I think there's some documents in this thread that will walk you through it. I think I mentioned previously to give VituixCAD a try if you can't get Jeff's spreadsheets going. VituixCAD can do the baffle diffraction sim as well as the splicing.
Let's just go over why we are doing this whole process. The far field measurement is intended to capture the baffle diffraction, but because it's in room it cannot contain any low frequency response, we have gated out the room reflections, and the low frequencies as a result.
The nearfield measurement includes the low frequency response without room interaction, but also without the diffraction interaction as well, so we must add that back in.
So we take a nearfield response with the diffraction added back in from a simulation (they're quite accurate esp at low freq), and splice it with the far field measurement and the result is a complete frequency response without any room influence, just like an anechoic chamber
As I am trying to follow this next step, i have a few questions. This is a MTM, but I have only one measurement of the woofer - the upper one. Can I do a baffle step with one, sum with the response to the summed port and NF Woofer response? Is this adequate? Does the Vertical placement of the drivers affect BS or horizontal is the most critical one?
I suppose for proper measurement, each individual woofers need to be measured, but is one woofer good enough?
Also, which software is preferable for simulating the bafflestep and merging with the summed port+woofer response? I don't see an option of merging the response in Diffraction and Boundary Simulator - do i merge in PCD?
This is Jeff's Diffraction and Boundary Simulator - Do I need to add in the room, or just the baffle diffraction is good for the XO modelling?
Baffle is 6.5 width x 20 height (4inch roundover at the top, 0.5 inch roundover at the sides). Driver radiating surface is 3.25 diameter, upper woofer is 15.5inch from the bottom.
I suppose for proper measurement, each individual woofers need to be measured, but is one woofer good enough?
Which brings up a related question of mine: How does one correctly measure one woofer far-field when two or more woofers share the same cabinet space? As was previously mentioned, the unused woofer acts as a passive radiator which affects the response. Plus, the single woofer is radiating into the entire cab volume by itself which also affects the response. Perhaps these are both near field effects that don't factor into the gated far field response?
I suppose for proper measurement, each individual woofers need to be measured, but is one woofer good enough?
Which brings up a related question of mine: How does one correctly measure one woofer far-field when two or more woofers share the same cabinet space? As was previously mentioned, the unused woofer acts as a passive radiator which affects the response. Plus, the single woofer is radiating into the entire cab volume by itself which also affects the response. Perhaps these are both near field effects that don't factor into the gated far field response?
Hi Ed
These are the response for the dual woofer. The first one is the port response with both woofer - the response is as expected. When the port measurement is taken with only one woofer connected - it goes wonky.
Same for the NF measurement. When only one woofer is connected and measure NF, the dip is not there, but when both woofers are connected, but only one woofer is measured, the dip is there - but the top end response varies i suppose due to some response bleeding in to the measurement from the other response?
Port response with both woofers connected (Blue), Port response with only the upper woofer connected
This is the image from earlier, when only one woofer is connected, there is no dip, with both woofers, connected there is the dip in the Woofers Near Field response (the green is with both woofers connected, but measured at the top woofer)
So what do we do about the far field - measure both woofers on the tweeter axis, but reduce by 6DB (if each woofer is to be entered individually in PCD) or measure with only woofer connected and have the PR issue with the other woofer.
But would it matter if only one woofer is connected if the splice is around 400 hz? But what about a TMWW? and suppose i am crossing around 150hz or so? Does side firing require any special consideration in PCD or measuring setup?
When I simulate, I give the 2 different drivers individual BS adjustment. The top will have an affect on the response.
To measure, you really need to fill the hole with a plate for the second if you want individual profiles. Or- you could try shorting the second woofer, electrically isolated/not connected to the circuit, and it will impact the signal much less. How much less is the question, as it will still move.
As I am trying to follow this next step, i have a few questions. This is a MTM, but I have only one measurement of the woofer - the upper one. Can I do a baffle step with one, sum with the response to the summed port and NF Woofer response? Is this adequate? Does the Vertical placement of the drivers affect BS or horizontal is the most critical one?
I suppose for proper measurement, each individual woofers need to be measured, but is one woofer good enough?
Also, which software is preferable for simulating the bafflestep and merging with the summed port+woofer response? I don't see an option of merging the response in Diffraction and Boundary Simulator - do i merge in PCD?
This is Jeff's Diffraction and Boundary Simulator - Do I need to add in the room, or just the baffle diffraction is good for the XO modelling?
Starting with your first questions. I'm going to assume because this is not a center channel you are OK with sum as one woofer measurement.
The top woofer FR far field measurement and then the baffle step estimate using the top woofer dimensions works best in my opinion to get a good match in the Blender.
Use the Blender to blend and add the baffle step to the nearfield. Response modeler does only a hard splice.
