1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Room Acoustics – Solutions for better Mid and Treble

Discussion in 'reference' started by Tenson, Jun 16, 2008.

  1. Tenson

    Tenson Trade: AudioSmile

    since I wrote a thread about room acoustics and bass I have now done a similar one for the mid and treble range.



    Room Acoustics – Solutions for better Mid and Treble

    Why doesn't the mid and treble sound as good as I would like? There are four possible answers:

    Too many sound reflective surfaces resulting in a long decay time with nowhere for the sound energy to escape to. This decaying energy smears over the next piece of music you want to hear.

    Too many sound absorptive surfaces resulting in a lifeless sound with no ambience.

    There are many absorptive surfaces, but they are not very deep/thick. Things like curtains, carpets, rugs and acoustic foam less than 50mm deep. This will create an unnatural sound because the high frequencies are absorbed and the mid and lows remain untouched. The greater difference between the ranges makes problems even more obvious to the ear.

    The decay time may be well balanced with absorptive and reflective surfaces, but their positioning is not ideal. First reflection points are left uncontrolled.



    There are solutions to the problems of maintaining a natural sounding room while controlling bad acoustics:

    1) Place broadband absorbers at the first reflection points* on the side walls, behind the listener and between the speakers. Or whichever of these you can practically get away with. Broadband absorbers are ones with sufficient depth to absorb throughout the high to lower mid range. Realistically this means acoustic foam of no less than 50mm thick, ideally 100mm. Heavy lined curtains partly bunched up and some distance from the wall can help. CD racks can help deflect the reflection as can book cases but they do little to improve the decay rate of energy in the room. Area of coverage depends on the size of the room. In an average room an area about 1m to 1.5m square on each reflection point should work well.

    The aim of doing this this is two fold. It creates a temporal gap between the arrival of sound from the speakers and reflected sound from the room. Your mind can then tell the difference more easily. It also helps to lower the decay time for sound energy in the room thus reducing its smearing effect on the next piece of sound you want to hear.

    *First reflection points are places on a room boundary where sound from a speaker only needs to reflect once to reach the listener. They exist on each and every wall, floor and ceiling.
    [​IMG]


    2) If the sound still lacks definition, and this is improved by moving nearer the speakers, you are essentially asking for a shorter decay time in the room.

    A shorter decay time can be bought about by adding more broadband absorption, or by adding diffusion.

    The addition of pours bass traps across the room corners can provide very broadband absorption that will also improve the bass. More absorptive panels can also be added either side of the listener and anywhere else in the room, spaced evenly. The downside of this is, as mentioned earlier, too much absorption (even if broadband) creates a lifeless and possibly psychologically uncomfortable atmosphere.

    Diffusion can be a substitute for additional absorption. Diffusors maintain the life of the room by reflecting the sound, but they do it in a very special way that spreads it out evenly. The 'diffuse' sound strikes many more surfaces around the room than a single reflection and so its energy decays at a faster rate. The 'spreading out' of the reflection in to many, many separate small reflections also produces a more even energy field and decay in the room, which pure absorption can not achieve.

    The result of a diffuse sound field is a gain in definition similar to sitting closer to the speakers because of the shorter decay. Yet the life of the room is maintained because the sound energy is still in the room, only now more evenly distributed.

    Do not make the mistake of thinking digital room correction (equalisation) can help in the mid and treble. It can not. It can't change the decay rate and it can not reduce reflections. It can only adjust the tonal balance.

    Do not use diffusors in a position close to the listener, place them at least 1.5m away. They need space to work their magic.


    As an aside, if you feel the sound is lacking in space and 'air' at the top end it could be down to the dispersion of the speakers, not just the room. A single driver full range speaker is never going to have much air and space around the treble range because it has very narrow dispersion of sound. This means that there is little high frequency energy in the room and the effect is the same as having too much high frequency absorption. Different tweeters, waveguides and horns all have differing dispersion. Wide dispersion provides more ambience and life in the sound.
     
