HOLMImpulse: Measurements in practice

So does using a loopback provide any real time timing controls when doing impulse measurements? From a tutorial I looked at on part express tech fourm, it seems that loopback is not a dynamic feature, it seems more for the ability to create a static calibration file using the make button under calibrations.
 
Another user published a YouTube video in the last year. He has stating that the loopback is somehow preventing clock drift when doing the time alignment with impulse response measurements. I have asked how he came to that conclusion and his response was his results using the loopback vs not using it when actively taking the measurements. All that I found in the documentation available online was using loopback to create a calibration with the "Make" button not that it is a dynamic feature of the software to prevent clock drift.

Should the loopback "Make" function be preformed before attempting to do impulse time alignment? Would that be a one time process or be preformed every time prior to doing impulse time alignment measurements?
 
diyAudio Moderator
Joined 2008
Paid Member
In the past and certainly with contemporaneous versions of windows, HolmImpulse only ever needed the single channel mic input and the timing was always correct, and this is unlike REW which needs both channels. There is a good explanation somewhere back in this thread, or maybe the sister thread. It may have been by gedlee.
 
Just don't plan on using a USB mic with locked timing ref lol.

There's much to be gained from using a real loopback timing ref, I think there is no question why other software (REW, ARTA, SE) has moved in this direction.
So does using a loopback provide any real time timing controls when doing impulse measurements? From a tutorial I looked at on part express tech fourm, it seems that loopback is not a dynamic feature, it seems more for the ability to create a static calibration file using the make button under calibrations.
That's correct, loopback cal "make" in Holm just creates frequency response calibration for your equipment, nothing to do with timing reference. It's the same as a "soundcard calibration" in REW. Optional process for the most part as the FR compensation only occurs at the very frequency extremes, calibration can be done once, as long as you use the same equipment for measurement all the time. Perhaps for some cheap class-D amp with pre-filter feedback, calibration should be done for every speaker connected, but better to avoid using non-linear amplifiers for measurement unless for permanent installation like a plate amp.
 
This is somewhat unclear to me, are you saying there is or is not any validity to connecting a loop back cable when preforming impulse response measurement for time/phase alignment? This setup is to measure alignment and phase of car audio speakers. XLR mic is on channel 1 and the left output is wyed between RCA to analog DSP inputs and the loopback plugged into channel 2 as was described in the following youtube video.
 
Last edited:
Description of video may be most informative: "kind of amateur". Holm is NOT a dual channel measurement system, despite the desciption of the setup at the start of your linked video. Loopback is unavailable for timing information in Holm. Is measures single channel only with "timing lock" which you'll see in the video around 8:45 to 9 minute mark. It uses the first measurement as timing reference and tries to keep the audio device active to maintain the timing. Close the application, disconnect USB interface, come back tomorrow and the timing reference is lost.

Use REW or ARTA if you want to use hard wired loopback, Recommend latest 5.20.14 early access release if using REW. I would recommend either over Holm.
 
  • Like
Reactions: 3 users
Administrator
Joined 2004
Paid Member
As stated above, HOLM is single channel.
I have run many hundreds of sweeps in HOLM (maybe 1000s) and the time lock has rarely failed me. When it did fail it was a full measurement crash, nothing subtle. I’ve gone so far as to test it over a dozen runs and it never varied by a single sample.

Caveat. I used M-Audio USB sound cards analog in/outs or the digital input of a crossover. Basically the same clock. Using different sound cards for in and out has always failed to give good results.
 
  • Like
Reactions: 1 user