Measuring Raw Driver Acoustic Data - How to

[Loren42]

Are there any how-to white papers for doing this?

I would like to do this for my 2235Hs and import that data into Bass Box Pro/Xover Pro.

I assume that this is done with the driver in "free air", but there is going to be more to it than that, i would think.

[Mr. Widget]

This is the best resource I know of:

http://www.audioxpress.com/bksprods/products/bkaa45.htm


Widget

[Robh3606]

Hello Widget

I second you recomendation and add Floyd Tooles book as well. Great foundation material in there. As far as measuring T/S Bass Box has a really good section in it and there is info in the manual as well. Both the Clio Manual and the Leap Manuals are available for download and there are sections in those worth reading as well.

Are you talking T/S or in room or both?? For you to do a crossover you need in your box measurements for the best results.

Rob:)

[Ruediger]

In Thiele's paper which I published here a few days ago there is a procedure how to measure the TS params.

For the other params such as voice coil inductivity etc. You don't need a guide, right?

Ruediger

[Loren42]

Hello Widget

I second you recomendation and add Floyd Tooles book as well. Great foundation material in there. As far as measuring T/S Bass Box has a really good section in it and there is info in the manual as well. Both the Clio Manual and the Leap Manuals are available for download and there are sections in those worth reading as well.

Are you talking T/S or in room or both?? For you to do a crossover you need in your box measurements for the best results.

Rob:)

Dickason's book does a pretty good job of explaining the T/S parameter harvesting. However, when it comes to BB Pro there is an option to load the raw driver acoustic data.

I am thinking that the raw driver is supported in some form of jig and the near field frequency plot is taken in free air. That would be my assumption, but I don't know.

As far as in-box measurements go, I can see that, too, but figured that the acoustic data for a free air driver would be the default standard.

I am alos looking for where the cone breaks up and how.

I will read the suggested links everyone left me. Thanks!

[Mr. Widget]

I am thinking that the raw driver is supported in some form of jig and the near field frequency plot is taken in free air. That would be my assumption, but I don't know.Even mounting a woofer on a small baffle will affect the measurements. For measuring Fs and getting free air impedance etc. along with the other TS parameters using CLIO, I always suspended the woofers on a steel cable. with a meter or more of free space around it on all sides. If it sits on a surface or is need a reflective surface it will affect your results.


Widget

[Loren42]

Even mounting a woofer on a small baffle will affect the measurements. For measuring Fs and getting free air impedance etc. along with the other TS parameters using CLIO, I always suspended the woofers on a steel cable. with a meter or more of free space around it on all sides. If it sits on a surface or is need a reflective surface it will affect your results.


Widget

Is that the same method to get the acoustic data for the driver?

Or am I missing something? Should the acoustic data be gathered with the driver installed in the target enclosure?

That would seem to be a catch-22 situation, no?

[4313B]

The Response tab contains the normalized acoustic response of the driver (it must be normalized to the predicted response of the driver). It is very important that the acoustic data entered for the driver not include the response of a test box or baffle. The acoustic data should include only the +/- variations which occur when the measured response differs from the predicted Thiele-Small response.


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm51.png

Why enter the driver's acoustic response? Because it can improve the accuracy of several performance graphs by showing how acoustic anomalies will affect the overall system.

There are two ways to enter the acoustic response. It can be manually entered using the individual "Level" input boxes for each data point or it can be imported from one of several measurement systems.

The preview graph in the upper right corner shows the shape of the response while you edit or import data. Notice also that the preview graph includes a "zero line" to help you adjust the overall level of the data so that it is normalized. A red vertical line is also provided to show what part of the response will be changed if the input box of the selected frequency is changed.

Level The normalized amplitude response in dB for each acoustic data point. The response of the driver can be manually entered or edited with the "Level" input boxes. There are 134 data points from 5 Hz to 20 kHz. Use the horizontal scroll bar to access the "Level" input boxes for all of the data points. Note: You do not have to enter a value for every acoustic data point. You can leave some blank and then use the "Interpolate" button later to calculate the missing values. The "Interpolate" button is explained later in this topic.

Normalize Options The acoustical data must be normalized to the predicted response of the driver. (Zero dB represents the predicted response of the driver as defined by its Thiele-Small parameters and, if present, its box parameters.) This usually involves a two-step process:

1 Adjust the overall level of the acoustic data so that the flat region of the response curve is level with the zero line in the preview graph. This can be accomplished with the second, third or fourth normalization options listed below.