You don't need need the room response for x-o work. However, I try to measure (far field) in place where there are boundary considerations like a center channel, near a back wall, or woofer near the floor.
Baffle is 6.5 width x 20 height (4inch roundover at the top, 0.5 inch roundover at the sides). Driver radiating surface is 3.25 diameter, upper woofer is 15.5inch from the bottom.
In the Blender you can add a 4 inch round over. The purpose of the baffle step added to the near field is to get a good match and blend with the far field. imho until you start running into match problems you should try to keep it simple.
So what do we do about the far field - measure both woofers on the tweeter axis, but reduce by 6DB (if each woofer is to be entered individually in PCD) or measure with only woofer connected and have the PR issue with the other woofer.
But would it matter if only one woofer is connected if the splice is around 400 hz? But what about a TMWW? and suppose i am crossing around 150hz or so? Does side firing require any special consideration in PCD or measuring setup?
Yeah we're back to a sum as one question. For MTM sum as one then use a TM in PCD woofer Y=0. TMWW = a 3 way TMW with the woofer sum in the middle in PCD. Other than Vituix Cad nothing will tell you what happens off axis with a side firing woofer.
Got Blender to work. Proceeding with the one woofer (the top one) NF and FF to do the response and the Port response is scaled to match the ends.
Using the Diffraction modeler within Blender and using the FF Upper woofer response and the NF Port+Upper Woofer NF summed, this is what I get. The blending area SPL is matched (though how good of a match is required?)
I didn't do anything with the tails - does the tails look ok, or what should i adjust them to?
This is the Diffraction data in Blender
This is the Zoomed in view - Does the match look ok?
This is the tails. What am I supposed to target - Jeff's Blender Tutorial is somewhat silent / vague on the slopes for the tails.
Baffle is 6.5 width x 20 height (4inch roundover at the top, 0.5 inch roundover at the sides). Driver radiating surface is 3.25 diameter, upper woofer is 15.5inch from the bottom.
In the Blender you can add a 4 inch round over. The purpose of the baffle step added to the near field is to get a good match and blend with the far field. imho until you start running into match problems you should try to keep it simple.
The round over seems to be applicable all around. I have a 0.5" rounder on the sides, none on the bottom and 4" on the top.
Also for non rectangular shaped baffles, i need to get Edge agree with Blender, so i can export from edge and import to blender to spply for non rectangular baffles.
In Edge, move the mic in front of the speaker. You'll find that it doesn't affect the response behaviour much from 0-1kHz which is all you should really be caring about (file was merged at 400Hz right?)
Tails...I think Step 4 Reads "Set the slope of the response tails to model asymptotic LF & HF behaviour". This means set the tails to match the slope of your response so the line continues on to infinity. This will improve the accuracy of the minimum phase extraction.
Better way? Everything is relative to what you're preferred workflow is. I'll say again...VituixCAD will do everything apart from the actual measurement if you want to stick with one software package for everything.
Holm and REW can both lock the timing reference, which is actually advantageous as it avoids the whole 3 measurement workaround to determine driver offsets, and allows you to measure all drivers on axis for greater accuracy of the simulation .
T= time and 0= whatever point in time you want it to be. It allows you to lock the impulse window start at a specific time to provide similar results to a 2 channel measurement. You lock the timing for the window start on the driver closest to the mic, usually the tweeter, and subsequent measurements will include the excess phase from the extra distance from the mic. In this way, you don't have minimum phase data that you have to measure and add a delay to, but the delay is built into the measured data from the start.
In Holm there's a button that says time lock or something to that effect.
In VituixCAD, with every response file you bring in there is a checkbox at the bottom of the screen for "minimum phase" and a delay value that you can enter manually, so you certainly don't need the "t=0" timing referenced data in all your files, it just makes for better off-axis data if you do measure all driver files completely on-axis, not with everything measured at the tweeter axis.
In the merger tool as well, there is a checkbox to export the data as minimum phase as well, otherwise it will retain the phase of the far field measurement.
I also notice VituixCAD has a "time align" screen where you can take the standard 3 measurements for determining offsets and you can determine the offset here by adding delay manually, or it has a "solver" that attempts to determine the delay on its own.
What is the t=0 timing and how and where do you lock it in Holm?
IIRC HolmImpulse users had problems with the time lock feature not giving repeatable results. And when I used HolmImpulse the time lock value never matched the value that the three measurement method gave me. So I stuck with the three measurement method until I moved to REW and a dual channel measurement setup. YMMV.
Thanks Dcibel. I am not familiar with VituixCad. will try it out.
I am planning to take another set of measurements, will try out the t=0 timing referenced. Just so I understand the process and what i am supposed to do...
