Kef LS50, Gallo Speakers, brightness and measurements?

Status
Not open for further replies.
Jump to Latest
Z

zenpmd

Member
Joined 2013
#1
How does one learn how to interpret the frequency graphs of speakers? I have recently bought the LS50, which I had to do blind as there is no way of trialling where I live. Everywhere the reviews were great. However, I found on some recordings that they were excessively bright. This forum, I have discovered is the only one that agrees with me, and its backed up by its measurements ( plus 3db in 1.5 to 7hz roughly) An opportunity has arisen to buy some gallo stradas and the sub cheaply as I was hoping someone would help me interpret the numbers... Thanks!
 
#5
Thanks - but what am I looking for in the graphs particulary? Is there a guide to interpretation at all anywhere?
Click to expand...
I prefer a declining frequency response, translated to graphs I prefer the range of 2k-10k about 3db lower compared to the range 40hz-1k.

Where can I read about the ear sensitivity point too?
Click to expand...
Wiki Equal-loudness_contour

A completely flat frequency response is necessary for studio monitors when critical mixing is done.
But at home you're mostly better off with a declining frequency response.


Regards,
Danny
 
Last edited:
#6
I've owned both the LS50s and Strada, though the Stradas too far back to remember well. They did have a more noticeable top end than the Kefs from memory.

I didnt find the LD50s bright by any means, fairly polite in fact. I was running my signal through Fabfilter but wasnt inclined to touch anything above the upper bass.


1.5-7k is quite a bandwidth, its possibly some peoples ears take that as the baseline and interpret everything above and bellow as rolled off, vs finding 1.5-7k elevated.

If you've got a PC source by any chance just Eq it to your liking, much cheaper than swapping out speakers.

Re the 'flat' issue, the language is often unspecific and confusing. I'd contend every speaker, studio monitors and domestic speakers, should and usually do aim aim for flat (on-an off axis aggregated) measures near field/anechoic. This will always translate to a falling FR measures at the listening seat.
 
#7
 
Last edited:
#8
A previous thread about LS50 http://www.diyaudio.com/forums/multi-way/218281-kef-ls50.html

KEF LS50 is a little bright, but I guess that it is intentional because that's just what people expect from a "monitor loudspeaker" It has rather smooth off-axis, and then we get good results with just eq shaping.

If a speaker has ragged off-axis, eq doesn't help because of difference in direct and reflected/power response. Well it may help a little but the difference remains and "spot" will sound very diffrerent than rest of the room.

The pic is taken from here http://www.harman.com/EN-US/OurCompany/Innovation/Documents/White%20Papers/LoudspeakersandRoomsPt2.pdf
 

Attachments

  • direct - radiated Toole.png
    direct - radiated Toole.png
    109.3 KB · Views: 254
Last edited:
#10
Re the 'flat' issue, the language is often unspecific and confusing. I'd contend every speaker, studio monitors and domestic speakers, should and usually do aim aim for flat (on-an off axis aggregated) measures near field/anechoic. This will always translate to a falling FR measures at the listening seat.
Click to expand...

Most of us don't have an anechoic chamber, so measuring in a normal room will give a different frequency response.
A speaker that measures flat in an anechoic chamber room will not measure flat in a living room.
So when comparing frequency responses you should first know where the measurements are taken.
But in my living room I prefer a speaker with a declining frequency response.
 
#11
danny_66 said:
I prefer a declining frequency response, translated to graphs I prefer the range of 2k-10k about 3db lower compared to the range 40hz-1k.


Wiki Equal-loudness_contour

A completely flat frequency response is necessary for studio monitors when critical mixing is done.
But at home you're mostly better off with a declining frequency response.


Regards,
Danny
Click to expand...


You will end up with a lack of energy in your speaker if you have a 3 db trough on axis like that ....
 
#13
danny_66 said:
Most of us don't have an anechoic chamber, so measuring in a normal room will give a different frequency response.
A speaker that measures flat in an anechoic chamber room will not measure flat in a living room.
So when comparing frequency responses you should first know where the measurements are taken.
But in my living room I prefer a speaker with a declining frequency response.
Click to expand...

You should be careful about making broad statements like those in your post. They are incomplete at best, and misleading at worst. You can get anechoic response (or quasi-anechoic) using modern measurement techniques in your living room. I suggest reading a few books, such as Joe D'Appolitto's Testing Loudspeakers.

You will get a falling in-room response if you measure the total sound incident upon the microphone, including the direct sound from the loudspeaker and reflected sound from all the room surfaces. This is because most surfaces will absorb some HF, and as a result, you get the falling response. But if you remove the late arriving reflections, and measure on-axis of the loudspeaker, you should see the pure response of the loudspeaker. In the HF, this is all that matters because the ear will reject late arriving reflections. Ideally, the direct on-axis response should be flat, with off-axis being smooth replica of the on-axis response and smoothly falling in the very high frequencies.

So, even though you may see in one measurement, a falling response, it doesn't mean that is the correct response. You have to provide the context for the measurement as well. Anechoic, quasi-anechoic, in-room, and so on.

The reason for explaining this is that the measurements on the Soundstage Network website are like golddust. You can tell almost completely how a loudspeaker might perform in a room by looking at them. The measurements are highly reliable and lot can be learned from them.

Here is the LS50:
SoundStageNetwork.com | SoundStage.com | NRC Measurements: KEF LS50 Loudspeakers

Look at the first graph. It shows on-axis, and 15 and 30 degrees off-axis. Off-axis looks reasonably smooth, but there is a 3db rise beginning at 1.5 kHz. Now look at the third curve, the listening window. The rise doesn't go away. This is the best approximation of how it might sound. Clearly, it will far too bright. You want a flattish response here.

Some might say that you should listen off-axis. But the bright nature of the sound will still dominate. Check out the KEF Reference 201/2. The same topology with dual concentric drivers, but look at the response:
SoundStage! Measurements - KEF Reference 201/2 Loudspeakers (12/2007)
 
#15
ra7 said:
You should be careful about making broad statements like those in your post. They are incomplete at best, and misleading at worst. You can get anechoic response (or quasi-anechoic) using modern measurement techniques in your living room. I suggest reading a few books, such as Joe D'Appolitto's Testing Loudspeakers.

You will get a falling in-room response if you measure the total sound incident upon the microphone, including the direct sound from the loudspeaker and reflected sound from all the room surfaces. This is because most surfaces will absorb some HF, and as a result, you get the falling response. But if you remove the late arriving reflections, and measure on-axis of the loudspeaker, you should see the pure response of the loudspeaker. In the HF, this is all that matters because the ear will reject late arriving reflections. Ideally, the direct on-axis response should be flat, with off-axis being smooth replica of the on-axis response and smoothly falling in the very high frequencies.

So, even though you may see in one measurement, a falling response, it doesn't mean that is the correct response. You have to provide the context for the measurement as well. Anechoic, quasi-anechoic, in-room, and so on.

The reason for explaining this is that the measurements on the Soundstage Network website are like golddust. You can tell almost completely how a loudspeaker might perform in a room by looking at them. The measurements are highly reliable and lot can be learned from them.

Here is the LS50:
SoundStageNetwork.com | SoundStage.com | NRC Measurements: KEF LS50 Loudspeakers

Look at the first graph. It shows on-axis, and 15 and 30 degrees off-axis. Off-axis looks reasonably smooth, but there is a 3db rise beginning at 1.5 kHz. Now look at the third curve, the listening window. The rise doesn't go away. This is the best approximation of how it might sound. Clearly, it will far too bright. You want a flattish response here.

Some might say that you should listen off-axis. But the bright nature of the sound will still dominate. Check out the KEF Reference 201/2. The same topology with dual concentric drivers, but look at the response:
SoundStage! Measurements - KEF Reference 201/2 Loudspeakers (12/2007)
Click to expand...

Try as you might there is No way to get any accuracy gated indoors below 500hz , unless you have the speaker in a pretty big, tall room full of fiberglass batons .....
 
#16
You will get a falling in-room response if you measure the total sound incident upon the microphone, including the direct sound from the loudspeaker and reflected sound from all the room surfaces. This is because most surfaces will absorb some HF, and as a result, you get the falling response.
Click to expand...

Yes, I'm aware of that, I should have mentioned that it's for measurements that are in-room.
When I make in-room measurements I always make measurement of some speakers of which I know the anechoic response, this to see the influence of the room and the measurements setup.
 
#17
Looking back at my Eq files in Fabfilter seems I forgot that I did end up liking some Eq on the KS50s. I remember not minding them without, but there is room for improvement.

Filters were 1.3k +3db Q=2.5; 2125hz -2db Q=4.4; 3430hz -1.5db Q=2.2

I think my ears were more sensitive to the 900-1.7k dip than the rise above that.
 
Status
Not open for further replies.
Top Bottom