HOLMImpulse: Measuring Frequency & Impulse Response

Yes, both. Or either. They should do the same thing.

Right, answered my own question just after I posed it, by exploring the ADC/DAC Calibration section of the second? tab.

How could I have found this out without asking? I realize there is no manual, and I started to read through all 50+ pages of this thread. Is there a HOLMImpulse forum? Or a knowledge base? Or some examples?

Regarding loopback sound card calibration, the one time ADC/DAC image cal load is fine for me (once I nail down how to use it). The ECM8000 to M-Audio preamp to UCA202 USB is clean enough. And one channel is sufficient(/perfect/preferred) for FR and IR measurments.
 
John

Make no mistake about it, and I know that you won't like this, but even one channel, HolmImpulse is the best loudspeaker measurement tool that I have used to date. Its all I use anymore. Sure, it could be improved, but it is still exceptional as it is.


It measure frequency response. What's the big deal, that it is free? I don't mean to sound sarcastic, but just what makes it so great?
 
It measure frequency response. What's the big deal, that it is free? I don't mean to sound sarcastic, but just what makes it so great?
Well, it is free, it is good depending on how critical one is about looking at data. Sometimes this is good help for getting people started in sound improvement. I had been happy with SpeakerWorkshop when I started out; but as I learned more, the desire for more bang for the buck came into consideration, and I moved to SoundEasy which pretty much is evolving continuously due to feedback from more experienced users. With consistent effort from the developer, I think it would be interesting to see how HOLMImpulse will evolve. It's always encouraging to see people making effort to produce better things.
 
Well, it is free, it is good depending on how critical one is about looking at data. Sometimes this is good help for getting people started in sound improvement. I had been happy with SpeakerWorkshop when I started out; but as I learned more, the desire for more bang for the buck came into consideration, and I moved to SoundEasy which pretty much is evolving continuously due to feedback from more experienced users. With consistent effort from the developer, I think it would be interesting to see how HOLMImpulse will evolve. It's always encouraging to see people making effort to produce better things.

Yes, I think the operative work is FREE. :) Certainly is seems easier to use than SpeakerWorkshop, but I don't think it is necessarily better as far as getting good data.
 
It measure frequency response. What's the big deal, that it is free? I don't mean to sound sarcastic, but just what makes it so great?

- Its free
- It uses a log sweeps which puts the nonlinearities at precise points on the impulse response as individual impulses - nothing else does this and its key to any view of nonlinearity that is going to mean anything
- its very easy to use
- its exports a whole set of impulse - such as at various off axis positions - into a single file - nothing else does that
- you do sound sarcastic
- and OHH did I mention that its free!
 
You know, I mis-spoke. That its free is irrelavent. I'd gladly buy it if I had to. I insisted that my Chinese client switch over to HolmImpulse since they were using old B & K and LMS stuff. Holm danced circles around both of those.
I agree that LMS has limitations, and that the key "old" is very critical. Lots of audio companies cannot afford new B&K equipment. Not that the new equipment is better, but that B&K seems to put more focus on noise related aspects rather than driver design related aspects.

- Its free
- It uses a log sweeps which puts the nonlinearities at precise points on the impulse response as individual impulses - nothing else does this and its key to any view of nonlinearity that is going to mean anything
...
I am quite interested in this point because even if it did put the nonlinearities at the precise points, it still takes more effort to allocate the cause of such nonlinearity in a way that an adequate driver fix can be predicted. I have never seen any paper on this subject just looking at impulse data. If any is available out there, I certainly would appreciate the information. Additionally, with only single channel impulse, how would one differentiate between nonlinearity in the driver, interaction between driver and amp, vs. driver. I don't believe there is any paper out there that addresses this either.
 
I have seens these before, and I browsed through them again. I see a different means of extacting data that is available via other commonly available means of measurement, but I still fail to see how it assists in further diagnosis of the DUT. Even if it provides more distinction under certain circumstances, is the cause due to complicated acoustic wave interaction? Is it due to the mic? or the DUT?

