HOLMImpulse: Measuring Frequency- & Impulse-Response

[gedlee]

Askbojesen

I have not tried it, but before I do, I have some questions.

When the log sweep is used does the impulse response that is calculated from the cross-correlation yield the impulse response of the higher harmonics preceding the main impulse as the theory says it should?

Is the signal file that is cross-correlated with the measured data corrected to give the proper phase and amplitude as is required per Farina?

Can the impulse response be written out to be read by another program? What about OLE?

Is 64 bit VISTA a problem?

In any case thanks for this valuable addition to the technology. Hopefully more people will do measurements and quite making decisions "by ear".

[Dr.EM]

I decided to test my soundcard, since I don't have my mics or speakers here. It's evidently not great! Creative Audigy thing. It'll work for frequency response measuring though, which is really what I need

[askbojesen]

Originally posted by Dr.EM
I decided to test my soundcard, since I don't have my mics or speakers here. It's evidently not great! Creative Audigy thing. It'll work for frequency response measuring though, which is really what I need

Your soundcard is out of phase with itself. Check of the "Invert recorded signal" under device settings. At the impulse response it is very clear that the impulse is going down instead of up.

[Dr.EM]

Ah yes, that's much better! Thought the phase was a bit odd

[askbojesen]

Originally posted by gedlee
I have not tried it, but before I do, I have some questions.

When the log sweep is used does the impulse response that is calculated from the cross-correlation yield the impulse response of the higher harmonics preceding the main impulse as the theory says it should?

Yes that is how I calculate 2-9th harmonics and sum them to THD
I have tested the magnitudes with a normal steady state sinus + FFT afterwards.

Is the signal file that is cross-correlated with the measured data corrected to give the proper phase and amplitude as is required per Farina?

Yes I know that a logsweep needs an inverse crosscorrelator (Unlike MLS sequences, that are they own inverse)

Can the impulse response be written out to be read by another program? What about OLE?

The measurements can be exported to text file in custom format (See attachment)
OLE - No.

Is 64 bit VISTA a problem?

If VC++ crosscompiler works then no. I have crosscompiled but not yet tested on 64bit HOLMImpulse now with 64 bit

In any case thanks for this valuable addition to the technology. Hopefully more people will do measurements and quite making decisions "by ear".

I agree that more people should measure more - They would be able to pinpoint what they like / dislike. We would have more quantities to talk about instead of just words like smooth/tight/warm/hard/cold/sloppy,... Ore maybe link those words to quantities

[gedlee]

Originally posted by askbojesen

Yes that is how I calculate 2-9th harmonics and sum them to THD
I have tested the magnitudes with a normal steady state sinus + FFT afterwards.


Yes I know that a logsweep needs an inverse crosscorrelator (Unlike MLS sequences, that are they own inverse)


The measurements can be exported to text file in custom format (See attachment)
OLE - No.

You misundertood a lot of the questions, so lets start again.

Does the program export the impulse response WITH the pre impulses of the harmonics? If it does this then I am very interested, if it doesn't then I am not. Frequency response with THD is not very interseting.

Farina talks about several things that have to be done to the "inverse crosscorrelator " to get rid of some artifacts. Its all in his latest AES paper. Do you do these things?

[BDP]

Any sound card? I selected my sound card and then selected measure and my computer rebooted . The computer I tried it on is 5 years old with a SB live sound card running XP. Any tips?

BDP

PS: You are working with a 1/2 geek!

[askbojesen]

Originally posted by gedlee
Does the program export the impulse response WITH the pre impulses of the harmonics?

Not as it is now - it would give huge amount of data.
But if you are interested I can easily make "special save" button with the direct cross-correlation in binary format.
(Sent me a direct message if of any interest)

If it does this then I am very interested, if it doesn't then I am not. Frequency response with THD is not very interseting.

With a timewindow I disregard the harmonics. so the THD is not included.

Farina talks about several things that have to be done to the "inverse crosscorrelator " to get rid of some artifacts. Its all in his latest AES paper. Do you do these things?

Not sure that I have read the latest (I am a AES member, but not subscriber), so sent me a link to the article.
I have tested with digital loopback and electric loopback to ensure I have no artifacts. But I might have forgotten something - I have an open mind :-)

[askbojesen]

Originally posted by BDP
Any sound card? I selected my sound card and then selected measure and my computer rebooted . The computer I tried it on is 5 years old with a SB live sound card running XP. Any tips?
PS: You are working with a 1/2 geek!

No errors before reboot? sounds weird
HOLMImpulse is using the .Net 2.0 windows framework
which is running in protected mode in windows.

Have you run a windows-update lately ?
http://update.microsoft.com/windowsupdate/

[BDP]

Yes, updated continously. .Net Framework 3.5 is loaded. There were no errror or prompts before the reboot. The system recovered fully.