Dirac Live - First Impressions

[JMac]

So my UMIK arrived today so I've spent the afternoon playing with Dirac Live and I have to say I'm not liking it.

There's no doubt its helped focus the image a little (my speakers were already quite well setup from an alignment perspective as I used a laser measurer to ensure they were both angle and distance equal from the 'sweet spot' ) and its move the stage more central.

Now, on some recordings its brilliant, for example the album Lana Del Rey, Born to Die sounded awful before but now sounds much better. BUT and its a BIG BUT, its robbing the system of virtually all the low and mid bass. Looking at my original graph there's a massive humps in the lower frequency range and the new graph gets most frequencies to pretty much flat response but the tracks just sound bad. I play a lot of electronica which have bass heavy rhythms and music where there used to be a low rumbling under bass now have nothing. Its like swapping out a cannon for a pop gun. The response may be flat, but thats not what the track should sound like.

I'll play around some more tomrrow and post my graphs but anyone have any suggestions?

 

[JMac]

Looking around the Web this seems to be quite a common complaint. It's not that something is wrong, just that flat bass sounds like no bass if you know what I mean. Most people seem to dial the humps back in as listening enjoyment doesn't always correlate with a flat response

 

[CnoEvil]

I haven't heard Dirac Live....but have not preferred any DSP that I have heard on 2 Channel.

On my AVR600, I much prefer Stereo Direct.

 

[MajorFubar]

I have no experience of Dirac Live, but your response seems so much like that of folks who hear a truly flat pair of speakers or headphones for the very first time. It's a bit of a shock to the system. When you can't (or do not want to) create stadium-level volumes in your living room, totally flat can definitely sound, well, totally flat, not least because our ears are less sensitive to frequency extremes at lower volume levels, hence the once-popular loudness/contour button. But hey there's no harm in dialing-in a little bit of lift to put the fun back in.

 

[ellisdj]

you don't want a flat target curve or response if you have manually set it to be flat that's problem number 1.

a nice target is +6db at 30hz rising from 500hz in a straight line.

now I don't know what speakers you have and how realistic that is but that's a nice target if they can do it.

If not just dial it back keep the target curve angle but stop it shorter than 30hz if your speakers roll off at 50hz for example keep the target line at that angle but not all the way to 30hz.

if you have big humps in the freq response your only going to get 1 or 2 note bass like that.
Big peaks in the bass response is the definition of one note bass. not trying be horrible trying to help

 

[ellisdj]

the dirac stock curve plays it very safe.
but if you set the curve correct you will get the best accurate bass you can from your speakers.

also make sure to take your measurements in a large area not a small area.

take the measurements right set the curve right it's perfect every time but it can't work a miracle at the same time.

of you post photos of what you have done I can help

 

[insider9]

I haven't heard Dirac Live....but have not preferred any DSP that I have heard on 2 Channel.

On my AVR600, I much prefer Stereo Direct.

In which case it was done wrong or you used a wrong solution. Just an impulse correction would improve what you have by a large margin not even touching frequency response.

 

[JMac]

No problem. I'll post my initial curve up tomorrow (just on my way to bed now) and you can tell me where to raise/drop. I was using the 'optimise' option and just letting it pretty much aim for zero.

When you say take your measurements in a large area though, they're quite specific on the mic positions on the chair which are all quite close together. Wouldnt widening that skew the intention.

 

[Strictly Stereo]

...The response may be flat, but thats not what the track should sound like.

I'll play around some more tomrrow and post my graphs but anyone have any suggestions?

It is impossible to be sure at this distance, but very possibly that is exactly what the track should sound like, within the constraints of your speakers’ frequency response. The peaky bass and the rumbling you are accustomed to are most likely caused by the room. The contrast can be quite jarring.

Did you dial in a completely flat target curve? Try the target curve suggested by Dirac Live instead.

 

[Strictly Stereo]

Looking around the Web this seems to be quite a common complaint. It's not that something is wrong, just that flat bass sounds like no bass if you know what I mean. Most people seem to dial the humps back in as listening enjoyment doesn't always correlate with a flat response

True. Can you name a track or two and point out where and when you feel there is something missing? Preferably something that I am likely to own or that I can find on TIDAL.

 

[ellisdj]

No problem. I'll post my initial curve up tomorrow (just on my way to bed now) and you can tell me where to raise/drop. I was using the 'optimise' option and just letting it pretty much aim for zero. 

