Target Curves

[Le Concombre Masqué]

12 hours ago, jaaptina said:

 

Hi, what color belongs to what measurement (resp. Synthesis, RR1, B & K, Bob Katz)? Thx! 

at 1  K Hz

 

dark green Katz

Purple B & K

Brown RR1

lighter green JBL Synthesis

 

To be thorough ; I eQ to measurements done with Moving MIc Method

 

 

[Le Concombre Masqué]

6 hours ago, cjf said:

For the time being, I've settled on a curve that has a 2db bump at 20hz and is flat to 1khz and then ends at -10db down at 20khz in my room which has zero physical treatments and is quite large and open. This seems to work pretty well so far but at some point I may reduce the correction a touch to let the main speakers return a little more of their natural sound which is a bit more realistic sounding than the current curve with room correction in place. All in all, I like the corrected room much better as a whole but some may find it on the duller side.

 

Target curve shown in Red via Audiolense...Actual measured in room response shown via multi colored REQW curve.

 

 

evaluation of dullness etc is not trivial

 

First to which target has the album been mastered to ? Apart from Bob Katz nobody I know of disclaims

 

ie Jamal s Ballades is unbearably bass heavy with any Harman type target : this record would be more enjoyed with B&K and maybe even Katz or ITU EBU type

 

ECM I have explored so far favour B&K too

 

But if you stay in Europe but go to a Savall recording, HArmans will suit it better. Same with MOFI SACD

 

So I think we need at least two targets ; B&K and RR1 should be first to have

 

yours seems an average of the two types

 

Then subjectivity may come back depending on criteria

 

ie Savall's Ibn Battuta has its charms with a RR1 eQued to a REW simple average ( eQ to the mandatory for phases Vector Average is unbearable IMO) if your criteria are surgical precision minute details of ambience etc. But in the hand it's stressful tiring etc when you listen to any kind of music.

 

That lets me very dubitative about commercial softwares and @mitchco' s offer based on average of several measures. I know the answer pertains to fine tuning of FDW based on room dimensions, steady state timing of first reflections etc etc  etc. The MMM based eQ is cumbersome but much better IMO and there's a several dB difference above 500 depending in the measurements depending on MMM vs Averages...

 

 

[Le Concombre Masqué]

On 12/4/2019 at 11:25 AM, STC said:

The measurements at 6 Finnish concert hall were the measurements of direct and indirect sound. What you want is measurement at your listening position to be like that after applying whatever position, room treatments, DSP, EQ , convolution filter and etc etc. 

 

But that doesn’t mean all your albums would sound good. Concert hall slopes are best for classical. 

sorry I missed your post. Yes you're right and it's even much too much damping with chamber music, be it piano or quartet. I gave such a target a try but never use it

 

It s interesting to note that a downward slope is not just natural for our living rooms but for professionally designed large concert halls as well. Maybe IRCAM at Beaubourg could twist the sound that much but I doubt any concert venue we go to measures flat in 2 dB

[Le Concombre Masqué]

what to eQ to is as important as the target.

 

About MMM : https://www.ohl.to/audio/downloads/MMM-moving-mic-measurement.pdf

 

not so great for LF 

 

I'll give 30/40 ms IR window a try ; looks like the enveloppe of sweeps is then in the MMM ballpark

[semente]

On 1/10/2020 at 3:05 AM, cjf said:

For the time being, I've settled on a curve that has a 2db bump at 20hz and is flat to 1khz and then ends at -10db down at 20khz in my room which has zero physical treatments and is quite large and open. This seems to work pretty well so far but at some point I may reduce the correction a touch to let the main speakers return a little more of their natural sound which is a bit more realistic sounding than the current curve with room correction in place. All in all, I like the corrected room much better as a whole but some may find it on the duller side.

 

Target curve shown in Red via Audiolense...Actual measured in room response shown via multi colored REQW curve.

 

 

 

 

May I suggest that you reduce the vertical (dB) scale to the more common/standard 50db-wide range?

 

It's difficult to assess any anomalies when everything looks flat.

[cjf]

9 hours ago, semente said:

 

May I suggest that you reduce the vertical (dB) scale to the more common/standard 50db-wide range?

 

It's difficult to assess any anomalies when everything looks flat.

The link in my Sig has other graphs including an un-smoothed version of the full sweep previously posted. Unfortunately  thats the best I have on hand at the moment. I've never considered shrinking the vertical  scale but will keep it in mind next time I take some measurements. The un-smoothed measurement shows the as the Blue Line being all channels combined and then the Red/Green lines being just Left or Right main with the paired Left or Right sub.

