Measuring the effect of acoustic treatments in the room

[AndyU]

So true: Measurements give you data. Messing about gives you a solution. The solution (one of many) can be scientifically non-intuitive. I love my first reflections off the side walls.

Also what you might think of as “science” when it comes to hifi - frequency response for example - isn’t necessarily so. If you recast the problem as wanting to maximise intelligibility - say whether you can discriminate, or even hear, the woodwind section in an orchestra, or what the cello is playing in a string quartet - then there may be a complex relationship between reverberation time and frequency response. Nobody I know of measures this for music, but there is a substantial body of research on speech intelligibility. This paper for example suggests that your preference for early reflections might have a measurable basis in fact when it comes to speech:

https://www.researchgate.net/public...ty_of_speech_in_classrooms_for_young_children

 

[tuga]

I only moved the mic for these measurements, not the speakers, and the measurements were made using Periodic Pink Noise with RTA so there is no time domain information. I can, however, repeat some measurements using impulse sweeps for the distances that are of interest.

I'll try listening at 180cm, 225cm, 240cm, 255cm, 270cm and 285cm to see what they sound like, however I think 285cm is going to push me too close to the speakers! Can I ask why 285cm is of particular interest to you?

I think that it would be interesting to compare the response in the time domain as well as in the frequency domain in two different places.

I understan that this is a very time-consuming task but I would also choose a couple of listening spot locations and move the speakers back and forth a bit.
In my previous post I meant to say the speakers and not the mic:

It might be interesting to pick a couple of mic-to-RW distances and try moving the speakers 10cm forward and 10cm backward.

 

[ToTo Man]

I spent last night listening at the "180RW" distance and it definitely sounded better than "160RW". Increasing the ratio of direct:reflected energy by moving closer to an equilateral triangle improved phantom centre imaging and soundstage depth. Lateral imaging still lacked the ultra-preciseness you get when the side walls are treated but it was still intelligible.

After dialling in my parametric EQ I'd say this is the best I've heard these speakers from a tonal POV. I perhaps reduced the 53Hz, 86Hz, 108Hz and 208Hz peaks a bit too much as a lot of familiar tracks sounded a little light on bass, but the presentation was wonderfully open and had a great sense of ambience. I now realise I was probably over-deadening my room with the amount of absorption I had in it, although many on Gearspace will argue otherwise!

However, after about 20 minutes or so listening fatigue set in. As there's nothing harsh about the way my system sounds, the fatigue is obviously a result of my brain having to process so many reflections, a problem I haven't had to deal with for many years! It would be great if I could somehow come up with a room treatment plan that reduces reflections and decay times to a manageable level and improves my low end response below 200Hz WITHOUT adversely affecting the frequency response above 200Hz. Surely this isn't too much to ask?!

Going forward, before I start experimenting with pulling the speakers away from the wall, I'm thinking it would be interesting to take time domain measurements 180cm from the rear wall and 180cm from the front wall. This will show how the two halves of the room behave w.r.t. reverb.

 

[tuga]

I spent last night listening at the "180RW" distance and it definitely sounded better than "160RW". Increasing the ratio of direct:reflected energy by moving closer to an equilateral triangle improved phantom centre imaging and soundstage depth. Lateral imaging still lacked the ultra-preciseness you get when the side walls are treated but it was still intelligible.

After dialling in my parametric EQ I'd say this is the best I've heard these speakers from a tonal POV. I perhaps reduced the 53Hz, 86Hz, 108Hz and 208Hz peaks a bit too much as a lot of familiar tracks sounded a little light on bass, but the presentation was wonderfully open and had a great sense of ambience. I now realise I was probably over-deadening my room with the amount of absorption I had in it, although many on Gearspace will argue otherwise!

However, after about 20 minutes or so listening fatigue set in. As there's nothing harsh about the way my system sounds, the fatigue is obviously a result of my brain having to process so many reflections, a problem I haven't had to deal with for many years! It would be great if I could somehow come up with a room treatment plan that reduces reflections and decay times to a manageable level and improves my low end response below 200Hz WITHOUT adversely affecting the frequency response above 200Hz. Surely this isn't too much to ask?!

