects the combined directivity of the box in the overlap region.

Dr. J
May 27th, 2011, 11:10 AM
Kip -- Good to hear back from you. I have been waiting for you to comment. Sounds like you made good decisions & if they sound great -- then it is great.

Yeah -- on my tops that 3k bump on the phase trace is on both of them. My tops are NOT bi-amped (soon to be) so I figured it was done by the manufacturer OR I am doing something wrong. I don't think so though..... I measured my monitors as well and they are not bi-amped either and found basically the same thing. Must be a FullRange internal XO thing.

On my 260 it has a section that allows you to alter the phase on the High outputs. I was thinking or hoping that this would allow me to achieve a flatter phase response BUT I couldn't change anything with it. It seemed to do nothing. I am told that this is the 260's version of an All-pass filter. Don't know.

Lately I have been measuring a different way from in the past and I have to say "What a difference this made". I do have to thank Dave Gunness for the insight and what a Smart guy...... Also Harry Brill, Jr for sending me to him. Thanks Harry!

I had been doing "Averages" measuring in several spots and shooting for that range. I thought my tops sounded pretty good BUT they were harsh at times. I now take 5 Snapshots Horizontally and 5 Vertically (different degrees) and display them all at once. This will tell you very quickly if a dip should be EQ'd or not. I was CLEARLY EQing dips that should have been left alone. If any dips occur ONAX that change OFFAX --- like go way up -- then they should probably be left alone.

This is the example Dave gave me: http://www.fulcrum-acoustic.com/wordpress/wp-content/uploads/2011/01/Prod-Spec-DX1565-v6.pdf

Look at the Vertical measurements and you can see a ONAX dip around 3k and from the other OFFAX measurements -- you can see why they left it.

I was goosing the ONAX by using averages. I guess the moral of the story is to not let any OFFAX measurements peak higher than the ONAX.

I also gradually roll of the top end from 8k on out as well. Although I am still very green with Smaart -- my system has never sounded better.

Can't wait to Bi-amp and do it all over again...

Dr. J
May 27th, 2011, 11:24 AM
Arthur -- Is there any other criteria you use to determine XO points on two drivers other than the Spectrograph? Well, besides specs.

Do you measure one driver at a time to see "Where" a possibility is? Or do you fire them both up and look? Kind confused on that.

One other thing -- Does measurement mic distance from the BOX influence this?

Arthur Skudra
May 27th, 2011, 01:28 PM
Arthur -- Is there any other criteria you use to determine XO points on two drivers other than the Spectrograph? Well, besides specs.

Do you measure one driver at a time to see "Where" a possibility is? Or do you fire them both up and look? Kind confused on that.
I don't have an objective criteria, I first do a level balance to get all the pass bands at the appropriate level, then look at where the response starts to dip down in the high end, or start "misbehaving" in the low end, and apply the appropriate filters to compensate for any strange anomalies that might be significant enough beyond the crossover point. Paying attention to each driver's coverage patterns in the spectrograph and the point where they naturally intersect in the magnitude response, I select an appropriate crossover point then work on getting timing alignment right between drivers using the phase curve. Many variables that you're manipulating at once, you feel like an octopus! :)


One other thing -- Does measurement mic distance from the BOX influence this?Distance is important, usually I tell my class that you need to position the mic at least 3 to 4 times the diagonal dimension of the front face of the box in order to be in the far field. Any closer than that, and you see one driver more than the other. Furthermore, you need to position the mic perfectly on axis between the drivers. Obviously with a 3 way box this can be rather challenging to pick the appropriate center point to measure from. A coaxial device is easy peasy!

Kip Conner
May 29th, 2011, 05:46 PM
I may be wrong with this, but I think the phase shift that you can manipulate is equivalent to the delay feature on the on the DriveRack 260. It seems to me that they are essentially the same thing. As you apply delay to the output, you are essentially changing the phase response. It's all related- even a high pass filter will shift your phase in a predictable manner.

Once you get into a Bi-amp situation and begin studying the drivers on a individual basis you will be surprised on how much you learn. As Arthur stated, the bi-amp situation is pretty easy when you get the hang of the basic principles. The tri-amp wouldn't be too bad if all of the components were in the same box. However we work mostly with bi-amp boxes stacked on a sub or even flown over a sub- making it harder to find a good reliable reference point. Whenever possible I try to measure the boxes before they take to the air if it's a box I'm not familiar intimate.

Thanks Ferrit for the addition... are you still running around with Preston?

