what when i like when i hear but FR is not flat

[mga2009]

hi,

i have a couple of diy speakers (diysg fusion 6) that i never really like how they sounded, as the highs sounded a little muffled, altough its FR looked relatively flat.

yesterday i had time and i messed up a little with the crossover , the same designer had a similar design with same components so i kind of made a mix of both.

i doubled the tweeter cap to 2.7uf (was 1uf) and i added an L pad on the tweeter (compression driver).

i know very very little on xo design, but i seem to had lowered a little the XO point (was 2400hz i think) and made the highs a little louder...

the thing is that i kind of more happy with the result. highs seem more edgy, razory, clear... i like it.

the thing its the FR obviously shows higher levels on the high end, up from the XO point, so its not ruler flat.

so the thing is... is there anything wrong? should i aim for a flat FR or for a sound i like more? is there anything wrong from the SQ point of view? i really like my plot running the highs a little hot compared to the mid and sub range.

IDK, just looking to your opinion. i can upload the plots if you want.

 

[Galu]

At the end of the day, no loudspeaker is perfect (why do you think there are so many different ones to choose from?) and you are at liberty to prefer the sound of your modified speakers.

As long as the downwardly extended treble response does not overload, or excite resonance in, your compression driver (you would hear the distortion) then you are good to go.

Many commercial speakers are designed to have a prominent treble which gives them 'showroom appeal'. Some may find the treble to be fatiguing when listened to over longer periods back home. Just rely on your own ears to tell you what is correct for you.

 

[PeteMcK]

"highs seem more edgy, razory, clear" - exaggerated high frequencies sounds appealing, but is likely to be aurally fatiguing after a while.... but if you like the overall balance, stick with it.
'a little hot' - how many dB?

 

[Douglas Blake]

Lowering the crossover point on a speaker can have some repercussions.

First it can lead to overheating a tweeter or driver that is not intended to operate in the new range.

Second it can lead to the woofer and tweeter (in a two way system) overlapping and, depending on crossover design, this can cause both phase problems and uneveness in the overlap range.

In the end ... so long as nothing is being damaged and you like the sound... go for it!

 

[mga2009]

thanks for your replies.

i really like the "new" sound, and the changes seemed minimal and based on a speaker with the same drivers so i hope i wont damage anything (also, the same CD driver is used in another designs with a much lower XO point)

the FR plot look nice although with a V shape (high low end and high high end).

regarding the phase... is there anything i should be aware of in the FR plot?

 

[oon_the_kid]

The technique measuring FR is actually controversial... most people would measure it with a mic 0.5 or 1m on axis. However much of the sound we hear are actually reflected sound. Put a piece of small plank in between you and the speaker and you realise the sound doesn't change all that much. There is also a high chance you are listening off axis.

Hence for some people ( stereophile magazine for example), frequency response is measured as an average across some angle. If you choose this method, then your on axis will a bit hotter than the guy who just measured on axis only.

Oon

 

[boswald]

Your room is probably absorbing or cancelling some of the highs. So you have made them work for you, in your room.
Try another speaker (that measures flat anechoically and has similar dispersion) in your room. You'll likely see the same issue.

 

[midrange]

If you like the sound, and you are not damaging anything that has to be the one to keep. Sitting listening and thinking "I do not like the sound, but I know the FR is flat", is not going to bring you satisfaction.

 

[mga2009]

The technique measuring FR is actually controversial... most people would measure it with a mic 0.5 or 1m on axis. However much of the sound we hear are actually reflected sound. Put a piece of small plank in between you and the speaker and you realise the sound doesn't change all that much. There is also a high chance you are listening off axis.

Hence for some people ( stereophile magazine for example), frequency response is measured as an average across some angle. If you choose this method, then your on axis will a bit hotter than the guy who just measured on axis only.

Oon

I actually measured just as you say... 0.5m on axis, but most of my listening is on axis with the center channel (the speaker in question is a center channel MTM horizontal positioned)... I am positive the measurements taken are not perfect, but I made measurements under the same condition with the default XO to see the impact of the modifications...

Your room is probably absorbing or cancelling some of the highs. So you have made them work for you, in your room.
Try another speaker (that measures flat anechoically and has similar dispersion) in your room. You'll likely see the same issue.