Going forward, before I start experimenting with pulling the speakers away from the wall, I'm thinking it would be interesting to take time domain measurements 180cm from the rear wall and 180cm from the front wall. This will show how the two halves of the room behave w.r.t. reverb.

Looking really good. I'd love to see the decay plot of one of your speakers. Please?

 

[tuga]

Also what you might think of as “science” when it comes to hifi - frequency response for example - isn’t necessarily so. If you recast the problem as wanting to maximise intelligibility - say whether you can discriminate, or even hear, the woodwind section in an orchestra, or what the cello is playing in a string quartet - then there may be a complex relationship between reverberation time and frequency response. Nobody I know of measures this for music, but there is a substantial body of research on speech intelligibility. This paper for example suggests that your preference for early reflections might have a measurable basis in fact when it comes to speech:

https://www.researchgate.net/public...ty_of_speech_in_classrooms_for_young_children

Toole has made this same mistake of confusing the acoustic qualities of a room for producing sound/music with those of a room to reproduce recroded music.

Recordings have embedded ambience cues and the reproduction has different requirements.

 

[ToTo Man]

Looking really good. I'd love to see the decay plot of one of your speakers. Please?

I can't do a decay plot with the EQ on. Well, I think it is possible, but I don't know how to apply the EQ settings I programmed into Audirvana's AU plugins into REW's EQ!

 

[tuga]

I can't do a decay plot with the EQ on. Well, I think it is possible, but I don't know how to apply the EQ settings I programmed into Audirvana's AU plugins into REW's EQ!

I don't know how to do that either, sorry. But maybe one without EQ would be fine.

 

[ToTo Man]

I just posted this update on Gearspace... should set the cat amongst the pigeons!

 

[ToTo Man]

@tuga, The low frequency reverb times appear to be better in the front half of the room, but the difference is smaller for the right speaker:

 

[tuga]

@tuga, The low frequency reverb times appear to be better in the front half of the room, but the difference is smaller for the right speaker:

I agree, they look quite similar. Looking forward to seing how the traps will affect the decay .

If you are not yet completely fed up with the process , it might be worth choosing a speaker and listening spot location then measure one speaker at 0cm, 10cm back and 10cm forward with the mic also at 0cm, 10cm back and 10cm forward (9 measurements in total using sweep).

 

[ToTo Man]

I tried the rule of thirds this afternoon, speaker baffles 138cm from front wall, listening position 138cm from rear wall.

It was terrible at first, you could clearly hear the sound coming from the speaker baffles and it was very coloured… I forgot to engage the EQ! This is what makes the process so damn tedious, every time I tweak my speaker or listening position I have to re-adjust the EQ!!!

After EQ the speakers disappeared better but I wasn’t convinced the soundstaging was any deeper than when I had the speakers against the front wall. I waited until it was dark and listened again with the lights out and it was much more holographic. Still not sure if significantly better than hard against the wall, I’ll need to move them back again and compare. I also need to try the two other arrangements in the link you provided. My dad’s gonna be pissed!…

This is how it measures with the rule of thirds. If I adopted this position permanently I’d definitely need to get my XXLS400 subs back into my room to fill in that wide crater between 65Hz-80Hz.

BTW - that EQ setting was my first hurried attempt to get rid of the colouration, I’ve since raised the level of the bass up by +2dB so it's no longer as lean as the graph suggests.

 

[ToTo Man]

I reintroduced the 8 GIK TriTraps, but this time I placed them at the floor-wall corners instead of the wall-wall corners. As you can see, they still eat the speakers' lower midrange in this position.

 

[Tony L]

That is a remarkably strange location for a pair of Ditton 66s! They are designed to be parked against the wall aren’t they? Not free-space. I’ d have thought doing so would bring up the 100-400Hz range and naturally shift them more towards a gently falling ‘target’ response. It makes sense that you found them rather forward and fatiguing.

PS I’m sure I only get away with my Tannoys out from the wall as I have to sit so close to the back wall so get reinforcement there (my room is pretty small). If I lean forward or sit on the floor in front of the sofa a lot of the bass weight and heft goes away. In a larger room they’d certainly need to be closer to the wall.

