Smaart 7 target curves

Re: Smaart 7 target curves

Do anybody here have a Reinforced Contour target curve for Smaart?

I wish I knew how to make it myself, but I can't find any description about how to do it.
I am not sure what you are asking for, but if it is a "target curve for reinforcement systems", then my "go to" is as flat as possible from around 100hz to as high as the cabinets can go without a lot of boosting on the top end.

This is using the transfer function-NOT RTA.

I feel this gives a "clean white canvas" that accurately reproduces what the sound guy is doing, so he can "paint" whatever sonic picture he wants, without having to "counteract" any specific "sounds" of the PA system.

I know others have different ideas.

For a PLAYBACK-system, this can often be different, depending on the particular type of playback, the level of the playback and so forth.

There is no "one size fits all" .
 
Re: Smaart 7 target curves

This is one of my standard cut and pastes of what I usually write regarding system tuning.

The problem is the definition of flat. Flat in relationship to what? That is why we use a program like SMAART, what you really want is linearity. Ideally what goes in is what comes out. That is how a system should be tuned, in my opinion. Then you can deal with tone shaping on the individual inputs or groups. I do my vocal mic tuning in a group for each type of mic to get the best gain before feedback. This way I get a much more accurate instrument and playback sound.
 
Re: Smaart 7 target curves

This is one of my standard cut and pastes of what I usually write regarding system tuning.

The problem is the definition of flat. Flat in relationship to what? That is why we use a program like SMAART, what you really want is linearity. Ideally what goes in is what comes out. That is how a system should be tuned, in my opinion. Then you can deal with tone shaping on the individual inputs or groups. I do my vocal mic tuning in a group for each type of mic to get the best gain before feedback. This way I get a much more accurate instrument and playback sound.
Agreed that each input should be adjusted as needed (tonally).

When I say flat-I mean typically +/-3dB (or less) from around 100Hz to at least 10Khz.

When people talk about a "house curve -or X curve etc" they usually have no idea what they are talking about, what the "house curve" is-why it is used and for what purpose.

If you want to be accurate, then having the higher freq at a lower level (house curve) than the lower freq means that the signal IS NOT ACCURATE, but rather being tailored by the speaker system.

But some people like a blurry copy----------------
 
Re: Smaart 7 target curves

For a speaker to sound correct it should measure flat in an anechoic chamber, if we use an FFT measurement with a sample time long enough to capture the low frequencies in a typical live sound environment it will capture other stuff as well.

Our hearing does all sorts of cleaver stuff and our perception does not quite match what Smaart measures. It’s tricky, we as sound engineers have to match our hearing’s perception with EQ.

That’s why we tend to prefer systems with in-field Smaart measurements that are not quite flat and appear to have transfer functions that have a gentle low frequency boost.

… and that's why some measurement programs are now starting include“Target Curves” that are not "flat".

I would just make sure what your are doing has a smooth Smaart trace, has a gentle boost in the low end, and sounds natural. There will of course be some things that you can not fix with EQ, and should not try to fix ... so listen.

FWIW In 2015 I'm seeing more and more people listening (and mixing) with their eyes :roll:
 
Last edited:
Re: Smaart 7 target curves

That’s why we tend to prefer systems with in-field Smaart measurements that are not quite flat and appear to have transfer functions that have a gentle low frequency boost.

… and that's why some measurement programs are now starting include“Target Curves” that are not "flat".

I would just make sure what your are doing has a smooth Smaart trace, has a gentle boost in the low end, and sounds natural. There will of course be some things that you can not fix with EQ, and should not try to fix ... so listen.

FWIW In 2015 I'm seeing more and more people listening (and mixing) with their eyes :roll:

Actually that low frequency "boost" you mention has more to do with the position of our listening plane (our ears) being close to the floor, consequently we are hearing the boundary effect on the low end response of the system. Our ears have grown accustomed to this phenomenon, hence why it's desirable to hear sound systems tuned this way. Get yourself a 12 ft stepladder, and listen/measure the system at a seated listening height vs. 12 ft in open space, and you will see a change in the LF, even a few dB.

Which opens up a measurement can of worms....put your measurement mics at ear height, so that you can see the boundary effects of the floor, but at the same time you're having to deal with the cancellations from the floor reflections you typically see in your measurement. It is more challenging to window out those reflections while getting sufficient resolution in the LF because the reflection path is relatively short. Or put the mics up really high off the floor, now you can window out the reflections without taking a "hit" on the LF resolution (because you can have a larger time window), but you no longer see the boundary effects in your measurement. Or simply use MTW in Smaart, put the mics at ear height (or do a ground plane measurement) and rely on the coherence curve to help you navigate your way around the limitations of your measurement.
 
Last edited:
Re: Smaart 7 target curves

It should be noted that the target curves in Smaart only work in Spectrum mode. You can create your own by simply modifying a captured spectrum measurement. Keep in mind that you want your mic levels (and target curves) to be properly calibrated to dB SPL.

Where I find target curves to be handy is in setting up cinema systems to the X-Curve, or to display NC rating contours to determine the noise floor of the room, or to properly optimize a noise masking system. Merlijn van Veen produced some interesting target curves that illustrate the effects of our hearing at different SPL levels (Fletcher Munson curves), you should check out his website.

If you want to do a target curve in a transfer function, simply capture a trace there, then edit it to your liking using the 0 dB line as your median. I've done stuff like doing a + and - transfer curve of varying tolerance for system commissioning reports, and superimposing the measured trace with upper and lower limits.