  2. Tenson

    Tenson Trade: AudioSmile

    I recently did a hi-fi show and took a number of diffusors and acoustic panels with me. On the forums over the months have been a number of threads about acoustics, and since I had to put all my treatments back in my room I thought I would document the effects as I built it up. Hopefully this can help you if you are thinking of using some treatments in your own room.

    Firstly here is the bare room, with the hi-fi, a sofa a few other bits, oh and one quadratic diffusor hanging on the wall above the seat - I couldn't be bothered to take it down just for these test. You can also see the measurement microphone hanging from a hook in the ceiling, this way the mic stand doesn't introduce its own reflections.

    [​IMG]
    [​IMG]

    Here (below) is the Decay Time of the room as in the above pictures. You can see it hovers around the 350ms mark from 100Hz-10KHz with a rise ing to 400ms around 2KHz. For a room of this size (10ft x 12ft) this is quite a high reverb time for critical listening. Although, as living rooms go it is not that high due to the big thick carpet with heavy underlay.

    [​IMG]

    Here (below) is the frequency response at the listening position. The top end rolls off quite fast because the mic is not on axis and its a far field measurement. Quite clear is the room mode at about 35Hz, but this is not a worry to me at this time since I am concentrating on the mid and treble acoustics. In this room I will simply use a digital equaliser to correct for low frequency resonances (the peak also adds a nice kick and makes you realise the Kensai goes damn deep for such a tiny speaker!). What is noticeable is a general rise of about 3dB all the way from ~300Hz to 3KHz comparative to the upper bass and treble. This can make the sound a bit fatiguing and mid strong.

    [​IMG]

    When adding reflection absorbing panels, the most useful measurement has to be the Energy Time Curve (ETC) this shows a simple plot of sound energy at the microphone vs. time. This plot allows you to see the direct sound from the speaker (first peak) and the individual reflections arriving afterwards and their relative level to the direct sound. It looks a mess I know, but as we start to add panels at the first reflection points and see what is happening, you will get the idea.

    [​IMG]

    Okay so now I add an absorptive panel to the back wall behind the listening seat. You get strong reflections here if you are close to the wall and they come very soon after the direct sound which is a bad thing.

    [​IMG]

    This is where the ETC plot is really good because we can see what reflections are reduced by placing this panel here. Below is an overlay of the ETC curve before the panel was added (red), with the new ETC after the panel is positioned (green). Looking carefully you will see a significant reduction of almost 20dB of a reflection arriving just after the 13ms mark on the graph. This was almost as strong as the direct sound! If that wouldn't mess up your sound localisation abilities (imaging) I don't know what would. Another is reduced (probably reflected from the side walls to the back and to the listening position) just before the 19ms marker. Clearly other reflections are reduced too because they would have reflected from that portion of wall somewhere along their path of travel.

    [​IMG]

    Following are the respective RT60 (decay rate) and Frequency Response plots. It can be seen that the decay of sound has been sped up over the range of ~200Hz-1KHz. Strangely there is a strong peak in the reverb time around 2KHz which I have not usually come across when treating rooms. I work out what is causing it later on.

    [​IMG]

    The Frequency response is also improved. With those big reflections gone the midrange is now much smoother from 200Hz up. However, there should be a tiny rise around 2KHz because that's what the speaker does.

    [​IMG]

    After this encouraging result we want to do more! The next step I took is to treat the first reflection point on the right hand wall with an absorptive panel.

    [​IMG]

    Here is another comparison of the ETC curves. Red is the last ETC with the back panel, and green is with the right hand panel added. There is not such a clear change this time, but if you look closely you will see a significant 12dB reduction of a reflection very early on. This is the first reflection arriving from the right hand wall, and although it looks insignificant on the plot, its early arrival after the direct sound and high relative level mean it has a strong effect on image localisation and stability. Accepted wisdom is to reduce the early reflections as much as possible within a 10-20ms period after the direct sound and then maintain a diffuse string of reflections that add spaciousness and life to the environment. So reductions in this early period are very good.