When I look at an impulse created using MLS test signal, I do see various effects as I tweak drivers. I can see reflections from mic clamp. Even with this kind understanding, there are still some areas where I hear differences, but cannot quite find the correlation in data. The method described in these papers desmonstrates some valid mathematical methods, but does not seem to write off what is caused by the measuring equipment. It also brings more uncertainties when only using single channel to do the measurement.
 
I am quite interested in this point because even if it did put the nonlinearities at the precise points, it still takes more effort to allocate the cause of such nonlinearity in a way that an adequate driver fix can be predicted. ... Additionally, with only single channel impulse, how would one differentiate between nonlinearity in the driver, interaction between driver and amp, vs. driver.

When one is dealing with any type of transducer, the nonlinearities in the transducer will swamp out those in the amp, unless its a very bad amp, which you should not be using in a measurement system.

You can check the amp with a simple load to be sure. The amp will tend towards higher orders of nonlinearity while the transducer will have very large lower orders.

And whose to say that the driver needs to be "fixed" because it has nonlinearities? There are papers on that and they imply that chasing nonlinearities in loudspeakers is a waste of time.
 
When one is dealing with any type of transducer, the nonlinearities in the transducer will swamp out those in the amp, unless its a very bad amp, which you should not be using in a measurement system.

You can check the amp with a simple load to be sure. The amp will tend towards higher orders of nonlinearity while the transducer will have very large lower orders.

And whose to say that the driver needs to be "fixed" because it has nonlinearities? There are papers on that and they imply that chasing nonlinearities in loudspeakers is a waste of time.
Interaction between the amp and the driver is still significant. Unless the amplifier has been verified under different types realistic nonlinear loading, I think the amplifier factor still needs to be taken out by using a reference channel. I have measured distortion right at the speaker terminals under mainy situations during one of the projects we've been working on. I have also seem many measurements Stereophile has made public. Basically the results have similar trends compared to mine.

I really think whether anything needs to be fixed or not is really a design tradeoff. But you have to have enough data showing what part is causing what kind of distortion before one makes that decision. Nothing is going to be perfect, just an engineering judgement as to whether it's worth fixing or not. No two engineers will have the same opinion. There are things that I can't determine through listening. But if data shows a certain part is producing higher levels of problems than other parts, that is surely an aspect to at least consider. If that improvement produces audible improvement, then we've learned one more thing. Whether it's worthwhile to make it into a product is a different issue. I'm sure there may be some forms of nonlinearity to be a waste of time to consider.
 
Last edited:
First you must PROVE that any of it even matters in the perception of the end product. In the case of loudspeakers and distortion, it doesn't. End of the need for any more data.
I like the improvement, implement it, and the customer likes it. To me, that is proof enough. Others can choose what they want to do. Life is too valuable to try and convince those beyond their will. :) If "chasing nonlinearities in loudspeakers is a waste of time", then "puting the nonlinearities at precise points on the impulse response as individual impulses" would be of no value either.
 
I like the improvement, implement it, and the customer likes it. To me, that is proof enough. Others can choose what they want to do. Life is too valuable to try and convince those beyond their will. :) If "chasing nonlinearities in loudspeakers is a waste of time", then "puting the nonlinearities at precise points on the impulse response as individual impulses" would be of no value either.

Yes, if you (think) that you hear something and that's good enough for you, that's fine. I'm more scientific than that. I don't believe something until it can be shown subjectively AND objectively. I'm not willing to except one OR the other.

I'd be interested in the nonlinearities more for electronics than loudspeakers because nonlinearities in electronics IS audible (loudspeaker nonlinearities CAN BE audible, but they can also be inaudible. In my speakers they are inaudible). But without having the distinct orders seperated out you can't tell the good from the bad for anything. Thats why HOLM is so valuable.
 
Administrator
Joined 2004
Paid Member
And another reasons its valuable is that it's free. Ooops!
Seriously, I tried many different measurement softwares - free, pay, share - and settled quickly on HOLM when it showed up.

It has really helped my measurements because it's both good and easy to use. Being both, it tends to spur further work, measurement technique and investigation. That's a good thing.