When you say take your measurements in a large area though, they're quite specific on the mic positions on the chair which are all quite close together. Wouldnt widening that skew the intention. 

yes you are right but the forward in the chair measurements ideally need to be forward and the high ones high.

they suggest a half metre area minimum a larger area will mean less is taken off as a basic rule of thumb.

the magic of dirac is it's flexibility in tailoring a sound to how you like it

the trick is to be sensible and not push things too crazy depending on how the initial measurment looks and what room effect you can and can't use

I assume you used the correct specific calibration file for your UMIK1 and the correct one 90degree version if you pointed the Mic vertical ??

 

[Deleted member 108165]

You are not alone with your initial reactions JMac, I tried it last year and didn't like the results either. I found it rather a blunt tool that needs too much tweaking to get it right... it's not a quick-fix tool. Personally I have two thoughts on the subject:

1. Unless you have the time to devote your undivided attention to it the two week trial period is nowhere near enough to fully investigate its advantages/disadvantages.

2. It's too expensive, £100 maybe, but not £340. I'd rather buy more music.

Insider and Ellis are the guys for this so I will now bow out.

 

[andyjm]

JMac,

As pointed out above, your reaction is very common. After decades of people saying they want a flat frequency response, and it to 'sound like it did when it was mixed' - now they can actually achieve it with DSP it turns out that they don't want it at all.

There are lots of arguments about why this may be the case. Familiarity with a 'boomy' system before the corrections were applied, an assumption in the mixing process that the end users system will be boomy (or perhaps the control room acoustics weren't that great) - and so on.

Discussions around this have been going on for a long time in home cinema circles where DSP correction has been much more common. A consensus has developed around a curve that generally sounds right, 'the house curve'. If you google this, you will find pages of comments and descriptions of the right curve to have. Rather as described above, it is a downward sloping curve with LF lift.

From an engineering perspective, it raises a number of questions. Should the recording engineer assume a flat response in the listeners system, or rather like RIAA equalisation, should they assume a specific cast to the sound and compensate accordingly. What about the specifics of the control room acoustics, or the bombed out hearing of the engineer and artists. Without understanding all the equalisation steps and compensation that has taken place along the way, it is almost impossible to say what is 'right'.

At the risk of joining the 'trust your ears' brigade, there probably is an argument for just adjusting it until it sound good to you.

 

[lindsayt]

Wouldn't the best sounding frequency response depend on the listening volume?

Fletcher Munson?

Turn the volume right up to rock concert levels and a flat bass response might sound fine?

At lower, normal home listening volumes, a rising bass, as Ellisdj pointed out, would make sense.

 

[ellisdj]

it is Fletcher and Munson a flat looking graph doesn't sound flat.

a house curve graph actually sounds flat.

£340 is cheap for such a product you don't realise how lucky we are to have this for this price.

you initially had to pay £20k for a product if you wanted dirac put that into perspective.

you can do a simple test play a 90hz test tone then a 30hz test tone.
the 30hz tone will seem much quieter.
The amount of db you need to increase the 30hz tone by in order for it to be perceived as the same volume as the 90hz test tone is the amout of +db you need on the graph / target curve at 30hz

 

[ellisdj]

there is also 2 lots settings to pay attention to.

In the DAP where you select your saved set of filters you want to use. There is a very small left arrow and right arrow.

the left arrow takes you to the gain control this can normally be raised a bit watch the clip indicator and raise the level until you have some clipping and back down slightly so there is no clipping.

the right arrow takes you to level gain and delay controls.
dirac sets things to be ultra focused but actually prefer to disable that section of what it does so it's another thing to listen to and check

 

[ellisdj]

one other thing to consider strictly stereo mentioned this briefly.

do you have you speakers in the best spot.

You potentially need to think differently in regards to speaker placement.

with HiFi the reaction to boomy bass is pull the speakers away from the boundary. I have mentioned this before this can be backwards thinking as while you reducing the boom your reducing the rest of the bass range at the same time.

with dirac you have bass management you don't need to worry about bass boom

in fact it could be the more bass boom the better you put the speakers in the spot that gives the bass that is closest to the target curve you want I.e. +6db at 30hz.
if some of the bass range is above the curve dirac will fix it but it can't really fix leaner bass not really - all it can do is put what's there into better balance.

I think once you know the sound curve you want in your room and what bass you get at your MLP this will influence what speakers buy.

sorry for keep going on but it's all important

 

[andyjm]

it is Fletcher and Munson a flat looking graph doesn't sound flat.

a house curve graph actually sounds flat.

£340 is cheap for such a product you don't realise how lucky we are to have this for this price.

you initially had to pay £20k for a product if you wanted dirac put that into perspective.

you can do a simple test play a 90hz test tone then a 30hz test tone. the 30hz tone will seem much quieter. The amount of db you need to increase the 30hz tone by in order for it to be perceived as the same volume as the 90hz test tone is the amout of +db you need on the graph / target curve at 30hz

Well yes, and no.

