TOPIC:

I got it working! So easy. I had it nearly right just had to set my REW preferences as shown. Now I can see that to boost my response up at 20Hz is using max Dirac headroom and will be reeaallly pushing my amp so I think I'm going to narrow the curtain to just above 20Hz to take some load off. Thanks so much for all the detailed and helpful replies guys. Here are my first three presets and their corresponding house curves in Dirac. I'm using max boost to get those low frequencies up so might attenuate them a bit. 

I'm pleased by posts turned out to be of some use @tallbeardedone  .

What are your speakers and what have you done in terms of trying to optimise speaker and listening position prior to do doing your Dirac Live measurements? It looks like you're trying to do rather too much in terms of extending bass range and boosting levels there to me...

Ultrasonic wrote:

I'm pleased by posts turned out to be of some use @tallbeardedone 

.

What are your speakers and what have you done in terms of trying to optimise speaker and listening position prior to do doing your Dirac Live measurements? It looks like you're trying to do rather too much in terms of extending bass range and boosting levels there to me...

I optimized my speaker placement using REW and sliding my speakers 1cm at a time out from rear wall and measuring until I got best bass response in room. I've written a detailed guide on how I did it. Speakers are Monitor Audio gx300. Final placement shown below. I slid the curtain up to 20hz so Dirac doesn't have to do so much work to boost SPL down there and now the gain is much lower (7dB max at 50hz). Happy with it now. See new measurements and curves below. 

If you're happy then great, but bear in mind that where the listening position is in the room is a huge factor. I'd personally be looking to find somewhere with more output at 40 Hz, which is not that uncommon in music.

Re. the low frequency limit bear in mind that 20 Hz could well still be too low, although if you're not using your system for movies you'll be unlikely to ever have much signal there to worry about anyway. 

Edit: and do some time experiment with only applying DL in the low-bass region. I only like using it up to 150-200 Hz myself.

Ultrasonic wrote:

If you're happy then great, but bear in mind that where the listening position is in the room is a huge factor. I'd personally be looking to find somewhere with more output at 40 Hz, which is not that uncommon in music.

Re. the low frequency limit bear in mind that 20 Hz could well still be too low, although if you're not using your system for movies you'll be unlikely to ever have much signal there to worry about anyway. 

Edit: and do some time experiment with only applying DL in the low-bass region. I only like using it up to 150-200 Hz myself.

That position has the best bass response at 40Hz (least suck out). I spent weeks moving speakers (and listening position) 1cm at a time to find best response in the room. The suck outs at other positions were much lower. Uncorrected response measured at tip of equilateral triangle and roughly 60cm inside that point (where 3-D imaging is most emmersive) are shown below. That 7dB gain there isn't too bad tbh. 

For info. if you want to share REW graphs it's best to use the camera symbol at the top left of each graph that lets you save it as a simple jpeg file. This is what I used above. Using the All SPL view rather than the distortion view also makes it easy to show different curves at once and if the legend is shown it also tells the reader what averaging was used.

I agree though that your last MLP graph does look pretty reasonable  . How do you think it sounds? Ultimately it's your personal preference that will determine what is best. 

Do be mindful that the last target curve you showed will be working the low end output of your system a lot more than you're used to so see how you get on. Note it's the overall shape of the DL filter that matters rather than specifically if something is a boost or a cut. As an exaggerated example, applying no bass boost but reducing the mid-range by 20 dB is the same as a 20 dB bass boost and no change at mid-range in terms of the demands placed on amp and speakers (there would be a difference in terms of concern re. digital clipping and therefore steps to manage this).

Tbh I’m currently finding myself always trending back to unprocessed raw digital to the DAC. I think I like the sound of my room. As shown above at my listening position it’s flat everywhere except that 60hz boost which I like. Unprocessed it just sounds more “real” atm. To give an example, if an upright bass was being played in my room, 60hz would be boosted, so when I hear that boost due to room it sounds more like the bass is really there. I will continue to tinker with Dirac but like I said at the moment the “wow” factor is diminished slightly with it on (using any of the filters I’ve made so far.) my system excelled at imaging and soundstage and the time domain correction of dirac miiiight have focused it a bit more but at the expense of realism. I will keep tinkering tho and see if I can’t find a filter that gives the wow and realism back. 

And yes, I’m aware that there was way too much boost in that +10dB Harmon curve, so to get a 10dB drop from 20Hz to 20khz I’ve instead boosted the bass only by 3dB and dropped it to -7dB at 20k for the same overall curve. Doing that requires more work from my amp but at ~500W into 4ohms it’s more than up for the job. I’m finding that level matching is sort of difficult and maybe some of the reason I’m preferring uncorrected is merely due to more headroom on my amp. 

so … yeah. Still tinkering but atm preferring no dirac. 

tallbeardedone wrote:

 To give an example, if an upright bass was being played in my room, 60hz would be boosted, so when I hear that boost due to room it sounds more like the bass is really there.
 

but do you want it to sound like the bass is in your room or do you want it to sound like you're  sat in front of the stage? Both are acceptable answers. 

tallbeardedone wrote:

so … yeah. Still tinkering but atm preferring no dirac. 

One of the strengths of Dirac Live is how flexible it is. Firstly, my initial thought on the last target curve and response you shared was that I'd personally almost certainly hate how it would sound. Different people have different tastes though which is why I didn't say this. 