2 Subtract the predicted response from the acoustic data. This can be accomplished with the first normalization option listed below.

Each normalization option is described next:

Use the predicted T-S response When this option is selected, the predicted response of the driver will be added to the preview graph with an orange plot line as shown below.


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm52.png

The predicted response should approximately match the shape of the acoustic response curve as shown in the left example in the illustration below. If it does not, then one of two problems exits: 1) either there is an error in the driver or box parameters, or 2) there is a problem with the acoustic data.

The predicted response includes the box response if box parameters are entered. Otherwise it represents the driver in an infinite baffle as shown above. The predicted response can be changed by changing the driver parameters on the Parameters (mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro. chm::/html/u075mx.htm) tab or by changing the box parameters on the Box Design (mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro. chm::/html/10i2kp1.htm) tab of the Box Properties window. Important: The box parameters should duplicate the test box used when the driver was measured. Note: An infinite baffle can be simulated by selecting a closed box and using a huge box volume (Vb) and huge QL value. If necessary, remember to restore the driver and box parameters after normalizing the acoustic response.

When the "Normalize" button is clicked, the difference between the predicted response and the measured acoustic response will be calculated. This "difference" will be entered as the "normalized" acoustic response. The illustration below shows the same acoustic data before and after normalization:


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm53.png

If desired, this method of normalization can be limited with F1 and F2. Use F1 to prevent low-frequency data from being normalized. Use F2 to prevent high-frequency data from being normalized. For example, enter "500" into F1 if you do NOT want acoustic data below 500 Hz to be normalized. Enter "7000" into F2 if you do NOT want acoustic data above 7 kHz to be normalized. Either or both F1 and F2 can be left blank if you do not want to limit normalization.

Tip: We do not recommend that you leave the "Use the predicted T-S response" option selected after you have finished normalizing the acoustic data because it can slow the operation of the program on some computers. This is because BassBox Pro recalculates the predicted response every time a change is made to a driver or box parameter. Instead, we recommend that you select either the second or third normalization option before you leave the Response tab after you have finished preparing the acoustic data.

Use the driver sensitivity This normalization option does not change the shape of the acoustic response like the first option. Instead, it adjusts the overall level of the acoustic data. For example, an acoustic measurement may represent a 1-watt, 1-meter sound pressure level. In cases like this, the acoustic response may not be visible in the preview graph because the acoustic level is too high. For example, the flat portion of the response may be between 85 and 95 dB. Use this normalization option to adjust the level of the acoustic data so that it is level with the zero line in the preview graph. This is best done before normalizing the data to the predicted response.

Depending on the nature of the acoustic data, it can sometimes be normalized to the zero line with the driver sensitivity rating. However, any level can be used---even a negative level if the acoustic data is below the zero line. Let's examine an arbitrary example:


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm54.png

In the above example, the flat region of the acoustic data is about 15 dB above the zero line as shown on the left. To correct this, we replaced the driver sensitivity value with "15" and clicked on the "Normalize" button. The result is shown above in the example on the right. The value entered in the "Driver sensitivity" input box is subtracted from all of the acoustic data points.

Use an averaged band This normalization option does not change the shape of the acoustic response like the first option. Instead, it adjusts the overall level of the acoustic data like the second opton. It does this by subtracting the average level of a selected frequency band from the acoustic data. The start and end of the frequency band is entered into F1 and F2 as shown below:


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm55.png

In the above example, F1 = 150 Hz and F2 = 300 Hz so the acoustic data from 150 to 300 Hz will be averaged and this level will be subtracted from each acoustic data point when the "Normalize" button is clicked.

Use a single level (not shown) The final option for normalizing the acoustic data is not listed in the "Normalize Options" box. Like the second and third options, it adjusts the overall level of the acoustic data without changing the shape of the response. It is executed by double-clicking on any of the "Level" input boxes. When you do this, a message box will appear to confirm that you want to normalize the acoustic data by subtracting the selected level from each acoustic data point.