Well... my room must be a problem. All of the testing I've done was in my living room, and I was quite happy with the "new" speaker's performance... so I moved everything to my TV room and connected it to my AVR aaaannnd back to reality... the speaker does not sound that good unless I go up on the volume. The TV room is untreated with lot's of reverb (I know, i am working on some sound absorption panels).

The thing is... is it possible that a room kill the highs on a speaker like it does here?? This is probably not 100% reflected on the measurements I think.

If you like the sound, and you are not damaging anything that has to be the one to keep. Sitting listening and thinking "I do not like the sound, but I know the FR is flat", is not going to bring you satisfaction.

Yep, you are right, but sometimes it's better to "educate" the listening habits hehe.

 

[oon_the_kid]

Hi,

What I meant to say is if it measured flat on axis, it will measure a little weak on the treble if you use the average method, which means you take the measurement at 0 degrees and 10, 20 and 30 degrees of axis. Average them. Since HF is weaker off axis unless you use a constant directivity horn or 1/2 inch tweeter, it will measure lower.

The microphone + software typically has a gated measurement that will only measure the first arrival sound and will disregard all reflected sound. What are you using, REW? You have to remember it is meant to measure speaker response without the room. However our ears will pick up everything. So the perception of sound is not that well correlated.

So if it is measured flat by average method it will be a bit hot on axis. However you would be using the method that is used by stereophile magazine, do it is not wrong and I would say more accurate. A normal musical instrument is more omnidirectional than a speaker.

Oon

 

[mga2009]

Hi,

What I meant to say is if it measured flat on axis, it will measure a little weak on the treble if you use the average method, which means you take the measurement at 0 degrees and 10, 20 and 30 degrees of axis. Average them. Since HF is weaker off axis unless you use a constant directivity horn or 1/2 inch tweeter, it will measure lower.

The microphone + software typically has a gated measurement that will only measure the first arrival sound and will disregard all reflected sound. What are you using, REW? You have to remember it is meant to measure speaker response without the room. However our ears will pick up everything. So the perception of sound is not that well correlated.

So if it is measured flat by average method it will be a bit hot on axis. However you would be using the method that is used by stereophile magazine, do it is not wrong and I would say more accurate. A normal musical instrument is more omnidirectional than a speaker.

Oon

Yes, I am using REW and a miniDSP UMIK.

I now switched from the compression driver to a ribbon tweeter that I had around, just for sake of testing.

Here is a measurement at 1 feet on axis (in front of the tweeter) 1/12 smoothed, i had to upper the tweeter cap to 6.4uf and put the 8ohm L-Pad resistor from PE at the max setting (reverse polarity measured better).

This seemed to measure flatter and it does sound good, maybe not as edgy and aggressive as the CD (with the modified cap and resistor), but more neutral and mellow, but still able to hear small stuff in movies.

The thing is that I can get rid of the dip at around 1500hz, but that's not tweeter problem, but woofer, correct? Maybe I need to start messing with the woofers XO (this is a MTM horizontal positioned central speaker). I will tray off axis measurement and in different positions.

 

[mga2009]

Here is the measurement.

 

[adason]

That response looks pretty good. I see one problem, there is no bass. You are almost 30dB bellow at 30Hz.

 

[mga2009]

That response looks pretty good. I see one problem, there is no bass. You are almost 30dB bellow at 30Hz.

Yes, yes, measurement was taken without subwoofers that take care of that, up to 80hz...

I am mostly concerned on the 10db dip that starts around 1khz and finishes at 1.5khz I might need to lower the inductor on the woofers (now 4mH)

 

[adason]

ok, i see, sorry i did not read carefully, yes with sub its fine

why don't you take an of axis readings (say 0, 10, 20, 30, 40 degrees) and see if you spot the problem at 1-1.5kHz.

 

[adason]

...and one more thing. can you in your fr response go to controls (upper right corner) and unwrap the phase? it will be much easier to see what is going on with phase

have you tried to measure with the tweeter in phase/out of phase and compare?
sometimes what we predict is different from what we measure
maybe the dip is the results of phase mismatch

 

[mga2009]

...and one more thing. can you in your fr response go to controls (upper right corner) and unwrap the phase? it will be much easier to see what is going on with phave

have you tried to measure with the tweeter in phase/out of phase and compare?
sometimes what we predict is different from what we measure
maybe the dip is the results of phase mismatch

adason,

thanks for your help. I will take more measurements. For now, this is the REW measurement taken on axis (onedrive link):

Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

 

[korpberget]

At the end of the day it is you who do the listening, not the microphone.