 

In any case, I can say its pretty smooth either way I shrink the scale but thanks for the suggestion and I will keep it in mind next go round.

[Le Concombre Masqué]

3 perspectives on the 3 convolutions I will use from now on (JBL Synthesis, RR1, B & K ; psychoacoustic smoothing of L+R)

 

All are smoothed, slope downward and have optimised step and impulse : none is bad whatever the recording and the target that might have been in use in the mastering room. Yet as dB differences show they don't sound alike and there are better convolution/mastering fits.

 

If I had only one to pick it could be good ol' B & K that appear to have been a long time standard. But with Harman's promotion of RR1 and Synthesis they have their best fits too

 

Maybe the most relevant view is the one bundling curves in 1 dB above 160 Hz : the obvious more or less bass and sub bass criteria is not that operational and how vocals are infatuated or not tends to be the go/no go point for RR1 (above the 2 others between 160 and 500 Hz) vs either BK or Synthesis that are equivalent in that region. Also, the extra mids from 800 to 4 K of the Synthesis often act as an appealing reason to go Synthesis even if there's a touch too much bass, because of extra presence and delineation of soundstage, instruments and vocals separation

[Le Concombre Masqué]

7 hours ago, cjf said:

The link in my Sig has other graphs including an un-smoothed version of the full sweep previously posted. Unfortunately  thats the best I have on hand at the moment. I've never considered shrinking the vertical  scale but will keep it in mind next time I take some measurements. The un-smoothed measurement shows the as the Blue Line being all channels combined and then the Red/Green lines being just Left or Right main with the paired Left or Right sub.

 

In any case, I can say its pretty smooth either way I shrink the scale but thanks for the suggestion and I will keep it in mind next go round.

FYI,

 

Left graphs show the results of convolving the Vector Average by an Impulse that smoothed it (big bass bump, straight line from 160 ; does not sound the best though has its charm if you like ultra analytical)

 

Right graphs show the results of convolving the Vector Average by an Impulse applying B & K target curve to the non FDW scalar average (impulse I use)

 

PS : sorry it got mixed up and names don't show : The Right graphs come first ; R phase shows the use of Pano shuffler in RePhase (ondulations) 

[Le Concombre Masqué]

HARMAN STANDARD RR as presented here : https://www.innerfidelity.com/content/warren-tenbrooks-summary-head-measurements-harman is a much worthy addition to my quiver.

Did not want to try for at first look it's awkward : big bass bump then flat

In fact it's a JBL Synthesis going flat from 3 to 20 K

Of course, no harshness (for stridency look in the 600 +- 200 Hz region), never understood the concept of dip in trebles. Trebles mean presence brillance and punch to the bass notes that do have attacks in that range. Now it does not suit every mastering, see the relevant thread : 

 

[Le Concombre Masqué]

Harman std is definitely a target to try if you have enough LF to balance the HF extension

 

Interestingly enough, it's the target that yields a non FDW (so including room's reflections) average looking like a FDW one, after convolution to said target

[Le Concombre Masqué]

17 hours ago, Le Concombre Masqué said:

Harman std is definitely a target to try if you have enough LF to balance the HF extension

 

Interestingly enough, it's the target that yields a non FDW (so including room's reflections) average looking like a FDW one, after convolution to said target

Consequently ? convolving the Vector Average by an Impulse applying the Harman std target curve to the non FDW scalar average yields better results than with ie B & K.

 

Beyond the amplitude correction those cleaner time domain results might account for my excitement for that overlooked target

 

graphs for the Right channel for comparison:

[Le Concombre Masqué]

In the following graphs (Left ch only) I convolve  my "sweep measure number 2 L" by an Impulse applying the Harman std target curve to the non FDW scalar average for Amplitude correction with phase corrected using the FDW Vector Average

 

I have found the use of FDW Vector Average mandatory to have a workable phase and so renounced to use sweep measure number 2

 

what makes sweep measure number 2 non FDW special is that it matches almost perfectly the scalar average of 15 non FDW measures as well as the measure obtained by MMM. 

 

So sweep measure number 2 is the measure I have with both time and frequency domains informations and that ressembles the most a relevant actual signal.

 

It thus seemed interesting to look at and share the graphs:

 

[Le Concombre Masqué]

CURIOUS TO KNOW HOW IT LOOKS LIKE if a FDW (see capture) has been applied to the convolved sweep measure number 2?