I'd think RightMark Audio Analyzer (RMAA) would be easier for that.
You could also download HolmImpulse (for free) and (i) make one unfiltered measurement of a driver, (ii) one filtered measurement of the same driver and (iii) use one of the built-in manipulation functions A-B=C for a presentation of the difference between the two measurements, i.e. the filter's transfer function. Just don't move the mic between the measurements. Assume you could export the result as an FRD file and run it through Phase Extractor to derive the filter's phase shift too:
http://audio.claub.net/software/jbabgy/FRM.html
You'd be up and running and done in 15-20 min altogether, faster than it would take you to figure out how to do it in ARTA
You could also download HolmImpulse (for free) and (i) make one unfiltered measurement of a driver, (ii) one filtered measurement of the same driver and (iii) use one of the built-in manipulation functions A-B=C for a presentation of the difference between the two measurements, i.e. the filter's transfer function. Just don't move the mic between the measurements. Assume you could export the result as an FRD file and run it through Phase Extractor to derive the filter's phase shift too:
http://audio.claub.net/software/jbabgy/FRM.html
You'd be up and running and done in 15-20 min altogether, faster than it would take you to figure out how to do it in ARTA
Last edited:
Provided it is only line level it is as simple as feeding your output signal to the crossover input and connecting the line-level output of the crossover to the input (where you would normally plug in your mic preamp out).
If measuring amplifier outputs you need to have a suitable resistive divider, and preferably some sort of robust buffer between the amp outputs and the sound card.
Here are some measurements I did of my active crossover using holm impulse doing as described. http://www.diyaudio.com/forums/analog-line-level/164886-synergy-active-crossover-4.html#post4003424
Tony.
If measuring amplifier outputs you need to have a suitable resistive divider, and preferably some sort of robust buffer between the amp outputs and the sound card.
Here are some measurements I did of my active crossover using holm impulse doing as described. http://www.diyaudio.com/forums/analog-line-level/164886-synergy-active-crossover-4.html#post4003424
Tony.
Last edited:
I had not thought of doing this in HOLMImpulse. That would certainly be easier than the method I've been using to try to match an active filter with a passive one. One concern I'd have would be that it seems like the latency in the measurement moves around a few samples from measuremenet to measurement. If it was too different from measurement A to measurement B, it seems like you could get some error in the calculation of A-B.
When I take measurements of raw driver responses to use for import into Frequency Allocator, I take an additional measurement with two of the drivers connected at the same time to get a summed response. When I compare that the measured summed response to summing the individual raw measurements, I always need to delay or advance one of the measurements a few loops to get the amplitude response of the sum to match the measurement. It's not immediately obvious to me how to get a check of A-B in the above case to make sure the calculation is accurate.
Dan
Last edited:
In ARTA, you'd have to pick dual channel measurement mode. Otherwise the time constant cannot be derived from the measurement setup. In single channel mode ARTA (and most other apps I presume) picks a rather random fixed time base at a certain percentage (10% I recall, but the manual points this out) on the rise of the total impulse response. That time base changes with the response characteristic of the DUT of course and thus gives you timing errors.
I do this kind of measurement all the time with ARTA in single channel mode. You need to be familiar with doing an impulse measurement in ARTA. I wrote a tutorial on this for loudspeaker measurements (Google: Charlie Laub ARTA impulse frequency response) that you might find helpful. The same idea applies for line level measurements, just without a mic.
Connect your line out to the crossover, and then its output to your line in. Next do an impulse measurement with the "sweep" (swept sine) mode selected (one of the tabs you get after you choose the play button when in impulse mode) and properly choose the time markers as shown in the tutorial. Finally, convert the gated impulse into frequency response.
Connect your line out to the crossover, and then its output to your line in. Next do an impulse measurement with the "sweep" (swept sine) mode selected (one of the tabs you get after you choose the play button when in impulse mode) and properly choose the time markers as shown in the tutorial. Finally, convert the gated impulse into frequency response.
I gladly refer to the ARTA manual, page 55 at the bottom.
Quick and dirty translated: single channel is fine if you're only interested in frequency vs. magnitude. If the phase is important, e.g. when you're designing loudspeaker crossovers, in single mode system (and DUT) dependent errors evolve and summing individual responses is not valid.
I gladly refer to the ARTA manual, page 55 at the bottom.
Quick and dirty translated: single channel is fine if you're only interested in frequency vs. magnitude. If the phase is important, e.g. when you're designing loudspeaker crossovers, in single mode system (and DUT) dependent errors evolve and summing individual responses is not valid.
Ah, not so fast to invalidate my approach my friend.
For designing loudspeaker crossovers I never use "measured phase", only minimum phase. It is very simple to use an interference technique to get the offset of acoustic center between drivers and that is all you need to design a crossover accurately. This has been spelled out by me in my ACD manual, and by others like Jeff Bagby, who is a well respected crossover guru. Google "find relative acoustic offset bagby" and you will find a white paper on the subject. This doesn't mean that other people's approaches (e.g. using measured phase) are invalid, but using min phase means that all of that time constant stuff doesn't matter. The minimum phase is simply extracted from the frequency response in the software using the Hilbert Transform. So, I would argue that single channel measurements are perfectly fine and really are all you ever need.
The OP did not mention phase, only frequency response. Still, for a line level measurement the minimum phase is plenty accurate.
I would not dare invalidating your approach. But lots of readers are unaware of the limitations of single channel measurement. And as you point out, you have to determine what the relationship is between two acoustic sources, be it by interference techniques or by dual channel measurement.
(and please do not start to point out that interference technique is more accurate... )
(and please do not start to point out that interference technique is more accurate...
)I would not dare invalidating your approach. But lots of readers are unaware of the limitations of single channel measurement. And as you point out, you have to determine what the relationship is between two acoustic sources, be it by interference techniques or by dual channel measurement.
(and please do not start to point out that interference technique is more accurate...
OK, thanks, I agree. That's fair enough for me!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.