TOPIC:

I don't know if it's been said before but I just hooked my new SHD to my scope: if you are using Dirac don't turn volume past -10dB otherwise it will clip (depending where and how much boost you applied to your curve). Below -10dB is safe.

If you are not using Dirac it will not clip even at 0dB. Probably Dirac needs about 10dB of headroom.

There's been some similar investigation for the DDRC-24 in the thread below, which also suggested at least 10dB headroom being a good idea, which can be guaranteed by using attenuation set in the plugin rather than risking ever going above -10 dB on the master volume. I also wonder if a bit more than 10 dB may be prudent to allow for phase change effects increasing peaks by more than the maximum boost applied.

www.minidsp.com/forum/dirac-series-suppo...nd-internal-headroom

It is likely more complicated than that. SHD dB is a gain, not a level, so where it clips depends on input level.
My recollection is that devteam said you can have SHD at 0dB and Dirac takes care of headroom, but is has always been ill defined just what Dirac does for headroom since Dirac 1 which was explicit.
Boosts of the Dirac curve are possibly your problem

In my case I'm adding about 4dB of attenuation on top of that because when master volume is at 0dB, the xlr analog will output 4V.
Being my amp input sensitivity about 2,5V anything more is going to clip the amp input.

entripy wrote:

It is likely more complicated than that. SHD dB is a gain, not a level, so where it clips depends on input level.
My recollection is that devteam said you can have SHD at 0dB and Dirac takes care of headroom, but is has always been ill defined just what Dirac does for headroom since Dirac 1 which was explicit.
Boosts of the Dirac curve are possibly your problem

I think it might have been that Dirac used to internally manage headroom but the experience of the OP and the evidence in the thread I linked to prove that this in no longer the case in v2. I suppose v3 might be different again though!

You're right that it relates to signal level but given that many recordings hit the 0 dB level I think it's prudent to just assume that some attenuation is needed, as the OP suggested. The cause is definitely boosts applied by Dirac.

entripy wrote:

It is likely more complicated than that. SHD dB is a gain, not a level, so where it clips depends on input level.
My recollection is that devteam said you can have SHD at 0dB and Dirac takes care of headroom, but is has always been ill defined just what Dirac does for headroom since Dirac 1 which was explicit.
Boosts of the Dirac curve are possibly your problem

Sure, as I said it depends on the boost you have in the target curve, but is unlikely you are using Dirac with no boost, at least in the low frequencies. I have about +6dB boost at 40Hz, if I play a 40Hz sinewave at 0db and observe it on my oscilloscope it is clearly clipped with Dirac on, turn Dirac off and it is fine.

Btw I started suspecting that because I was using the SHD into a preamp, with the SHD master volume at 0dB as I was controlling the volume with the preamp. With this setup I started to hear distortion on some tracks (I play loud most of the time ) so I fired up the oscilloscope to check.

I use the default Dirac target, the slight slope which Harmon and everybody have advised for decades. No boosts, no cuts, just the default. Of course I fully expect Dirac is providing up to 10dB boost to achieve that and my experiences strongly suggest it cannot boost any more than 10dB, So something somewhere needs to attenuate by 10dB to prevent clipping with full level inputs.

All I can say is that Dirac 1 had an explicit 10dB (and user variable) attenuation. Dirac 2 lost that but my levels didn't obviously change when I switched from Dirac 1 to Dirac 2 so I've always believed Dirac 2 still has a hidden 10dB attenuation.

I don't think I've ever set my SHD above -15dB, partly because my amps only require 0.775V nominally, so maybe that's avoided the problem for me. Also, it is likely that the music that I play doesn't often reach 0dB if at all.

Gain structure is hard. Fortunately the SHD has an enormous dynamic range which means even less than optimal gain structure normally works very well. There should be many ways of avoiding clipping with all the input and output gain/attenuation available in the plugin. Before I purchased my SHD I worried about level mismatches and clipping but in practice setting the SHD to -15dB to -25dB everything just works fine without obvious clipping.

cberardelli wrote:

Btw I started suspecting that because I was using the SHD into a preamp, with the SHD master volume at 0dB as I was controlling the volume with the preamp. With this setup I started to hear distortion on some tracks (I play loud most of the time ) so I fired up the oscilloscope to check.

Others have had similar issues and I do think this is something it would be worth miniDSP highlighting in their manuals.

Have you tried using your miniDSP without the external pre-amp?

If you use the default target curve (I wouldn't say it's exactly the Harman one but that's for another post) you will have boosts and cuts to correct your measured FR to the target curve unless your uncorrected measurements are perfectly coincident with the target, in that case you don't need Dirac at all.

Ultrasonic wrote:

cberardelli wrote:

Btw I started suspecting that because I was using the SHD into a preamp, with the SHD master volume at 0dB as I was controlling the volume with the preamp. With this setup I started to hear distortion on some tracks (I play loud most of the time ) so I fired up the oscilloscope to check.

Others have had similar issues and I do think this is something it would be worth miniDSP highlighting in their manuals.

Have you tried using your miniDSP without the external pre-amp?

Yes, now I'm using it directly into the amp but that doesn't change anything. I measured the SHD at the output terminals (XLR) and with some tracks or certain test signals it clips when the master volume is above -10dB and Dirac on. Once is clipped is clipped regardless you are feeding the signal to a preamp or to the amp. That is not a problem as you can attenuate the output in the SHD plugin but is good to know...

My room needed all the help it could get and even Dirac couldn't fill in the 15dB dips (which is why I concluded it had a maximum boost of 10db).

I used to use my SHD with a preamp but gradually all the input sources moved to the SHD and the preamp was left with a single input so I just connected that to the SHD too and got rid of the preamp. My SHD was betwen preamp and amp though, not before preamp.

cberardelli wrote:

Yes, now I'm using it directly into the amp but that doesn't change anything.

I wasn't suggesting it changed anything in relation to clipping . Rather just suggesting you explored this different option if you had't already.

Ultrasonic wrote:

cberardelli wrote:

Yes, now I'm using it directly into the amp but that doesn't change anything.

I wasn't suggesting it changed anything in relation to clipping . Rather just suggesting you explored this different option if you had't already.

Yup, I ditched the preamp (was using it just because of integration with the HT like an HT bypass but I found another way to do it).

I experienced the same. With SHD at 0dB it clips at high output on the frequencies that are boosted more than about 2dB. I heard this at different frequencies where the target curve crossed deep valleys in the measured curve. A solution is to have the target curve follow the deepest valleys.