HOLMImpulse: Measurements in practice

[DQ828]

Originally Posted by Nrik

DQ828: Your phase looks quite right for a 3-way.

Every crossover point with a 4th order filter, makes for a phase shift of 360 degrees. The phase response starts to change approx. 10x below and 10x higher than the attack frequency.
Normally all frequency analysers anticipates 360degrees = 0 degrees, because you have moved to the same point at the next cyclus. Your signal will actually be delayed 1 cycle, which means that the delay in mS will depend on the actual frequency.

Normally the phase shifts are not directly audible, so you should just base your choices for level of delay on each frequency area, on the impulse response. Even that looks nice allready on your graphs. How does it sound?

I hope it makes sense.

I have just this minute finished setting them up to start listening, as it is late & I am knackered I'm wont get much listening done tonight, but will report back in my build thread, in the coming days, OB Project Design Started, Help Requested..

Thanks for the reassurance, I'm not sure how you worked all of that out from the graph, I wish I new what it meant. Thanks again

David

[juz400]

Hey everyone,

I want to make a cable to do the Sound Card measurement

From Xonar Essence STX phono (OUT) to Blue Icicle XLR (IN)

I have a twin core shielded cable, XLR and (1) Phono connectors

looking at the Diagrams here I think I should wire as No. 17

Sound System Interconnection

I see in other places I should wire as No. 6

as the Icicle provides the 48v Phantom automaticaly (no switch) to the ECM8000,
I dont want it powering my Sound Card

Im eager to start playing with this great bit of software!
Thanks
Justin

[jtalden]

Originally Posted by DQ828

I have tried to take some time measurements of the new speakers, so I could apply a delay for the Mid-Tweeter drivers using the MiniDSP, as expected the Mid & Tweeter appear to be aligned well enough (from my reading) but the bass driver appears to be in the next room I would say the physical horizontal distance between the MT & bass driver is about 40mm (bass behind MT). Does anyone know how to interpret these measurements.

Oh and I have no idea what the phase is trying to tell me so some help there would be greatly appreciated.

David

Since you have the MiniDSP there is no reason not to fine tune the phase handoff of your setup. The mic should be on axis and back from the speaker far enough to minimize the relative distance error between the drivers and close enough to avoid excessive multipath issues. You may want to try about 2 m back and move closer if needed.

Assuming your graph reflects this type of setup the phase can be improved for both the W-MR and the MR-TW XOs. The dotted phase curves should overlay as much a possible through the acoustic XO ranges or from about 200-450 Hz and 1.8k to 4k Hz.

Note that the W phase at the 380 Hz acoustic handoff is about 40 deg and the MR is about 180 deg. This is a 140 deg differential. If you delay the MR more relative to the W the two curves will overlay much better. For the MR-TW the phase curves are closer, but can still be improved a little. The TW needs a little less delay relative to the MR for the best match.

Your system can be setup within 10 to 20 deg throughout the XO ranges with some fine tuning of the delays.

[juz400]

Well, I knocked up a cable as Diagram No.17

Tip of Phono to pin 2 XLR
Case of Phono to pin 3 XLR

I havent pluged it into the soundcard but an output of another device,
the Phantom power light (Ring around the gain knob) lights up as if to say its sending 48v somewhere...

maybe I`ll try connecting pin 1 and 3 of the XLR together...

Still unsure if Im barking up the wrong tree here and Im not going to get a result either way..

Do I NEED a transformer between the phono and XLR connector?

[AllenB]

I've been doing HF measurements with Holm using swept sine and all has been good.

I'm now trying to do LF measurements and the MLS signal seems to lack low frequency energy (it sounds like white noise). I am having trouble getting over the noise floor which is also a problem. The noise floor is tilted downward as shown in the picture. This was taken using the MLS signal but I removed the mic and muted the amps driving the speakers before I took this shot so it would show the floor. Does this look normal?

[wintermute]

It doesn't look particularly normal to me, 50db is an awfully big difference between your low freq noise floor and your higher frequency noise floor!

Have you done a loop-back measurement on your soundcard? That is cable from output looped back to input? It should give a very flat response. Or is that what the plot up the top is? if so was it a sampling rate of only 2048 Hz?

Below is a plot of my noise floor almost as you have described. Blue plot is a loopback measurement. Red trace is with a Mic preamp connected via RCA cable plugged into the sound card but no mic connected. Green trace is with Mic preamp plugged in, turned on and microphone attached (just ambient background noise).

If you did take the measurement at only 2048 Hz sampling rate (to get more low frequency resolution) I'd try bumping it up to at least 16K (I used to use that with speakerworkshop). I can't select anything lower than 44.1Khz in Holm though, so I'm using one of my card's two native resolutions ie 48Khz...

edit: also the plot below is raw response, and sort of smoothing I turned on changed it quite a bit.

Tony.

[AllenB]

The red trace is the loopback response with the signal lowpassed at 1000Hz. The blue trace is with the input coming from the phantom power supply with no mic and amps turned off, the green trace is the same but with Holm set to lowpass the signal at 16,384Hz...although I can't see where the lowpass frequency can make much of a difference the way I have it set up at the moment. I just hope this data can show something useful.

[wintermute]

hmmm do you get the ~20Hz harmonics (between 20 and 100Hz) if you do a measurement with Nothing plugged into the sound card at all? It does seem like there is a wiggle there even in the loopback measurement....

Also have you done a DAC-ADC calibration with the loopback cable connected?

Is your phantom power for your mic battery operated?

I'm also wondering if there is any form of earth loop between the amp and your pc (though if the amps were turned off maybe not)...

edit: also I just lowpassed my loopback (post measurement) at 1000Hz and I get a quite different result to what you have shown, starts to roll off at 1000Hz smoothy going to nothing over the range of 1000Hz (as per the second parameter). I think something is definitely screwy.... you could try clearing all of your measurements and shutting down and restarting your pc

Tony.

[AllenB]

Originally Posted by wintermute

hmmm do you get the ~20Hz harmonics (between 20 and 100Hz) if you do a measurement with Nothing plugged into the sound card at all? It does seem like there is a wiggle there even in the loopback measurement....

Same with in/out removed at the soundcard.

Also have you done a DAC-ADC calibration with the loopback cable connected?

Yes.

Is your phantom power for your mic battery operated?

Yes. My phantom power supply/preamp is simply batteries and capacitors.

I'll just overhaul the phantom box, needed or not. I once had an issue with the joints within it.

[AllenB]

Batteries are good but have a 0.6V difference. Solder joints are good. I could do some DeOxit work on my patch cables.