Ferrit37
May 30th, 2011, 04:37 PM
Hey Kip,
No I'm actually running off to Denmark for a training session then off around euroland for a bit.
I love the part about looking at OFAX dips DaveG is god :)

Rasmus Rosenberg
May 31st, 2011, 02:41 AM
Where/When in DK?? PM me,

Dr. J
June 1st, 2011, 11:20 AM
I may be wrong with this, but I think the phase shift that you can manipulate is equivalent to the delay feature on the on the DriveRack 260. It seems to me that they are essentially the same thing. As you apply delay to the output, you are essentially changing the phase response. It's all related- even a high pass filter will shift your phase in a predictable manner.

Once you get into a Bi-amp situation and begin studying the drivers on a individual basis you will be surprised on how much you learn. As Arthur stated, the bi-amp situation is pretty easy when you get the hang of the basic principles. The tri-amp wouldn't be too bad if all of the components were in the same box. However we work mostly with bi-amp boxes stacked on a sub or even flown over a sub- making it harder to find a good reliable reference point. Whenever possible I try to measure the boxes before they take to the air if it's a box I'm not familiar intimate.

Thanks Ferrit for the addition... are you still running around with Preston?

Hey Kip -- Yeah I have messed with the Phase Shifter / Rotator in the Driverack 260 but with Fullrange tops I couldn't get it to do what I was hoping for. I had better luck nudging the delay locator and delay spinner.

Maybe if my tops were bi-amped I would have better results. It only applies to the high outputs on the DR which is where my Tops are routed to.

Does anyone know if Smaart 6 will go smaller than 2ms increments? Well -- I have punched in 1ms increments but I can't seem to go any smaller than that. I was shocked to see how much smoother the phase trace or should I say how much better position the phase trace responded to 1ms than 2ms.

Kip Conner
June 1st, 2011, 05:27 PM
At very least you should see the whole phase trace shift to the right as you alter phase value, especially when you approach the 180 degree mark. I never quite understood the purpose. Since it's not an all pass filter it should do the exact same thing as delay.

As for the "less than 2ms"... I'm assuming you are talking about the delay locator? Short answer, No. I think that it's represented in v6 as .02ms which would be smaller than you think. That would be 2/100th of a ms, right? Help me!

Dr. J
June 1st, 2011, 06:21 PM
Yes - the delay locator. I think you are right Kip. My bad....... I don't have my laptop in front of me to look right now. :D

Kip -- when you EQ past the XO point on your horn drivers -- how far out should you go? Seems like the -6dB point may be far enough. Not sure though.

PaulTucci
June 1st, 2011, 10:24 PM
The smallest increment that the offset delay can be incrementally moved is one sample rate. Because we mostly work at 48K SR, the time equivalent is 1/48,000 of a second, or .002 mSec. If a 96k Sample Rate were in use the incremental delay spinner would react in 1/96,000 of a second differences or .001mSec. The 96 K devices are more costly.

PT

Dr. J
June 1st, 2011, 11:02 PM
The smallest increment that the offset delay can be incrementally moved is one sample rate. Because we mostly work at 48K SR, the time equivalent is 1/48,000 of a second, or .002 mSec. If a 96k Sample Rate were in use the incremental delay spinner would react in 1/96,000 of a second differences or .001mSec. The 96 K devices are more costly.

PT

Thank you Paul!

DennisA
June 3rd, 2011, 11:34 AM
At very least you should see the whole phase trace shift to the right as you alter phase value, especially when you approach the 180 degree mark. I never quite understood the purpose. Since it's not an all pass filter it should do the exact same thing as delay.

As a test, a change could be made to the value of the "phase" function and then a measurement could be made to determine if delay had been added (or subtracted) as a result of this change.

PaulTucci
June 3rd, 2011, 10:50 PM
As a test, a change could be made to the value of the "phase" function and then a measurement could be made to determine if delay had been added (or subtracted) as a result of this change.

From page 38 of the user manual:

4.12 Output
The output section of the DriveRack 260 provides the user with the ability to control output levels of the unit and adjust phase compensation of loudspeakers within the signal path. The output level control is located in the signal path before the limiters; this means that the limiter
directly affects the output level. The parameters for the Phase/Polarity Compensation effect are as follows and are user adjustable.

Shift On/Off
This parameter is used to turn the selected output phase parameters on or off. When Shift is
off, phase is 00 and polarity is positive (+).
Polarity +/-
This parameter is used to reverse the polarity of the output signal.
Phase 0 to -175 Degrees
This parameter is variable in 5 degree increments and sets the amount of phase shift within the
selected output path.

