TOPIC:

Tw99 wrote:

Thanks @Tony_J for the detailed research. I've recently acquired an SHD, which I guess has more or less the same signal processing architecture, and this is the best description I've found so far about how it all fits together, and has clarified a couple of things I was confused about.

Something I learnt about recently is that you can use REW to directly measure what the SHD is doing, via USB. So the chain is:

REW -> USB Output -> SHD processing -> USB Input -> REW .

Obviously you don't see what's happening directly in the analogue domain, but I assume any clipping would show up in REW anyway.

Attached is a graph from this method using a measurement sweep that shows the frequency response of the Dirac filters in my setup. It looks like the maximum boost applied for me is around 5db.

Hey mate I've just tried to determine the dB gain added by Dirac using this method today by connecting my macbook to my SHD via USB. Any idea what I'm doing wrong? See the Macbook soundcard and REW settings, as well as the REW graph below. 

When I used the minidsp as input and output to REW it was with the DDRC-24. I tried it with SHD but couldn't get it to behave as it did with the DDRC-24. What I do remember with the DDRC-24 was I had to disable outputs 3 and 4 for it to work. 

asx77 wrote:

When I used the minidsp as input and output to REW it was with the DDRC-24. I tried it with SHD but couldn't get it to behave as it did with the DDRC-24. What I do remember with the DDRC-24 was I had to disable outputs 3 and 4 for it to work. 

Hmm, so you couldn't get it to work in the SHD at all? Odd. Any guesses as to why? Do you mean disable outputs 3 and 4 in the miniDSP? 

This discussion is split across two threads but just to confirm I got it working on my SHD. 

As it may be helpful here, this is the response I posted in the other thread:

Two quick thoughts are that you want to measure each channel separately not L+R, and I also wonder if you might have a high-pass filter active?

Below I'll copy and paste my own notes on how to do this, but they're for Windows not Mac. 

How to Make SHD ‘loop-back’ USB Measurements in REW

• Connect SHD to laptop with USB cable and select USB input on SHD (can be done via Volumio)
• Before starting it can save time to check the SHD works as a normal audio output device, so e.g. play a YouTube video and check sound plays OK. I have had occasions when I've not been able to get even this to work (after adjusting Windows audio sample rate etc) and switching my SHD on/off resolved it.
• In REW have things configured as in the following image and this works OK. Note that the sample rate chosen matches that defined via the Windows control panel (Hardware and Sounds -> Sound and then check Properties and Advanced for the SHD)

 

• When making measurements the left and right speakers correspond to the Dirac 1 and Dirac 2 inputs in the SHD routing, and the (usually microphone) Input number corresponds to the SHD output channel number. In the image above I had 4 selected as this what was in use for my right speaker. These channels can be set in the measurement window. (I don’t think the ‘loopback input’ is relevant.)
• Note that when working properly it is possible to use the signal generator and RTA and see results but I think using normal measurement sweeps works best. Both run OK with the SHD plugin open and connected to monitor what is going on there too.
• Be mindful of the digital headroom needed for the Dirac Live filters when running sweeps. I also generally switch my power amp off when doing this so there are no demand placed on the speakers, especially if the sweep is run to very low frequencies.

 

tallbeardedone wrote:

Thanks so much. So you got it to work with the SHD specifically? @asx77 said he couldn't get it to work on the SHD, only the DDRC-24, and he couldn't understand why. I don't see how I could have a high pass filter in. I only have Dirac in the chain. I will set output to single channel and see if that changes anything. Hmmm . . . 
 

Everything in the plugin is active as well, not just Dirac Live. It may well be you've never set a crossover filter though. (I have crossovers set on my SHD and so need to disable them to get a sensible loopback 'measurement'.) Have you sorted the sample rates as described above?

Should you ultimately not be able to get this to work there is the fall-back I used to use. Make a measurement (per channel) with Dirac on and another with Dirac off. Divide the former by he latter and you'll get the correction applied. You'll need to use the same acoustic timing reference (e.g. left speaker) for both measurements.

Ultrasonic wrote:

Should you ultimately not be able to get this to work there is the fall-back I used to use. Make a measurement (per channel) with Dirac on and another with Dirac off. Divide the former by he latter and you'll get the correction applied. You'll need to use the same acoustic timing reference (e.g. left speaker) for both measurements.

@Ultrasonic
Thanks for the reminder on the SHD method, I remembered we had discussed it in the past, I might get the laptop out at the weekend and try it again if I have time. One thing about your original method, you're measuring the output after the signal has gone through the amp and speakers and bounced around. It may be Dirac applies a lesser of greater correction than actually makes it to our ears and mic no? For example it might try to boost a null by 10dB but it only lifts what is measured by say 3dB hence why we all liked the method of looping in and out of the SHD, or have I misunderstood?

 

asx77 wrote:

@Ultrasonic
Thanks for the reminder on the SHD method, I remembered we had discussed it in the past, I might get the laptop out at the weekend and try it again if I have time. One thing about your original method, you're measuring the output after the signal has gone through the amp and speakers and bounced around. It may be Dirac applies a lesser of greater correction than actually makes it to our ears and mic no? 
 

No. The impact of the room should essentially be identical with and without Dirac Live active, so the difference will be what DL is doing. If this weren't the case then the whole concept of 'room correction' would fall apart actually, as it wouldn't be possible to predict what the result of a particular set of filters would be.