My strongest recommendation of something to try is what I mentioned above: bring the high-frequency curtain down to somewhere in the 150-200 Hz region so that you're only making any adjustments in the low bass region. EQ in this region is far 'safer' in terms of not messing up the things your speakers are probably doing right (watch the excellent video below for a decent discussion of why - it's long but very watchable, particularly for someone with your background). EQ across the whole frequency range based on a single measurement position is definitely 'wrong' but by making measurements across a much wider range of measurement positions DL attempts to overcome the limitations. Some like the results and some don't, and neither group is 'wrong' as far as I'm concerned. The goal here is for each individual to do whatever means they enjoy the sound of their system most  .



It's also worth noting that the distribution of measurement positions used will affect the results. In your position I know some would deliberately make the measurements with the listening chair removed too. 

Hi @Ultrasonic,
What is it about correcting the full frequency range you don't think works so well? Is it something specific that Dirac tries to correct that is the problem? For example, if you had speakers that were a bit hotter on the highs than you want, could you apply a basic EQ with REW that doesn't have all of Dirac's bells and whistles and leave Dirac to the lower end? Or would you warn against any form of EQ above a certain point? But if you did use dirac full range would a benefit be that it helps get the timing of tweeter and woofer at the mlp better?
Thanks
Alex

asx77 wrote:

Hi @Ultrasonic,
What is it about correcting the full frequency range you don't think works so well? Is it something specific that Dirac tries to correct that is the problem? For example, if you had speakers that were a bit hotter on the highs than you want, could you apply a basic EQ with REW that doesn't have all of Dirac's bells and whistles and leave Dirac to the lower end? Or would you warn against any form of EQ above a certain point? But if you did use dirac full range would a benefit be that it helps get the timing of tweeter and woofer at the mlp better?
Thanks
Alex

Please take the time to watch the video I posted. Re. Dirac the concern about EQ adjustments at higher frequencies based on in-room measurements applies to all speakers. If a speaker truly was too bright for a room then rolling this off using simple EQ in the miniDSP plugin would be an alternative way to try this than in principle seems more logical to me. Note that the room has a big impact on how bright a system may sound too, rather than is necessarily being a speaker 'problem'.

You are right that the phase/timing adjustments of DL are only applied over the defined frequency range. 

Please let me stress I'm just giving ideas for you to experiment with. Different people like different things, and what any individual enjoys most is what really matters  .

Yep, but it's good to understand where you're coming from. I'll watch the vid (with Mr Toole it will be good).

I give the 'loop through SHD' method another go hopefully later. Can you remind me, it's a case of sweep with and without dirac then do I need to do anything with trace arithmetic? If everything is bypassed I'm guessing not but if I leave the filters on my mains rather than bypassing them do I need trace arithmetic to show me the difference?
Thanks

asx77 wrote:

Yep, but it's good to understand where you're coming from. 

A two sentence summary would go something like this...

Our brains distinguish between direct and reflected sound at higher frequencies but it's impossible for EQ to change the reflected sound without also changing the direct sound. In the low-bass region the reverberant sound field in the room totally dominates what we hear so a simple measurement and correction works well here. (The video covers this and other things far better whilst honestly still being a pretty engaging watch. Not everyone would agree with everything Floyd Toole says mind.)

Do bear in mind my comment about about Dirac Live's multiple measurement position approach being far 'smarter' and so it's not a clear-cut right or wrong type of situation which is why listening and going with what each individual enjoys most is the way to go. 

asx77 wrote:

tallbeardedone wrote:

 To give an example, if an upright bass was being played in my room, 60hz would be boosted, so when I hear that boost due to room it sounds more like the bass is really there.

 

but do you want it to sound like the bass is in your room or do you want it to sound like you're  sat in front of the stage? Both are acceptable answers. 

Good question, and from my last eat listening sessions I think the answer is my room. It sounds more real, like it’s RIGHT THERE. And in regards to your other point of only using dirac for the bass (up to ~250hz), then to me there’s not much point. I like the bass response at my current speaker positioning, always have, it was the flat frequency response at the listening position between 2-8khz that I want to experiment with attenuating. But as both you and Toole (I’ve read his entire article on the topic of room correction) say,, Dirac can’t differentiate between the direct and reflected sound at high frequencies so it seems to me that although, yes, it attenuates the highs, it also takes the “realness” away by over processing the signal in some way. The three dimensional imaging I was getting, the feeling of stepping “through the screen” and “into the room” of the recording and “seeing” the performers in three dimensional space is just … gone. It’s a cliche, but it sounds like a veil has been thrown over the music. 
  If anyone has experienced a similar phenomena and has any suggests on what I can do to fix that, (I.e. remeasure with more points, broader area, smaller area, different curve, etc) please let me know. 

tallbeardedone wrote:

But as both you and Toole (I’ve read his entire article on the topic of room correction) say,, Dirac can’t differentiate between the direct and reflected sound at high frequencies 


 

To be clear, I did not say this, and nor have I ever seen any specific comments from Floyd Toole on Dirac LIve. As I tried to above, it is very important to distinguish between making a measurement at a single point and EQing the response to be a desired shape (a bad idea) and what Dirac Live does. The multiple measurement locations(and I believe some time-windowing) are key, and allow some degree of differentiation between direct sound and the room's contribution.

Re. being happy with your bass response just try what I suggested. You won't know if you can get a sound you like more if you don't try.

What number and spacing have you used between your measurement positions? Too tight a distribution is likely to exacerbate the main issue you've described. A minimum of 50 cm from the central measurement position is recommended to the others a believe.