Which would be the input for the mic and which for the reference input? It doesn't explicitly show that. Based on what I read, HOLM only uses one input for the mic, and one output for the amp.
Last edited:
There is no reference input. HOLM creates and outputs a test signal, records the resulting signal from the mic and analyzes the recorded data with respect to the test signal to determine the impulse response of the system. The "system" is everything between the D/A converter and the A/D converter - amplifiers, speaker, air path, room effects, microphone etc.
Correct. In that way HOLM is smart. It needs no reference input. Just a single input is all you get, all you need. HOLM, by default, removes delay and centers your impulse at time 0. You can change that or lock the time, but the default works fine for most things.
You can send more than one output, but it's all the same signal.
You can send more than one output, but it's all the same signal.
So I guess the impulse response is to find out the delay from the the moment the signal goes from the D/A then back to the A/D which will be used to calibrate out from the total delay to get the final minimum phase of the driver?
More than just that. The impulse response shows the round-trip time delay in the system, but it also completely characterizes the system. The frequency and phase response are obtained by taking the Fourier transform of the impulse. Because the impulse is the time-domain response, things like room reflections show up later than the primary impulse and can be masked out, at the cost of losing information about low frequencies. This allows you to measure, above a certain frequency, the anechoic system response in a non-anechoic environment. You will see that if you get HOLM working and play with it a bit.
No, you can't. HOLM is either a sweep or MLS noise. Given those options, I don't know why you would want to do single tones. You can limit to sweep range, if you need to. The calculations in HOLM are very good and very fast on either sweeps or MLS. You can filter after the measurement to see things in a way that works for you.
As SHC says, play around with HOLM a bit, you'll soon get the hang of it and understand what's going on.
As SHC says, play around with HOLM a bit, you'll soon get the hang of it and understand what's going on.
Thanks for all the responses. Another question I have is, although it's probably not critical, if there is no reference signal, there is no way HOLM can measure the actual amplitude at the input of the driver (or can it?)? Can the sound card measure its own output terminal voltage (with some losses on the cables) and HOLM will use that as the reference signal?
It's awkward to do absolute level measurements, partly because there are so many adjustable level controls in effect (e.g. HOLM output level, computer volume settings, amplifier settings, input gain settings). I use a meter or scope to measure the level going into the drivers. The calibration data for my measurement mic includes a nominal X-volts-at-Y-dB value. I loop the computer output back to its input and use a meter/scope to measure the level of the looped signal and figure out how recorded levels relate to sound pressure levels. That produces plausible-looking sensitivity values, but it's a pain to set up. Usually I don't bother because I really only care about the level into the drivers.
If You absolutely must measure absolute levels, take a high quality midrange with known freq response and sensitivity and use this as reference for absolute levels.
Lets say that the reference driver has a sensitivity of 90dB at 2 Khz at one meter. Measure it with Holm, and adjust the volume or mic so that the measurement lands on 0 dB. Now You know that 0 dB on the Holm graph means 90 dB when measuring Your own drivers.
Lets say that the reference driver has a sensitivity of 90dB at 2 Khz at one meter. Measure it with Holm, and adjust the volume or mic so that the measurement lands on 0 dB. Now You know that 0 dB on the Holm graph means 90 dB when measuring Your own drivers.
I gave up on trying to get holm to work in virtualbox or vmware. I don't think that the virtualization of the hardware works very well for getting clean audio. I admittedly haven't tried for a few years, so the later versions may be better.
Interesting that ARTA was ok. I found that I had problems with ARTA with pops and clicks running on physical hardware (and holm was ok)......
Tony.
Interesting that ARTA was ok. I found that I had problems with ARTA with pops and clicks running on physical hardware (and holm was ok)......
Tony.
I have been unable to get Windows installed on this imac so my only choice is to run it in VirtualBox. In any case, the response readings that I get from ARTA match well with the readings from REW run under OSX so I think the problem might not be with VirtualBox. I really would like to get Holm working.
Anyone has had any problems like this, even running natively?
Anyone has had any problems like this, even running natively?
I take it that the "response readings" in Holm do not look right (you didn't actually say if they did.) If they look OK then the pops are not a problem. If the pops are a problem then I would guess that the responses terrible and are useless. Just get a Windows platform, like a cheap the minibox PCs from Amazon (there are 100's of them, the Intels work best, but aren't the cheapest. Some are < $200, with the OS.) Holm is not very demanding on the hardware so these small PCs work just fine.