TOPIC:

So if it helps us compare:

Device: DDRC24
Setup: 2.1
Room : 3.3m x 3.5m (not including bay widow) x2.6(h)
Spk L-R dist: 1.5m apart backing onto bay
Listening Dist: 2.9m
Crossover: 160Hz
Dirac listening area: Wide but
Measurement area: only 0.9x0.6x0.3m
Filter applied to: ~20hz-18khz

Tony_J wrote:

I will have a play later on and post some full range plots, but at least one difference may be that I am applying Dirac over the full FR (~20 Hz to ~20KHz), and given that there is a limit to the number of filters Dirac can fit into the DDRC's memory, that would result in less fine detail than if you apply Dirac correction over a smaller range.

My own example was full frequency range , as was @asx77's. @tw99's example looks like it was limited to a bit over 200 Hz.
Mine was for an SHD rather than DDRC 24 though. It would be interesting if these tests did show up a significant difference in performance between the two, as I wasn't sure there was one. As you may have seen there was another recent thread that highlighted the DDRC 24 seemingly uses a 48 kHz internal sampling rate vs 96 kHz for the SHD though.

Having said all of this, I'd expect any corrections of major bass peaks to be prioritised as at least to me I think these have the most obvious impact on the sound. My bet is therefore a combination of your room and spatial averaging means Dirac isn't 'seeing' large bass peaks to try to manage in your case.

Tony_J wrote:

With this particular system I am working in a relatively small room and the measurement volume is similar top what you are describing - measurements approx 1 metre from the central (first) measurement, and using the "chair" configuration in Dirac, the listening position about 2 metres from the speakers. It is far from being a good listening setup as the room doubles as a study and workshop!

My points were for a 1 m diameter measurement sphere centered around the MLP, so about half the measurement range to you I believe.

Edit: crossposted with @asx77.

asx77 wrote:

Device: DDRC24

Not a device difference then .

On the correction range front I'll just mention something I've spotted myself but don't recall see discussed anywhere. This is that DL appears to apply phase correction over the full frequency range no matter where the correction limiting 'curtains' are set.

This was, err, wrong .

That's an interesting observation.

Interesting observation on phase. Is it just the impact of IIR filters being applied? Or does it look like more targeted phase corrections above the filter window? Would you mind posting the REW output?

For example below is an example of a IIR low shelf filter centered at 100 Hz, as you can see there are phase impacts (although small) for many octaves above 100 Hz.




Michael

mdsimon2 wrote:

Interesting observation on phase. Is it just the impact of IIR filters being applied?

It's the Dirac filters rather than anything else I think. If your move the frequency limit 'curtains' and look at the bottom of the screen you'll see it says, "Designing magnitude response filter". I spotted this after spotting the phase effect but this message supports it.

My SHD is current refusing to connect to REW in microphone mode to allow me to demonstrate this well. I'll post back if it does. I've been experimenting with trying PEQ filters before Dirac this afternoon too which I'll post about somewhere later as well.

Ultrasonic wrote:

Tony_J wrote:

I will have a play later on and post some full range plots, but at least one difference may be that I am applying Dirac over the full FR (~20 Hz to ~20KHz), and given that there is a limit to the number of filters Dirac can fit into the DDRC's memory, that would result in less fine detail than if you apply Dirac correction over a smaller range.

My own example was full frequency range , as was @asx77's. @tw99's example looks like it was limited to a bit over 200 Hz.
Mine was for an SHD rather than DDRC 24 though. It would be interesting if these tests did show up a significant difference in performance between the two, as I wasn't sure there was one. As you may have seen there was another recent thread that highlighted the DDRC 24 seemingly uses a 48 kHz internal sampling rate vs 96 kHz for the SHD though.

Having said all of this, I'd expect any corrections of major bass peaks to be prioritised as at least to me I think these have the most obvious impact on the sound. My bet is therefore a combination of your room and spatial averaging means Dirac isn't 'seeing' large bass peaks to try to manage in your case.

Tony_J wrote:

With this particular system I am working in a relatively small room and the measurement volume is similar top what you are describing - measurements approx 1 metre from the central (first) measurement, and using the "chair" configuration in Dirac, the listening position about 2 metres from the speakers. It is far from being a good listening setup as the room doubles as a study and workshop!

My points were for a 1 m diameter measurement sphere centered around the MLP, so about half the measurement range to you I believe.

Edit: crossposted with @asx77.

My measurements are definitely not as focused as that - 1M radius (2M diameter) - mic held at arm's length gives roughly a metre from fingers to nose.