I should point out that you do not need to enter in all the data points for 48th octave resolution, unless you're really anal about entering data and have lots of time to burn. I often enter data in at 1/3 octave or 1 octave intervals, Smaart connects the dots in between.
 
Re: Smaart 7 target curves

I found a way of making my own :)

What I'm using it for is as target curve in RTA mode as a visual reference. I have a curve I target for when I mix shows and need to rely on Smaart for spectral balance(read: terrible mix position) and I can get a mic in a good position to capture the PA. Basically it's what L-Acoustic calls Reinforced Contour, flat from 1K and up, +3dB/oct from 1K and down.

This is not a cuve I use as a target for tuning a sound system, although it must be said that if I tune for my own shows, I prefer a linear sound system.
Usually if there is someone else mixing, I give them more low mid and sub than I prefer, Its usually what people ask for anyway. You will often find two EQ overlays in my Lake, one linear and one boosted version.

We can probably argue untill the cows come home about the "correct" tuning, I find that giving people what they want instead of arguing keeps people happy and pays my bills :)
 
Re: Smaart 7 target curves

Over the past 10 years we have cataloged the end result of many engineer's system tuning. The average result is an inverse Fletcher-Munson curve.
 
Re: Smaart 7 target curves

Actually that low frequency "boost" you mention has more to do with the position of our listening plane (our ears) being close to the floor, consequently we are hearing the boundary effect on the low end response of the system. Our ears have grown accustomed to this phenomenon, hence why it's desirable to hear sound systems tuned this way. Get yourself a 12 ft stepladder, and listen/measure the system at a seated listening height vs. 12 ft in open space, and you will see a change in the LF, even a few dB.

Which opens up a measurement can of worms....put your measurement mics at ear height, so that you can see the boundary effects of the floor, but at the same time you're having to deal with the cancellations from the floor reflections you typically see in your measurement. It is more challenging to window out those reflections while getting sufficient resolution in the LF because the reflection path is relatively short. Or put the mics up really high off the floor, now you can window out the reflections without taking a "hit" on the LF resolution (because you can have a larger time window), but you no longer see the boundary effects in your measurement. Or simply use MTW in Smaart, put the mics at ear height (or do a ground plane measurement) and rely on the coherence curve to help you navigate your way around the limitations of your measurement.

Exactly - when measuring or developing new DSP settings for a speaker I hang it 4 m (12 ft) in the air for all of those reasons ... and when EQ-ing a system I also try to listen as well as measure :)

Do you have any screen shots that compare your multi time window measurement with a standard FFT measurements of different window sizes?
 
Last edited:
Re: Smaart 7 target curves

When I say flat-I mean typically +/-3dB (or less) from around 100Hz to at least 10Khz.

Come on Ivan! :D~:-D~:grin: You know that +/-3db doesn't mean anything without indicating what the smoothing is.... :razz:
(And the resolution of the measurement affects the resulting smoothness as well).

Over the past 10 years we have cataloged the end result of many engineer's system tuning. The average result is an inverse Fletcher-Munson curve.

Ed, can you elaborate on this? I'm curious how you cataloged and averaged. Also, did you keep track of the average db level of the various mixes? That's probably a hindsight kind of thought...
 
Re: Smaart 7 target curves

Come on Ivan! :D~:-D~:grin: You know that +/-3db doesn't mean anything without indicating what the smoothing is.... :razz:
(And the resolution of the measurement affects the resulting smoothness as well).


.
True. I dare anybody to get any loudspeaker to +/-3dB with no smoothing. Go ahead-use as many filters as you want----------

Yes it takes more "non simple" numbers to describe something.
 
Re: Smaart 7 target curves

True. I dare anybody to get any loudspeaker to +/-3dB with no smoothing. Go ahead-use as many filters as you want----------

Well, you CAN get some pretty smooth looking low frequency measures if you put the mic really close to the speaker. But what are you measuring at that point? Headphones?

Sorry for the ribbing. Trying to be... comically pedantic?

For the purpose of this thread the "smoothing" should actually be very wide. A reference curve is not likely to have any narrow peaks or dips.
 
Re: Smaart 7 target curves

Well, you CAN get some pretty smooth looking low frequency measures if you put the mic really close to the speaker. But what are you measuring at that point? Headphones?

Sorry for the ribbing. Trying to be... comically pedantic?

For the purpose of this thread the "smoothing" should actually be very wide. A reference curve is not likely to have any narrow peaks or dips.
If you take all the smoothing off,even subs will get pretty ragged.
 
Re: Smaart 7 target curves

If you want to be accurate, then having the higher freq at a lower level (house curve) than the lower freq means that the signal IS NOT ACCURATE, but rather being tailored by the speaker system.

But some people like a blurry copy----------------

Hmm...I would think it means that the system/room relationship is not optimal. Input signals from most sources are pretty 'balanced' from the provider....music tracks, keyboard patches, voices...rtf.
 
Re: Smaart 7 target curves

Over the past 10 years we have cataloged the end result of many engineer's system tuning. The average result is an inverse Fletcher-Munson curve.

Fletcher-Munson has been debunked as derived from a poor sample: All men, and all men who had been exposed to ordnance!
 
Re: Smaart 7 target curves

Fletcher-Munson has been debunked as derived from a poor sample: All men, and all men who had been exposed to ordnance!
Well...all the engineers were men and most engineers hearing could be categorized as having been exposed to ordinances!!!:lol:

To answer an earlier question, we took Smaart transfer function snapshots when the mix hovers around 100dbA.

Anecdotally , these results were from various multi-day music festivals. When we took measurements for Jazz and Classical shows, the average results were "more flat"