Thanks for the screenshot of Holms Response. Should i take measurement of each driver on it's own axis? So if I start with the tweeter
1. Measure on Tweeter Axis 2. Set the Time zero locked and click on use 3. Measure Tweeter again? on tweeter axis - this is step 5 in the Holms Impulse help - how do i know it works? 4. Measure upper Woofer on Upper Woofer axis 5. Measure lower Woofer on lower Woofer axis
No need to measure all three together? The NF still needs to be merged with the individual Woofer responses? So upper NF merge with Upper FF Time locked Woofer and Lower NF merged with Lower FF Woofer. Would the time lock affect the merge?
Does these tails look better? Both the HF anf LF are at 48db slope... but the graph is squished, so can't make out how well the tails join the response
Comments
On the Diff tab my preference would be to enter the coordinates for the top woofer. Then click the Use Diffraction data at the right.
On the LF tab load your new near field with the port.
On the HF tab load you far field
Go the the control tab and follow the instruction. Take a screenshot if you run into problems...
InDIYana Event Website
Wolfy's Photobucket
InDIYana Event Website
Wolfy's Photobucket
I suppose for proper measurement, each individual woofers need to be measured, but is one woofer good enough?
Also, which software is preferable for simulating the bafflestep and merging with the summed port+woofer response? I don't see an option of merging the response in Diffraction and Boundary Simulator - do i merge in PCD?
This is Jeff's Diffraction and Boundary Simulator - Do I need to add in the room, or just the baffle diffraction is good for the XO modelling?
CAD drawing of the side profile - the left side is the front.
Baffle is 6.5 width x 20 height (4inch roundover at the top, 0.5 inch roundover at the sides). Driver radiating surface is 3.25 diameter, upper woofer is 15.5inch from the bottom.
In process of settinh up my virtaul machine - if there is a copy of the Blend or the response modeler that works in office 365, please send me a link or upload in my Drive folder folder: https://drive.google.com/drive/folders/1KpSwd8eiFf5Kb46I_pThX29eCN0Mlbvb?usp=sharing
These are the response for the dual woofer. The first one is the port response with both woofer - the response is as expected. When the port measurement is taken with only one woofer connected - it goes wonky.
Same for the NF measurement. When only one woofer is connected and measure NF, the dip is not there, but when both woofers are connected, but only one woofer is measured, the dip is there - but the top end response varies i suppose due to some response bleeding in to the measurement from the other response?
Port response with both woofers connected (Blue), Port response with only the upper woofer connected
This is the image from earlier, when only one woofer is connected, there is no dip, with both woofers, connected there is the dip in the Woofers Near Field response (the green is with both woofers connected, but measured at the top woofer)
But would it matter if only one woofer is connected if the splice is around 400 hz? But what about a TMWW? and suppose i am crossing around 150hz or so? Does side firing require any special consideration in PCD or measuring setup?
To measure, you really need to fill the hole with a plate for the second if you want individual profiles. Or- you could try shorting the second woofer, electrically isolated/not connected to the circuit, and it will impact the signal much less. How much less is the question, as it will still move.
InDIYana Event Website
Wolfy's Photobucket
The top woofer FR far field measurement and then the baffle step estimate using the top woofer dimensions works best in my opinion to get a good match in the Blender.
Use the Blender to blend and add the baffle step to the nearfield. Response modeler does only a hard splice.
You don't need need the room response for x-o work. However, I try to measure (far field) in place where there are boundary considerations like a center channel, near a back wall, or woofer near the floor.
Using the Diffraction modeler within Blender and using the FF Upper woofer response and the NF Port+Upper Woofer NF summed, this is what I get. The blending area SPL is matched (though how good of a match is required?)
I didn't do anything with the tails - does the tails look ok, or what should i adjust them to?
This is the Diffraction data in Blender
This is the Zoomed in view - Does the match look ok?
This is the tails. What am I supposed to target - Jeff's Blender Tutorial is somewhat silent / vague on the slopes for the tails.
Also for non rectangular shaped baffles, i need to get Edge agree with Blender, so i can export from edge and import to blender to spply for non rectangular baffles.
Or is there a better way?
https://drive.google.com/file/d/11HFVuXGKOWykJL9e86jKUf218441tUEy/view?usp=sharing
I am planning to take another set of measurements, will try out the t=0 timing referenced. Just so I understand the process and what i am supposed to do...
Thanks for the screenshot of Holms Response. Should i take measurement of each driver on it's own axis? So if I start with the tweeter
1. Measure on Tweeter Axis
2. Set the Time zero locked and click on use
3. Measure Tweeter again? on tweeter axis - this is step 5 in the Holms Impulse help - how do i know it works?
4. Measure upper Woofer on Upper Woofer axis
5. Measure lower Woofer on lower Woofer axis
No need to measure all three together?
The NF still needs to be merged with the individual Woofer responses? So upper NF merge with Upper FF Time locked Woofer and Lower NF merged with Lower FF Woofer. Would the time lock affect the merge?