I've done some FR sweeps (disabling the crossover) - again, the mostly flat green line shows the transfer function without the Dirac filters and shows most of the PEQ I have applied - it doesn't show a small notch filter I inserted at 1250 Hz though because that appears on the other channel, but that doesn't really make that much difference. I've expanded the traces as much as was possible within the screen, so there is a bit more detail showing now.

Tony_J wrote:

My measurements are definitely not as focused as that - 1M radius (2M diameter) - mic held at arm's length gives roughly a metre from fingers to nose.

That was probably just indicative but if not you should treat yourself to a microphone stand .

I used measured positions (in 3D) for my dataset above, so I'd say the points were probably all within 2cm of being on a 1 m diameter sphere. This is the minimum recommended for Dirac Live I believe, which is why I used it.

Is your PEQ contribution around the crossover region in your active speakers?

Ultrasonic wrote:

Tony_J wrote:

My measurements are definitely not as focused as that - 1M radius (2M diameter) - mic held at arm's length gives roughly a metre from fingers to nose.

That was probably just indicative but if not you should treat yourself to a microphone stand .

I used measured positions (in 3D) for my dataset above, so I'd say the points were probably all within 2cm of being on a 1 m diameter sphere. This is the minimum recommended for Dirac Live I believe, which is why I used it.

Is your PEQ contribution around the crossover region in your active speakers?

I do have a mic stand - well, a camera tripod but same difference.

The PEQ contribution is mostly above the mid-to-tweeter crossover point (2000 Hz).

It's the Dirac filters rather than anything else I think. If your move the frequency limit 'curtains' and look at the bottom of the screen you'll see it says, "Designing magnitude response filter". I spotted this after spotting the phase effect but this message supports it.

I understand it is Dirac that is applying the filters, but Dirac uses a mix of IIR and FIR. I just checked it myself and it looks like the phase impact outside of the curtain is the result of IIR filters being applied by Dirac within the curtain and not targeted phase correction outside of the curtain.



Michael

Hmmm... I could have sworn it behaved as I said but double checking data now I see that it doesn't. I may have got some files mixed up before .

Here's a comparison of the different phase corrections for a full range filter (blue) and only up to 150 Hz (red) showing the additional changes made:



When I'd looked before I'd convinced myself that they look like the blue example in both cases. As for your point @mdisimon2 this shows the difference between the large scale effect you've highlighted and the targeted corrections superposed on this. I am left wondering whether what is going on above 5 kHz is having a negative effect?

(These examples are based on dividing microphone measurements rather than @tw99's method which is why they are noisier but they show the effect OK.)

Ultrasonic wrote:

I'd say the points were probably all within 2cm of being on a 1 m diameter sphere. This is the minimum recommended for Dirac Live I believe, which is why I used it.

I thought you just had to make sure measurements were a min of 30cm / 12” apart. One centre position with all others 50cm away seems excessively far. Where have you read that?

I am left wondering whether what is going on above 5 kHz is having a negative effect?

That looks like a timing issue to me and is most likely a measurement artifact. Are you using "Use loopback as timing reference" as your timing setting in REW? When measuring I will run a completely unfiltered sweep and adjust the timing offset in REW as necessary to make sure the phase response is flat. You will definitely need to make some adjustments to the timing offset if you have delays implemented.

Michael

mdsimon2 wrote:

That looks like a timing issue to me and is most likely a measurement artifact. Are you using "Use loopback as timing reference" as your timing setting in REW?

I'm not as I don't think it's possible with my laptop as opposed to a dedicated soundcard? I always use an acoustic timing reference.

I did spot an issue where the impulse responses that I measure change shape with the measurement length (with my 2x4 HD), which when I queried on the REW forum John Mulcahy said was probably due to a slight clock synchonisation error . Possibly the same thing as above? This was the thread, including graphs of the effect if anyone is curious:

www.avnirvana.com/threads/measurement-le...ation-question.4475/

mdsimon2 wrote:

When measuring I will run a completely unfiltered sweep and adjust the timing offset in REW as necessary to make sure the phase response is flat. You will definitely need to make some adjustments to the timing offset if you have delays implemented.

A couple of days ago I got @tw99's method of using my SHD as both source and 'microphone' to work and with no filters applied both the amplitude and phase responses were perfectly flat. Is this related to what you are suggesting or not? Apologies if this is a dumb question . If there is a way I can improve my measurements I'd like to understand how .

Edit: I do have timing offsets applied for subwoofer integration. Currently 9.56 ms to each main speaker. I'm not sure this should matter though?