 

[ToTo Man]

That is a remarkably strange location for a pair of Ditton 66s! They are designed to be parked against the wall aren’t they? Not free-space. I’ d have thought doing so would bring up the 100-400Hz range and naturally shift them more towards a gently falling ‘target’ response. It makes sense that you found them rather forward and fatiguing.

I can't find any literature that states the Ditton 66 is designed for wall placement but I guess most designs from this era were designed to be put against the wall as having speakers in free space wasn't domestically acceptable?! There isn't any baffle-step correction on the crossover that I'm aware of. The anechoic response of these speakers is rather peculiar as maximum output seems to be between 400Hz-1kHz, this is very noticeable when the speakers are out from the wall.

FWIW - the least amount of EQ cut needed to get the midrange sounding transparent is when the speakers are hard against the front wall and close to the side wall. The physics of this is beyond me but my best guess is that the early reflections arriving from both walls helps to cancel/smooth out the mid's output. But how did the designers know this is what would happen?

I'd love to know how far away from the side walls these speakers were designed to be placed and also the intended listening axis. For me the magic doesn't start happening until you're on-axis with the mid driver and the speakers are toed-in a bit.

My reason for wanting to get them away from the walls is to: a) reduce the amount of room treatment I need; and b) experience a fraction of the "20-ft deep soundstage" that certain audiophiles keep harping on about getting from their systems! I suspect I neither have the right speakers or a big enough listening space to achieve this without serious compromises elsewhere...

 

[ToTo Man]

@Tony L , the excerpt of the HiFi News review featured on the last page of the Ditton 66 brochure mentions listening to the speakers in a 110ft x 55ft x 25ft hall. I can't imagine they'd have been placed hard against the front wall in this size of venue?

 

[ToTo Man]

I’ d have thought doing so would bring up the 100-400Hz range and naturally shift them more towards a gently falling ‘target’ response.

You mean pulling them out from the wall should increase 100Hz-400Hz, or moving them towards the wall should increase 100Hz-400Hz?

 

[Tony L]

You mean pulling them out from the wall should increase 100Hz-400Hz, or moving them towards the wall should increase 100Hz-400Hz?

I’d guess placing them against the wall whould lift and flatten that 100-400Hz area, maybe a bit higher too. I am sure the 66, like all Dittons, will be a speaker designed for half-space (i.e. backed against the wall). You have to look very long and hard to find any speaker of their era which weren’t other than say ESLs. I’d expect all the Celestions, Rogers, Spendors etc to have been used within a foot of the wall maximum, i.e. maybe pulled just a little forward if there was a bookshelf or sideboard between them. Have you had a good search for brochures, period reviews etc. Might be some in-room pics. Pulling speakers right out into the room just wasn’t a thing back then.

 

[rag987]

I have not been able to follow all you have done and said, and admire the time and energy you are putting into this. This caught my eye:

It was terrible at first, you could clearly hear the sound coming from the speaker baffles and it was very coloured… I forgot to engage the EQ! This is what makes the process so damn tedious, every time I tweak my speaker or listening position I have to re-adjust the EQ

Are you not meant to first find the optimum spot through placement of speakers, then add appropriate treatments, only only when these give you the best then you add EQ / DSP?

 

[tuga]

I reintroduced the 8 GIK TriTraps, but this time I placed them at the floor-wall corners instead of the wall-wall corners. As you can see, they still eat the speakers' lower midrange in this position.

Wow, that's brutal.

 

[ToTo Man]

Wow, that's brutal.

If you look at the Ditton's published anechoic response there is a broad dip between 100Hz-300Hz. When I first saw this graph I thought it was only a -2dB dip but I now realise it's -4dB (each interval on the scale is 2dB, not 1dB). HOWEVER, many other speakers I've measured when the TriTraps have been in this room also show a dip in the lower mids / upper bass, including my Tannoy Edinburgh and the Falcon Q7, so the treatment is definitely making this area leaner. But it's probably more obvious with the Dittons because they're already recessed in this area to begin with.