    [​IMG]

    Here (below) is the decay time after that panel was added. A further reduction in decay time is evident from 200Hz and up, which is good as small rooms tend to sound better with short decay times around 250-300ms. However the peak at 2KHz has become even better defined, despite adding more absorption, and this has me thinking there is a specific area of the room that is ringing away at this frequency. If tamed, the decay rate should be pretty smooth right across the range. I didn't track it down yet though, and continued to add the last reflection panel for this room.

    [​IMG]

    The mandatory Frequency Response plot below. The lowering of the decay rate in the 200Hz-1KHz range has lowered the frequency response in this range too, and made it more in line with the upper bass level, which good, but it has also produced a bit of unevenness around the 2KHz point. Not to worry though as I suspect when I nail that 2KHz reverb spike the response will fall in line too.

    [​IMG]

    The final reflection panel I add is on the first reflection point of the opposite (left) side wall.

    [​IMG]

    The result is below. Another reflection has been clearly reduced

    [​IMG]

    The decay rate (below) has been reduced again around 300Hz, but not much elsewhere which was really bugging me when I was doing this!

    [​IMG]

    The Freqency Response is improved though, with the peak around 2KHz coming down by about 1.5dB. This has the overall response more in line with what the speaker actually produces. i.e. much more balanced relative to other frequencies with sinificantly less midrange energy.

    [​IMG]

    Having got the first reflection points sorted out I now want to ensure that the remaining later arriving reflections are as diffuse as they can be. This means there should be as many reflections as possible, all arriving shortly one after the other and exponentially falling in level. Note this is late reflections we are talking about now, not early ones. Diffusors help to do this by 'splitting up' any reflections that fall on them in to many smaller reflections and spitting them out in all differnt directions, which means they all take a slightly differnt path around the room before arriving one after the other at the listening position.

    [​IMG]

    It is actaully quite hard to show this in measurements. The best I can do is show a smoothed ETC curve over a longer period than the others here. Where the reflections are more diffuse, arriving close together and smoothly decreasing in level, the plot should look more like a smooth line. Conversely, where the reflections are less diffuse you will see a more jagged, spiky line. Below is a comparison of the room before the diffusors were added (orange), and after (blue). I think it can be seen that the 'after' plot is decreasing in level more evenly, while the other initially drops, then goes along a bit then drops again. It can also be seen I think, that the 'after' plot has less jagged sections.

    [​IMG]

    Finally, I also found where that damn 2KHz peak in the decay plot was coming from! There are two parallel walls either side of the seating position where the sound was evidently bouncing back and forth at this specific frequency. I had assumed that because the speakers where firing at these walls from an angle it would not be an issue, but I guess reflections still arrived here at the required angle and started to build up. Having put another diffusor on the wall to the side of the seat (in the gap to the right of the panel in the photo below, beside the plant) the decay plot suddenly smoothed out as it should have been in the first place with all those reflections banished away and the rear diffusors in place. Its a shame I didn't figure that one out earlier as it would have made the other decay plots more revealing of the differences the panels made but never mind. You can see how comapred to the very first plots it is now a lot better in all regards.
    [​IMG]

    [​IMG]



    For comparisons sake, here are the first plots for the bare room, with the equivalent plots after the room was treated.

    [​IMG]
    [​IMG]

    [​IMG]

    Anyway if you got this far I hope you found it interesting!

    As always there are a few anomalies in the measurements. For one I notice a very early reflection actually got stronger around the time I added the second side wall absorber, however it can't be due to this as the reflection arrives way too soon for that panel to affect it. I guess I must have moved something small close to the speaker without thinking about it when moving the panels around. Also ignore decay rate data below about 200Hz and above about 5KHz as it is not accurate for a small room like this and when there is noise outside from traffic etc..
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice