Room measurement, what to learn?

[murphythecat]

They start a gradual roll off from around 40hz. I've got the REL set to do nothing above 38hz.

Ive tried two rel sub. be careful, my rel t5 and rel t2 both, even at lowest xo setting (30 hz) actually was -6db at 60hz and -3db at about 50hz.



Im looking at your measurements. can you post a measurements without the sub on?
I also realy wonder why your -10db at 100hz and serious dip from 100hz to 400hz and huge bump between 900 and 2 khz.
+10db between 1 to 2khz will likely be very obvious.

can you post measurements from about 1.5 meter away from the speaker and also at the listening position so to know if its the room that colours the response so much or its the speakers themselves...



Ideally, you want to be from 200hz and up, about +-5db in room response. With a lot of treatment you should be able to extend that from 80hz to 20khz to be +-5 db at the listening position.

 

[Mr_Sukebe]

Murphy,
Will do. Hadn't thought of taking near-field measurements, good idea.
I also wondered about the mid-peak. I've been told that it's a trademark of my speakers. For all that, I'd like to understand how much of it is because of the room as I was wondering about putting some GIK 242s on the wall behind my seating position, i.e. pretty much directly opposite the position of my speakers. Thought that it might help clean things up a little.

I'll take some time out later today to grab the measurements. Interesting stuff.

 

[murphythecat]

take three measurements with only one speaker at a time. a measurements at 1 meter, 1.5 meter and at listening position always at about tweeter height.
Id sell a speaker that has a built in bump at 1khz though.

the gik 242 wont remove the FR anomalies, but will help for first reflections. Ideally, you should have a panel at ear level on the back wall, the side walls and the ceiling.

 

[Mr_Sukebe]

take three measurements with only one speaker at a time. a measurements at 1 meter, 1.5 meter and at listening position always at about tweeter height.
Id sell a speaker that has a built in bump at 1khz though.

the gik 242 wont remove the FR anomalies, but will help for first reflections. Ideally, you should have a panel at ear level on the back wall, the side walls and the ceiling.

Thanks for your thoughts on the 242s. I was thinking of a pair of the "picture" versions, such that they're an easier sell for the missus. Guess I'd better go find some suitable pics.

As for selling my speakers, interesting. I did think that one thing I'd learned was that pretty much every component in a system is a set of compromises, speakers included. I did think that these had a fairly good set of compromises.
What do you use?

 

[murphythecat]

Thanks for your thoughts on the 242s. I was thinking of a pair of the "picture" versions, such that they're an easier sell for the missus. Guess I'd better go find some suitable pics.

As for selling my speakers, interesting. I did think that one thing I'd learned was that pretty much every component in a system is a set of compromises, speakers included. I did think that these had a fairly good set of compromises.
What do you use?

oh, I dont even know what speakers you use. I said Id sell a speaker that has a 10 db boost around 1khz but I thought you said that speaker was designed that way. if so, this is not a compromise but serious coloration.

I use amphion one18, harbeth p3esr and tannoy gold 12inch

 

[Mr_Sukebe]

oh, I dont even know what speakers you use. I said Id sell a speaker that has a 10 db boost around 1khz but I thought you said that speaker was designed that way. if so, this is not a compromise but serious coloration.

I use amphion one18, harbeth p3esr and tannoy gold 12inch

Maybe I'm getting too pedantic in my old age, but you didn't mention a 10db peak yesterdays 4:12 post, just "a built in bump".

Agreed however that a +10db boost would be ridiculous, and probably also sound terrible.

 

[murphythecat]

Maybe I'm getting too pedantic in my old age, but you didn't mention a 10db peak yesterdays 4:12 post, just "a built in bump".

Agreed however that a +10db boost would be ridiculous, and probably also sound terrible.

haha no problem
in my post 201, Ive said that your measurements seems to show a 10 db bump between 1 to 2 khz....
i dont its built in the speakers but we never know. Ive measured some very well reputed speaker that measured like a roller coaster.

have you measured the speakers again but closer?

 

[davidsrsb]

Ive said that your measurements seems to show a 10 db bump between 1 to 2 khz....
i dont its built in the speakers but we never know. Ive measured some very well reputed speaker that measured like a roller coaster.

It wouldn't be the first time a crossover wasn't built correctly

 

[ToTo Man]

I finally got round to playing with Dirac Live. It was an '11th hour' task in so much as I had friends and family visiting the next day for a listening session and I just couldn't get my Lockwood Tannoys to play nicely in my room regardless of positioning. Conscious of time, I "cheated" the software at first by only taking a single reading instead of the recommended nine readings. This was the result:

In hindsight, doing only one measurement was a mistake. Although the resulting room correction gave me a measured in-room response which was spectacularly flat at the listening position (the flattest I have ever seen in REW), my ears detected very strange happenings in the lower mid-range akin to a smeary/boxy resonance that just didn't sound right, and the problem was exacerbated as I moved away from the listening position.

Disappointed with the result, I repeated the process but this time took the nine recommended readings, accurately spaced out over an invisible 56inch wide x 18inch deep 'sofa'. Each measurement was spaced 14 inches apart left to right, 12 inches forward and 6 inches back from the 'sweet spot' mic location. I chose not to vary the height of the mic as this would have made the process even more time consuming. The difference in the result this time was very noticeable. While the 'corrected' response did not look as flat this time, all of the previous objectionable peculiarities in the sound of had to my ears disappeared:

I don't have enough knowledge of how Dirac works to explain why taking only one reading did not give a good result, but I'm assuming it has to do with the implementation of its mixed phase filters, for which it requires measurements in varied locations in order to gain enough information about the loudspeaker/room interaction to optimise. Or perhaps my Tannoys were not amenable to such radical correction as was implemented based on the single measurement, I honestly don't know.

Regardless of how it works, work it does! The improvement in the sound has been nothing short of jaw-dropping. It is by far the biggest improvement in my system I have heard, and all visitors were in agreement. The smoothness and openness of the sound from top to bottom is an absolute joy to listen to. The clarity and detail that has been fleshed out in the upper mids has been the biggest revelation. Vocals and instruments have a real presence and no longer sound scooped. I honestly don't remember my system sounding this good, and I've had some pretty decent kit in my room over the years.

Before installing Dirac I spent a full day experimenting with positioning to try and minimise the bass nulls caused by SBIR. It turned out that hard against the wall was the best location, which is an added bonus as I can reclaim some floor space! After implementing Dirac, close your eyes and you'd never know the speakers are hard against the wall, such is depth of soundstage and clarity of image being produced.

I went with the default Target curve in Dirac but then raised the bass curve <100Hz by 1dB as it sounded slightly bass light (probably because I'd become accustomed to the +10dB peaks in the uncorrected response). I may boost the bass a little further just to restore a touch more impact, we'll see. This is what the response looks like with Psychoacoustic smoothing applied:

For fun I switch the Dirac off now and then to compare and the difference is so stark that even my poor hard-of-hearing mum can tell the difference. I honestly can't see myself going back to listening to music without full room correction on the Tannoys, in my case the improvement is just too big to ignore. Trouble is I am now wondering what improvement it will make to my various other loudspeakers. I see a tedious summer of sine-wave sweeping ahead of me!…

 

[Tony L]

Interesting stuff. I assume the digital room correction thingy is stereo and does one speaker at a time independently of the other? If so I guess it would help balance out the notoriously poor pair-matching in vintage Tannoys somewhat!

Any negatives? Whenever I've heard electronic room correction I've not liked the system, but I've no idea if that is the correction aspect or just that I didn't like the sound regardless. I remember sitting in on a high-profile expensive brand dem (I don't like slagging product given I'm the site owner, so I'll not name it) and thinking it was dull as ditchwater, just no groove, flow or enthusiasm to the music. It was just not fun. Sure it sounded flat, but in a bad 2d and boring way, not like good headphones or whatever.

 

[ToTo Man]

Interesting stuff. I assume the digital room correction thingy is stereo and does one speaker at a time independently of the other? If so I guess it would help balance out the notoriously poor pair-matching in vintage Tannoys somewhat!

Yes I believe it does, which is indeed very useful, because Tannoys aside, almost every vintage pair of speakers I've owned has had a left-right imbalance at various frequencies which is not room related (swap the left and right speakers over and the imbalance swaps too).

Any negatives? Whenever I've heard electronic room correction I've not liked the system, but I've no idea if that is the correction aspect or just that I didn't like the sound regardless. I remember sitting in on a high-profile expensive brand dem (I don't like slagging product given I'm the site owner, so I'll not name it) and thinking it was dull as ditchwater, just no groove, flow or enthusiasm to the music. It was just not fun. Sure it sounded flat, but in a bad 2d and boring way, not like good headphones or whatever.

It's very difficult to give an objective opinion on this, as I find the interpretation of qualities such as dynamics etc is very subjective as it means different things to people. e.g. now that the bass peaks have been ameliorated you could argue there is less overall impact, but I would argue the opposite as the listening experience is now no longer overwhelmed with 'room loading' bass. But getting the bass right isn't the be all and end all, at least it wasn't for me with the Lockwoods. Now that the upper mids have been lifted and the toppy highs tamed, everything just sounds much more fluid, open and revealing, but not in a sterile 'studio sounding' way if that makes sense, i.e. there is still enough coloration there to know that you're listening to a big behemoth pair of vintage Tannoys!

You do have to be patient and accurate with the measurements required for the room correction. If you're not satisfied with the result then repeat the process is the advice I have received, as apparently the algorithms aren't infallible. I wash't happy with the result I got when I 'cheated' and took only one measurement, as even when sitting precisely at the sweet spot the result sounded weirdly phasey as if certain frequencies had been boosted and/or cut too severely. Repeating the process with 9 measurements representing the footprint of a sofa produced a less extreme 'corrected' version which sounded much more natural at the sweet spot, and also surprisingly good either side of it. I do know of other Dirac users who like to repeat the same set of measurements numerous times before A/B/C'ing the resulting filters and choosing the one they most prefer, as even tiny shifts in mic position can result in audibly different corrections. Perhaps I got lucky first time with mine!