TOPIC:

hi Guys, 
so far i thought SHD sound filtering leaves no artefacts in the sound. After recent ceiling treatment I readjusted the Dirac profile. Though treatment left mostly positive outcomes it for some reason revealed even deeper dip at 48-60Hz range. (see profiles screens attached). What I'm trying to achieve mostly by Dirac is to remove the spike at 32Hz. And the dip I'm filling up with a subwoofer located in a specific place of the room (hence you see in the target curve I'm not forcing huge boost, which itself would distort the sound). I'd probably leave the right curtain at 45Hz only but between 100-200 Hz there is strange stuff going on as well so I keep the curtain at 200 hz. 
In result some tracks with a lot of low frequencies get crackling distortion. Example: Janet Seidel, You belong to me, at 21st second of the track. 

Please advise how to fix it.
thanks

Dirac can only apply a maximum boost of 10dB which is why it hasn't completely filled in the dip around 48-60Hz. The typical solution is to lower your target curve so no dip is more than 10dB below it, then Dirac will cope. Dirac can apply lots of cut but only 10dB boost.

Crackling is possibly distortion due to clipping, does it go away if you reduce the SHD gain control?

it helps. thanks. the thing is I've always kept it at minus 3 dB, but after the changes I moved it all the way to zero. Going back to minus 3 seams to solve the issue.

It seems like your amplifier or speakers aren't sensitive enough. With Dirac providing 10dB gain and SHD gain near zero you will get clipping on any inputs near full scale. Depending on what sort of music you listen to this could happen often or never. You would probably benefit from more amplifier gain or more sensitive speakers, it seems you are using the headroom of the SHD to the maximum.

I use ATC SCM 100 active speakers. This is top notch equipment. 

The quality of the equipment is irrelevant if it doesn't have enough gain.

Perhaps check the Input Sensitivity control on the back of the speakers to see if you can increase the amplifier gain to get more headroom from the SHD.

no, no, I have enough gain. (I have also a preamp with 12dB gain) I can run even if SHD is at minus 10 or less. Just if you go down at SHD you actually limit the dynamic range of the source. I'm trying to avoid that too. 

Being at -3dB on SHD and at 0dB before suggests your gain structure is wrong. Gain is fine as long as its in the right place. In the wrong place it causes problems.

In case it isn't obvious, if you reduce SHD gain to avoid SHD clipping the only place you can add gain to restore the same SPL is after the SHD

I think you're getting me wrong. My system can go without clipping to huge volumes. Have a look at ATC. Your ears will start bleeding before ATC would start distorting. They are famous for this aspect. I know how clipping sounds like. The clipping would be related to the loudest sounds. The crackliing apears on rother quiet passage with some specific combination of low frequencies. If play again at the same volume with Dirac filter switched off, there is no cracling. If I switch on Dirac there is distortion. Hence I believe Dirac filter thakes credit for this. and this may have something to do with the dip that I have described in my original post (including 2 files attached). 
Any ideas?

Keep your SHD gain below -10dB and you should  hear no crackling. I suspect it occurs due to Dirac boosting certain low frequencies by 10dB. If your music content happens to have large amounts of those particular frequencies it may slightly clip even at overall low music levels. Its more about the frequencies. Once you have clipping at specific frequencies it can sound very different to typical wideband clipping.

If you have your SHD gain above -10dB your gain structure is wrong. You all but proved that by reducing the SHD gain from 0dB to -3dB.

hm... I treat SHD as component which should be transparent from gain perspective. I had a PC based source, which supplied a DAC via USB, then from DAC to Preamp by XLRs and from Pream to Speakers by XLRs. SHD was inserted between PC (USB) and the DAC (AES/EBU). Does SHD give any gain byitself, when it's above -10dB? What is the zero gain level for SHD? I thought I should just leave it some head room for the dips which I'm asking it to pull up. The dip which I have at ~50-60Hz is however a strong room mode so I can't be solved by SHD increasing the volume (hence the subwoofer in a different spot). I also understand (correct me if I'm wrong here) that cranking down SHD by 10dB reduces the signal dynamic range by 10dB. so why would I crank SHD down more than the necessary minimum?

Have you looked at the screen shots I attached? I was actually more curious about your opinion on how I modeled the room curve with this dip. as alternative I considered lowering the whole room curve by 10dB, but then I would simply loose dynamic range and second, I would have to apply Dirac to full frequency range (I'm still sceptical about using DSP for the full frequency spectrum)

The SHD is a sampled data system, digital in common parlance. It has enormous dynamic range so you can afford to lose quite a lot without it being audible.
The downside is that any such system clips if it runs out of headroom within that dynamic range.
Without Dirac you can set the SHD gain to 0dB and a 0dB input will output 0dB, all fine with no clipping.
With Dirac, Dirac *may* add 10dB boost at certain frequencies. With an input of 0dB that means the SHD level *may* cause clipping if set above -10dB.
Not all music will reach 0dB, and not necessaarily at the frequencies that Dirac has boosted, so you can often get away with SHD gains above -10dB for most music.
It comes down to how much headroom the music has and how much gain Dirac applied and which frequencies it applied it at, mostly things you don't really know.
What is certain is that with SHD gain below -10dB even a Dirac gain of +10dB cannot overload and distort even with the loudest possible music at 0dB.

I did look at your images but wasn't clear what you had done or why so tried to talk in generalities mentioning the 10dB thing hoping it was relevant.

Don't get hung up on wasting 10dB dynamic range, it really will be inaudible. You are obviously far closer to clipping problems than dynamic range problems at the moment.

so in such case would you simply straighten up the room curve near the 50-60Hz and go down to -10dB? Or should I keep this dip as it is and compensate with subwoofer? (subwoofer fills up the dip nicely but it introduces other issues)

I used to use SHD Studio at 0dB with an integrated amp. Dirac boosting certain frequencies resulted in crackling sounds with a few tracks. Reducing the maximum Dirac target curve boost to 3dB (following the measured curve) solved the problem. 

Hello,

I do not understand the Dirac and SHD behavior regarding boosting and reducing the volume. I would expect that Dirac does this completely automatically. Dirac should know the maximum boost that it applies. So it could reduce the gain by exactly this amount, I think.

Also it is always said that when activating Dirac, the gain of -10db is applied. So it should be safe to boost some frequencies by less than 10db. If I look at the screenshot of the 2nd target curve the boost levels are below this value.

But as a side node: You cannot boost a hole in the curve because the louder frequency will cancel itself out the same as before. So I think it would be best to fill the hole with the subwoofer as far as possible and then use Dirac to correct the sum of speaker and subwoofer frequency curve.