You could:
> Set: Keep the stream active
> Set: Lock time zero
> Make a loopback measurement on Channel A
> Select: Auto Detect the largest peak
> reLock time zero to that sample value
[Remake this Channel A loopback measurement one or more times to assure the it is consistently locking at the IR peak. Mine setup sometimes drifts before becoming resonably stable so you may need to redetect the peak and reset the time zero lock accordingly.]
Once it is stable/repeatable:
> Remove the loopback cable and connect the mic
> Make a measurement on the Channel B
The distance between the 2 IR peaks is the absolute time of flight. This assumes you have no digital components in your system after the soundcard. If you do, then the latency of those units is also included.
To maximize the accuracy / clarity of the IR reading it may be best to do all the measurements (including Loopback) with a start frequency of 4000 Hz, or some other higher number as it will help give a clean/sharp IR that is relatively unaffected by any XO in the system. In practice, it probably won't make a significant difference to the measured result however.
Will dig out the gear and have a try.back to basics - do these "loopback" measurements look sane
Just been playing around in "back to basics" mode with no mic or speakers in the circuit, as prep work for measuring some power amps using a voltage divider. Before I go there, I'm just measuring the ADC/DAC, the USB SPDIF interface in loopback, and my preamp (ESP p88b):
I am using a MiniDSP Ministreamer (can do 48k/24 bit fill duplex) USB SPDIF in and out.
ADC is a Mic2496
DAC is a Hifimediy Hifimediy DAA1853
Asio4All drivers on an old WindowsXP laptop.
Main thing is the noise+THD seems relatively high for such a minimal setup, or am I expecting too much?
I don't have a particularly great cable connecting SPDIF from the Mic2496 to the MiniStreamer.
Do these plots look "sane" for what should be a very clean 24 bit 48khz recording environment?
blue line is digital loopback (dead flat!)
red line is DAC->ADC connected with RCA<->XLR adapter
green line is my ESP p88B preamp in line.. I can see it has a bit of attenuation in the LF but otherwise very "clean".
Just been playing around in "back to basics" mode with no mic or speakers in the circuit, as prep work for measuring some power amps using a voltage divider. Before I go there, I'm just measuring the ADC/DAC, the USB SPDIF interface in loopback, and my preamp (ESP p88b):
I am using a MiniDSP Ministreamer (can do 48k/24 bit fill duplex) USB SPDIF in and out.
ADC is a Mic2496
DAC is a Hifimediy Hifimediy DAA1853
Asio4All drivers on an old WindowsXP laptop.
Main thing is the noise+THD seems relatively high for such a minimal setup, or am I expecting too much?
I don't have a particularly great cable connecting SPDIF from the Mic2496 to the MiniStreamer.
Do these plots look "sane" for what should be a very clean 24 bit 48khz recording environment?
blue line is digital loopback (dead flat!)
red line is DAC->ADC connected with RCA<->XLR adapter
green line is my ESP p88B preamp in line.. I can see it has a bit of attenuation in the LF but otherwise very "clean".
Last edited:
Worked out the LF attenuation is impedance mismatch, my preamp is intended for amps of 22K input impedance or greater.. the ADC I'm using here has an input impedance of 2K.
Also reduced the noise considerably re-arranging cables a bit and using a different SPDIF cable. the 50/100 hz spikes I think are AC noise, not there using a different DAC.
The D'Appolito book, Testing Loudspeakers, http://www.amazon.com/Testing-Loudspeakers-Joseph-DAppolito/dp/1882580176 [/URL]
is very helpful and highly recommended.
is very helpful and highly recommended.
Great program, question about display options
Hi All,
My application for HolmImpulse is a little bit different than what it was intended for. I recently completed a DIY clone of an EQP-1A tube equalizer, and used HolmImpulse to characterize the frequency response of the finished unit... As it turned out, it worked really well for this application, and I couldn't be happier. Well, except for one minor thing. Since I typically want to display several sweeps at once corresponding to typical positions of control knobs, I found the default limitation of only displaying three measurements at once to be quite bothersome. Especially since there is ample room on that tab in the software to display six, seven or maybe more measurements at once.
So, my question - is there a way to change the default number of displayed measurements to increase it above three?
If not, is it possible to request display of additional measurements in a future release? It seems like it should require only a very small tweak in the code to enable. And I think that having this option would really open up HolmImpulse to new applications!
Thanks for your great software.
Hi All,
My application for HolmImpulse is a little bit different than what it was intended for. I recently completed a DIY clone of an EQP-1A tube equalizer, and used HolmImpulse to characterize the frequency response of the finished unit... As it turned out, it worked really well for this application, and I couldn't be happier. Well, except for one minor thing. Since I typically want to display several sweeps at once corresponding to typical positions of control knobs, I found the default limitation of only displaying three measurements at once to be quite bothersome. Especially since there is ample room on that tab in the software to display six, seven or maybe more measurements at once.
So, my question - is there a way to change the default number of displayed measurements to increase it above three?
If not, is it possible to request display of additional measurements in a future release? It seems like it should require only a very small tweak in the code to enable. And I think that having this option would really open up HolmImpulse to new applications!
Thanks for your great software.
I don't think that Ask is reading these threads any more 
I also would like to be able to display more than three traces. For when I want that, I export the measurements as frd files and import them into speaker workshop, I'm not sure what it's limitation is, but I've displayed at least 7.
Tony.
I also would like to be able to display more than three traces. For when I want that, I export the measurements as frd files and import them into speaker workshop, I'm not sure what it's limitation is, but I've displayed at least 7. Tony.
I don't think that Ask is reading these threads any more
Tony.
I talked with his boss and confirmed that Ask is on another project. He may also have left the company. At any rate this is what happens when the software is free. I do not anticipate that it will ever be improved from this point.
I wanted to work with them to impliment my polar amp plotting directly into the program. Holm thought it was a good idea, but didn't have a programmer to work on it.
Next time you speak to Ask's boss you might suggest that he consider making HOLMImpulse available under an appropriate free/open source license.
Some companies have successfully taken "dead" software and made it available as a FOSS licensed project. (That may not be appropriate here if, for example, the HOLMimpulse software shares a common heritage with other HOLM products).
Making the code available under a FOSS license doesn't always work-often software dies because it was no good to begin with-but it gives the software a fighting chance. And at least other smart programmers can learn something from it even if they don't develop it further.
Terry
Some companies have successfully taken "dead" software and made it available as a FOSS licensed project. (That may not be appropriate here if, for example, the HOLMimpulse software shares a common heritage with other HOLM products).
Making the code available under a FOSS license doesn't always work-often software dies because it was no good to begin with-but it gives the software a fighting chance. And at least other smart programmers can learn something from it even if they don't develop it further.
Terry
I think the main problem is HolmImpulse contains much of the coding (knowhow) of their commercial products, so I am pretty sure the code will never become open source (unfortunately). The only possibility is they produce a commercial version of HolmImpulse, but as software development is not their core business, I do not see that happening soon also...
Edit: corrected some typos