Presumably the mixing engineer was human, and therefore subject to the same Fletcher / Munson effects. So by adjusting your system to level this out, you are now boosting the LF relative to the experience of the engineer - unless his system had an equivalent lift.

Of course F/M is level dependent, so perhaps the engineer was blasting it out in the control room, and at home the more modest listening levels means that bass sounds lean. Then of course there is the guy who EQ'd the track for the CD who mave have had an entirely different listening level / setup.

The much maligned 'loudness' button was specifically included to compensate for this, so in the DSP world, one could consider different house curves for different listening levels.

If there is any merit to the MQA system, it is that these multiple steps in the chain are controlled and compensated for so that the final listening experience does bear some realtionship to the original artisit's performance. Allegedly.

 

[CnoEvil]

I haven't heard Dirac Live....but have not preferred any DSP that I have heard on 2 Channel.

On my AVR600, I much prefer Stereo Direct.

In which case it was done wrong or you used a wrong solution. Just an impulse correction would improve what you have by a large margin not even touching frequency response.

Possibly.

As a family, we all played musical instruments and would occasionally play stuff as a family. The sound didn't need DSP to sound as good as our talents would allow.

In my case, I have used the EQ provided by my AV receiver, through its microphone.

I judge a sound by characteristics that can't be measured.

 

[ellisdj]

your balancing the sound to your own ears
in order to hear a flat bass all bass frequencies must be at the same percieved volume
you can probably take the time and check 20hz 30hz 40hz even 21hz 22hz if you want to be totally exact.

if they are all the same perceived volume as 100hz then you have a flat sounding bass.

don't worry about what the mixing engineer did you can't you have no control over any of that.
Having bass like this at home is about as good as you can set it up. the rest is then down the amount and quality

 

[ellisdj]

it's back to what I posted a few back.
the trick is knowing what curve you want i.e I want +6db at 30 Hz or +4db at 50hz as an example

you then look at your measurements does where I have my speakers bass response give me close to the bass response I know I need.
If they don't you need to move them or get speakers with a bass output more in line with what you need.

that's what I meant by it will have an influence on what you do.

 

[JMac]

...The response may be flat, but thats not what the track should sound like.

I'll play around some more tomrrow and post my graphs but anyone have any suggestions?

It is impossible to be sure at this distance, but very possibly that is exactly what the track should sound like, within the constraints of your speakers’ frequency response. The peaky bass and the rumbling you are accustomed to are most likely caused by the room. The contrast can be quite jarring.

Did you dial in a completely flat target curve? Try the target curve suggested by Dirac Live instead.

Thats quite possible, I used the curve suggested by Dirac to 'optimise' and it dialled most of the bass down to near flat. The problem is, that whilst numerically correct, that probably isnt how the music is meant ot sound. Most of what I listen to has been mastered to be played in clubs via DJ's over large multi speaker systems. I know it may not be pure hifi in the truest sense of the work but those bass humps are necessary to a degree. A pure flat frequency response may be perfect from a theory perspective but it probably isnt how the artist intended it to be listended to back in the real world. If that makes sense.

 

[JMac]

Looking around the Web this seems to be quite a common complaint. It's not that something is wrong, just that flat bass sounds like no bass if you know what I mean. Most people seem to dial the humps back in as listening enjoyment doesn't always correlate with a flat response

True. Can you name a track or two and point out where and when you feel there is something missing? Preferably something that I am likely to own or that I can find on TIDAL.

I'll have a look around on tidal and see what I can find.

 

[JMac]

[st my initial curve up tomorrow (just on my way to bed now) and you can tell me where to raise/drop. I was using the 'optimise' option and just letting it pretty much aim for zero.

When you say take your measurements in a large area though, they're quite specific on the mic positions on the chair which are all quite close together. Wouldnt widening that skew the intention.

yes you are right but the forward in the chair measurements ideally need to be forward and the high ones high.

they suggest a half metre area minimum a larger area will mean less is taken off as a basic rule of thumb.

the magic of dirac is it's flexibility in tailoring a sound to how you like it

the trick is to be sensible and not push things too crazy depending on how the initial measurment looks and what room effect you can and can't use

I assume you used the correct specific calibration file for your UMIK1 and the correct one 90degree version if you pointed the Mic vertical ??

[/quote]

Yes, calibration file loaded for standard horizontal positioning.

 

[JMac]

Whats the best site these days for hosting and linking free images to the site?