Ultrasonic wrote:

No. The impact of the room should essentially be identical with and without Dirac Live active, so the difference will be what DL is doing. If this weren't the case then the whole concept of 'room correction' would fall apart actually, as it wouldn't be possible to predict what the result of a particular set of filters would be.

Ok, just thinking about that... I kind of do and dont understand. If I boost a null, I thought I was boosting both the direct and reflected sound so the null remians, or at least doesnt change by as much as we'd want. 

Ultrasonic wrote:

No. The impact of the room should essentially be identical with and without Dirac Live active, so the difference will be what DL is doing. If this weren't the case then the whole concept of 'room correction' would fall apart actually, as it wouldn't be possible to predict what the result of a particular set of filters would be.

Ok, just thinking about that... I kind of do and dont understand. If I boost a null, I thought I was boosting both the direct and reflected sound so the null remains, or at least doesnt change by as much as we'd want. 

asx77 wrote:

Ultrasonic wrote:

No. The impact of the room should essentially be identical with and without Dirac Live active, so the difference will be what DL is doing. If this weren't the case then the whole concept of 'room correction' would fall apart actually, as it wouldn't be possible to predict what the result of a particular set of filters would be.

Ok, just thinking about that... I kind of do and dont understand. If I boost a null, I thought I was boosting both the direct and reflected sound so the null remians, or at least doesnt change by as much as we'd want. 

That would be true for a perfect null but in practice they don't exist. I understand where you're coming from though. I might see about doing a comparison some time to prove/disprove my point...

tallbeardedone wrote:

Tw99 wrote:

Thanks @Tony_J for the detailed research. I've recently acquired an SHD, which I guess has more or less the same signal processing architecture, and this is the best description I've found so far about how it all fits together, and has clarified a couple of things I was confused about.

Something I learnt about recently is that you can use REW to directly measure what the SHD is doing, via USB. So the chain is:

REW -> USB Output -> SHD processing -> USB Input -> REW .

Obviously you don't see what's happening directly in the analogue domain, but I assume any clipping would show up in REW anyway.

Attached is a graph from this method using a measurement sweep that shows the frequency response of the Dirac filters in my setup. It looks like the maximum boost applied for me is around 5db.

Hey mate I've just tried to determine the dB gain added by Dirac using this method today by connecting my macbook to my SHD via USB. Any idea what I'm doing wrong? See the Macbook soundcard and REW settings, as well as the REW graph below. 

Were you able to get it working? Your results look like you did not have USB selected as the input on the SHD or the output you were trying to measure did not correspond to your input in your routing matrix.

It isn't a big deal but I would set your REW sample rate to 96 kHz to match the SHD and set the SHD input/output sample rate to 96 kHz in Audio MIDI Setup to ensure you see the full bandwidth of the SHD. Note REW does not like it if you change the sample rate in Audio MIDI Setup after you start REW so make sure you make any changes to Audio MIDI Setup prior to starting REW.

Good luck!

Michael

I've just had a go at comparing the loopback method to my acoustic one for measuring Dirac Live filters and I'm happy to say I feel I've comprehensively proved that the acoustic method holds up very well  . My actual setup is complicated by crossovers and delays so for this test I purely used my main speakers full range, and for both speed and to give the largest signal boosts I just used a single measurement at the main listening position (MLP) to generate the filters. For context, at the MLP the Dirac Live On/Off spectra comparison looks like this (for the right speaker):

 

I just left the target curve at the default DL came up with. In hindsight I should have brought the low frequency limit up as I wouldn't normally want to boost below the limits of the speakers output as shown here, although it ended up being a more complete test because I didn't. I then had two goes at measuring the DL filter by my acoustic method of making a measurement with DL On, a measurement with DL Off, and then dividing the first by the second (in REW). I say two goes as in principle the measurement can be made anywhere in the room. I first did this at the MLP and the comparison of my acoustic method to the loopback method looks as follows, where the graph shows the amplitude changes of he Dirac Live filter: 

 

Overall it's a pretty good match but above 6 kHz there are some notable differences which could be significant for anyone who uses Dirac Live full-range (I don't). There's no meaningful signal at 10 Hz from my main speakers so I'd pay no attention to the difference there.

I then tried with the microphone moved to about 25 cm in front of the speaker. Using data measured there the calculated DL filter matches that for the loopback method extremely well:

 

For completeness, here is how the phase adjustments of the Dirac Live filters compare as measured at this same nearfield location:

 

Don't get me wrong, the loopback method is clearly the one to use provided that someone can get it to work, but for anyone that can't my acoustic method is a pretty decent substitute I think  .

Well... That's pretty clear. 👍👍

In case it helps anyone else get this working I'll just add the following. I was having another play about this morning to see if I could get the loopback method to work with a 96 kHz signal. After initial attempts I got into a position where I couldn't get the test to work at any sample rate, and I also couldn't play audio via USB from my laptop (where I run REW) either. Rebooting my laptop made no difference, but turning my SHD off and back on again fixed it. Below is what USB loopback 'measurements' look like run at 96 kHz sample rate. What I was plotting above what the result of dividing the DL On by the DL Off 'measurements' to give the actual filter (i.e. so that the dB levels were the actual changes applied).

 

For info. I've now gone back and edited my 'how to' post above, adding a picture of the config. and simplifying the rest to focus on what I now believe is necessary. As is probably obvious I'm far more knowledgeable when it comes to acoustics than computer audio, but this works for me:

www.minidsp.com/forum/dirac-series-suppo...droom?start=75#63051