Restore The last three acoustic response curves are remembered so that you can "undo" changes when you attempt to normalize the acoustic data. Click the "Restore" button to return to a previous response curve. Clicking the "Restore" button again after the oldest curve has been restored will return you to the most recently stored curve. An acoustic response curve is stored into this "undo" buffer when either the "Interpolate" or "Normalize" button is clicked. The "undo" buffer is not affected by changes to individual acoustic data points with the Level input boxes.

Interpolate Clicking the "Interpolate" button will cause BassBox Pro to search the acoustic data for all zero values. It will then replace the zero values with estimated values by calculating the slope of surrounding non-zero data points.

Normalize Click the "Normalize" button to execute the selected normalization option. These options are described above and usually cause the level of the acoustic data to be adjusted to either a predicted response or to the zero line. Note: The "Normalize" button has one special feature when the "Use the predicted T-S response" option is selected. This feature is activated by holding down the Shift key when the "Normalize" button is clicked to cause the predicted response to be added to the acoustic data rather than subtracted from it as is normal.

Omit Driver Response in Graphs Check this checkbox to cause the graphs to ignore the driver acoustic data. This can also be controlled with a checkbox in the Graph Properties window and directly from the graph window by right-clicking on the graph and turning off "Include > Driver Acoustic Response" in the popup menu.

Import Acoustic Data Frequency-domain data from several popular measurement systems (including Brüel & Kjaer, CLIO, IMP, LMS, JBL/SIA SmaartLive, MLSSA, Sample Champion and TEF®-20) can be imported with the "Import Acoustic Data" button. Clicking it will open the Import Acoustic Response File window so you can select the data file. Several of the file types are identified by the file name extension. Use the "Files of type" drop-down list to select the file type as shown below:


http://audioheritage.org/vbulletin/mk:@MSITStore:C:\Loudspeakers\HT%20Audio\Bbx6pro.c hm::/images/bm56.png

If a file is chosen whose extension is not recognized, BassBox Pro will open a Select Data File Type window so you can select the file type from a list.

The acoustic data import filter works in the following manner:

• The import filter will interpolate between data points if some are missing.
• The import filter will average data points if there are extra ones.
• The import filter will give you the option of extrapolating the beginning or end of the data if it begins at a frequency above 5 Hz or ends below 20 kHz. The slope between the first or last five data points will be used if you choose to extrapolate.
• The import filter for some file types will automatically adjust the overall level of the data so that it will be visible in the preview graph.

Note: In most cases, the data will need to be normalized after it has been imported. Please read the preceding instructions about normalization.

Caution: Most acoustic measurements do not include accurate data at very low frequencies. For example, many measurements are grossly inaccurate below 100 Hz. The lowest usable frequency of the measurement will depend on the measurement's frequency resolution, "time window" and the location of nearby reflecting surfaces. These problems can sometimes be overcome with special measurement techniques like the "near field" technique. If the low-frequency portion of your acoustic data is not accurate, you should enter zeros for that portion of the acoustic data after the rest of the data has been normalized.

Clear Use this button to set all "Level" settings to zero dB.

Note: The acoustic response of a driver does not have to be entered in order to complete a speaker design. Most manufactures do not provide it, leaving it up to you, the designer, to measure it yourself.

[Loren42]

The Response tab ... Most manufactures do not provide it, leaving it up to you, the designer, to measure it yourself.[/I]

Thanks. That plugs a few holes in my knowledge base.

I guess the crux of the biscuit is how do I actually measure that acoustic data of my driver? There must be a procedure to do it, but I don't know what is the best way to get it.

Then once I get it I need to put it into a format that friendly for BB Pro.

I have looked a Smith & Larson's Woofer Tester web page. Very nice device and it looks like I could use that tool to capture the data, but I did not see a procedure or best practice to use to capture the driver response.

[herki the cat]

I have looked a Smith & Larson's Woofer Tester web page. but I did not see a procedure or best practice to use to capture the driver response.

Loren42, driver frequency response has to be measured in an acoustic environment totaly free of boundary surface reflections, aka: like from the floor, The walls & ceiling unless you have an "RTA "pulse system" that that performs fast enough to elude room reflection ramp-up & propogation to the test microphone. The type of microphone employed also is critical if you are using a manual sweep frequency test signal source. I have no clues regarding your test equipment.

At your convenience, could you email me description of your test equipment & the procedure employed to measure your speaker-system frequency response.

cheers herki