From what this states Dennis''s observation should be true.
The amount of delay added is related to the upper frequency of the bandpass you're working in. ie, the mids and the xover point is 1000Hz, then the amount of delay added would be a maximum of half the time period of 1KHz. (180 degrees of 360 degrees of the 1000Hz) About half a millisecond.

The 5 degree increments of delay would measure 5/360 of 1mSec
5/360 x 1 = .01388 mSec.

Someone measure that for us?

PT

Kip Conner
June 5th, 2011, 06:32 PM
i guess I should also add that they only time that I have used a 260 I set it up so output 1 and 2 were the lows. Thus the phase shifter was assigned to the lows, not the highs. not that it makes a big difference.

by engaging and moving the phase you will get a change, but it will be across the entire pass band.

What I am trying to figure is why the manufacturer put both the phase shift option and the delay option in that set of outputs. Furthermore why not put that feature on all the outputs- looks like it would only make it confusing to the end user.

DennisA
June 5th, 2011, 09:49 PM
What I am trying to figure is why the manufacturer put both the phase shift option and the delay option in that set of outputs. Furthermore why not put that feature on all the outputs- looks like it would only make it confusing to the end user.

The Phase Shift function is on all 6 outputs...I just double checked

Dr. J
June 10th, 2011, 11:16 AM
The Phase Shift function is on all 6 outputs...I just double checked


Yes it is & I can't believe I overlooked that.....

DennisA
June 11th, 2011, 11:09 PM
And the Phase Shift is an all-pass filter but DBX is hesitant to call it that because of it's location on the output of a crossover pass-band rather than a full range source.

Dr. J
July 6th, 2011, 03:45 PM
I will post here since this may be a continuation of the Phase Shifter in the DR 260.

My internal crossovers are going back to the factory for repair and I have made the move to Bi-Amp Mode. The Phase Shifter works very well on the DR 260 & for some reason with the Top in Full Range mode -- it doesn't seem to do much.

Since the Shifter is on the "Upper" frequencies of the bandpass you are working with -- it really has an effect on the phase trace. I was able to apply a certain degree of offset which I believe is nothing more than applying delay to that part of the phase response -- and actually make the phase trace go from climbing upwards (from left to right) to laying nice and flat.

So...... I don't want to do anything stupid here but adding delay into the phase response in this manner adds delay into the overall system right?

Let me say it this way: McCarthy explains in his book that "ALL PASS Filters" basically shouldn't precede anything else. They should be more or less used as a last step to -- not in his words -- "to fine polish" the system.

IF I were to do this in the upper frequencies of my mid speaker AFTER I aligned it -- wouldn't my alignment then be off again?

I know there is an order with this stuff so you aren't twisting it back up as you go. Oh -- I know the 260 doesn't have what is defined as a true All Pass Filter.

Any comments??

DennisA
July 7th, 2011, 11:17 AM
IF I were to do this in the upper frequencies of my mid speaker AFTER I aligned it -- wouldn't my alignment then be off again?

After this topic came up, I spoke with Jim Ure at DBX. He defined the "Delay" control as coarse adjustment and "Phase Shift" as fine tuning. According to that line of thought, alignment with the delay control would be incomplete until fine tuning was applied with the "Phase Shift" control, so no, your alignment "wouldn't be off again". Keep in mind, this fine tuning would probably be affected by moving the mic six inches but that is a whole different conversation.

Kip Conner
July 10th, 2011, 07:40 PM
I have a buddy who owns one of these 260's. I'm going to have to borrow it and see what is up with these features! Something about it seems counter intuitive to me. I suppose that the 260 is supposed to be their flagship model so it will have all the bells and other bells. :)

Dr. J
July 11th, 2011, 11:14 AM
I have a buddy who owns one of these 260's. I'm going to have to borrow it and see what is up with these features! Something about it seems counter intuitive to me. I suppose that the 260 is supposed to be their flagship model so it will have all the bells and other bells. :)


Hey Kip -- That would be cool if you could do that. Since I bi-amped my tops -- you can certainly see the change.

It seems strange to me to purposely move a phase response of a driver for example 120 degrees BUT then again do drivers like a horn and a mid ever fall exactly together in phase from the start? It is either a horn inverted or normal polarity gets you closer to matching the mid speaker in phase -- that seems pretty easy to understand and then the "phase Shifter" allows you to make smaller or finer adjustments to the phase response.

Maybe I could post some screen shots and demonstrate this and maybe some of you on here can comment whether or not it is a good move or a destructive one. In any case -- DELAY is being added to the system.

For me -- the phase thing is the hardest for me to wrap my head around. I mean I think I know what the goal is -- just making the right move to do it where I don't screw it up is another thing....:D