TOPIC:

The 'Auto EQ with REW' document (SHD manual on page 92 has a link to this) says:

"Firstly, set the Equalizer to "MiniDSP" for plugins running at 48 kHz, and "MiniDSP-96k" for plugins running at 96 kHz".

However, this setting only allows a maximum of 5 filters and the SHD plugin has a maximum of 10.

REW documentation says this:

MiniDSP-96k - The MiniDSP-96k equaliser setting supports the same filter types and resolutions as the Generic setting, but for 5 filters. It is aimed at MiniDSP plug-ins that operate at 96 kHz.
MiniDSP 2x4 HD - The MiniDSP 2x4 HD equaliser setting supports the same filter types and resolutions as the Generic setting, but for 10 filters at 96 kHz.
nanoAVR - The MiniDSP nanoAVR equaliser setting supports the same filter types and resolutions as the Generic setting, but for 10 filters at 96 kHz.

My reading of this is that there is no difference (in REW) between the 2x4HD and nanoAVR settings and that either of these could be used to allow a maximum of 10 biquad filters to be exported from REW and imported into SHD.

Am I correct - will using the 2x4HD or nanoAVR settings work with SHD?

I've used the 2x4 HD setting in REW to successfully do what you'd want with my SHD.

I had a 2x4 HD already and so just did this without thinking about it rather than looking at the manuals, which by the sound of it confuses things!

For info. the only reason I've used the PEQ option on my SHD was to make a direct comparison of performance to my 2x4 HD. Normally I would use Dirac Live instead.

Thanks for the feedback.

I'm looking at using a combination of PEQ and Dirac. I'm implementing a 2.2 system (i.e. one subwoofer per channel) and using the crossover in SHD. If I just use Dirac (I set the crossover in SHD and then run Dirac calibration) I find that Dirac over-compensates below 80Hz caused, I think, by a big peak at 60Hz - Dirac tries reduces this but overdoes it.

So, what I'm trying is this:

1. Set crossover in SHD. I use 24dB/Octave low pass for subwoofer - the subwoofer filters are switched off and their measured response is pretty much flat from 100Hz to 1KHz. For the main speakers I use and 12dB/Octave high pass - they have a measured LF roll-off of 12dB/Octave so this gives, combined with the SHD filter, gives an overall acoustic filter of 24dB/Octave.

2. Measure Subwoofers and Main Speakers separately in REW and equalise so that (a) peak at 60Hz is reduced and (b) measured response closely follows 24bB/Octave slope between 50 and 300 Hz (crossover is at 100Hz).

Next stage (which I'll do over next few days) is to then see what happens if Dirac calibration is run with the PEQ filters active. Will be interesting to see if Dirac is more (or less!) effective because it has to do less when some of the EQ is done with PEQ.

You've presumably tried this but can't you just set a target curve in Dirac that is higher below 80 Hz?

Yes - tried something like that - problem is that Dirac shows that it has fitted to target curve but measured response shows that actual response is still too 'lumpy' below 100Hz.

Listening confirms the measurement. I use a track by Rebecca Pidgeon (Spanish Harlem - www.digido.com/portfolio-item/subwoofers/) to listen. Listening with PEQ applied (and no Dirac) is much better than with Dirac (and no PEQ).

miket wrote:

Yes - tried something like that - problem is that Dirac shows that it has fitted to target curve but measured response shows that actual response is still too 'lumpy' below 100Hz.

Listening confirms the measurement. I use a track by Rebecca Pidgeon (Spanish Harlem - www.digido.com/portfolio-item/subwoofers/) to listen. Listening with PEQ applied (and no Dirac) is much better than with Dirac (and no PEQ).

One way to 'trick' Dirac to do what you want is to only make a measurement at the central measurement location, rather than all of the different positions. The curve with Dirac applied will then very closely match the target curve. I've done exactly this myself in the past. In principle Dirac is being 'smarter' by using the data from the multiple measurement locations but if you don't like the result then obviously its not as good for you.

Thanks - good suggestion - I'll give it a try and compare:

(a) Dirac Only with full measurement.
(b) Dirac Only with one measurement.
(c) PEQ Only.
(d) Dirac with PEQ and full measurement.
(e) Dirac with PEQ and one measurement.

Shame that there are only 4 presets - would ideally like to be able to switch between all of the above!

[quote="Ultrasonic" post=48896
One way to 'trick' Dirac to do what you want is to only make a measurement at the central measurement location, rather than all of the different positions. The curve with Dirac applied will then very closely match the target curve. I've done exactly this myself in the past. In principle Dirac is being 'smarter' by using the data from the multiple measurement locations but if you don't like the result then obviously its not as good for you.[/quote]

Interesting. How do you get Dirac to cycle through to the optimize target page without going through the whole Dirac position sequence?

rodrigaj wrote:

Interesting. How do you get Dirac to cycle through to the optimize target page without going through the whole Dirac position sequence?

Simply measure the central point and then click on the filter design tab .

Did some testing with a,b,c,e. Initial reaction is that order of preference is:

1. Dirac with PEQ and one measurement (e)
2. PEQ only (c)
3. Dirac Only with one measurement (b)
4. Dirac Only with full measurement

Having said that, only very small differences between 1,2, and 3 - but all of these substantially better than 4.

Wonder if using PEQ do do most of the 'heavy lifting' means that Dirac doesn't have to do big corrections and, therefore, works better?

How do the measured frequency responses compare for your different options?

Dirac with full measurement (and no PEQ, just crossover set in SHD) has 6 dB dip centred around 56Hz.

Other combinations are within 2dB from 40Hz to 100Hz.

Next is to try PEQ + Full (multipoint) Dirac Measurement next and also do some measurements to look at time domain.

miket wrote:

Dirac with full measurement (and no PEQ, just crossover set in SHD) has 6 dB dip centred around 56Hz.

Other combinations are flat (within 1dB) from 40Hz to 100Hz.

Next is to try PEQ + Full (multipoint) Dirac Measurement next and also do some measurements to look at time domain.

By flat do you mean horizontal? For info. most people prefer a response that increases to some degree as the frequency decreases.

I asked about the responses as if they're different then this is likely what you're hearing/preferring, and by adjusting the target curve I still suspect you can do all you want just with Dirac.

One other quick thought. When you're comparing frequency responses are you making and looking at full frequency range measurements? If not it's possible to have similarly flat bass response but the relative level to be different to the rest of the frequency range. I got caught out by this once as frustratingly Dirac only works with relative levels rather than absolute ones.