help please understanding results / improvement tips

[seangoodman]

Hi guys, am new to REW and have been spending the last few weeks measuring my room and making a few acoustic improvements, I have got as far as I can and am beginning to feel that I am now going round in circles and could do with a bit of help to understand what is going on and what I could realistically do to improve things if indeed it is deemed that improvements should be made.

my system is as follows:
Musical Fidelity Ms5i amp
Dali Opticon 8 Floorstanders
Input via Technics 1210 with Denon DL110 and digital via Cambridge can streamer +Qobuz

room treatment so far 6 GIK 242 panels at first reflection points on side walls and ceiling plus 2 home made triangle bass traps 40cm x 40cm x 120cm high in both front corners, filled with Isover TP1 insulation

the problems that need addressing (remember I am novice so this may be incorrect) is as follows:

1) left speaker bit of a null at 40hz, quite a wide dip from about 125hz up til about 700hz, what could be causing this and what could I realistically do about it?
2) right speaker big null at 68hz and again the dip between 125hz to 700hz, what could be causing this and again what could I realistically do about it?
3) I think my impulse measurements actually look quite could but I would like somebody more qualified opinion as to if they actually are acceptable or not

regarding the deep null at 68hz on the right speaker, I did have this on the left side as well but it went away when I moved the armchair back 20cm, doing this on the right side had sadly not effect.

regarding the dips in the middle, moving speakers a bit left / right, toe in or straight doesn't seem to have any effect nor does moving listening sofa back or forwards

perhaps somebody could have a look over my Mdat file and give me a heads up on how its looking and what I could realisably do to get anymore out of my system if at all possible

many thanks
Sean

 

[sm52]

The impulse is abnormal. It should be at first = 0, then it will sharply increase to the maximum amplitude, then gradually fade out. Your momentum is slowly increasing, so it's not clear where the starting point is. Why so, I do not know.

 

[John Mulcahy]

The measurements were made with a very long sweep, 4M, so any clock rate difference between the replay device clock and the USB clock from the port the UMIK is connected to will cause spreading of the impulse response. There is no need to use such long sweeps, 256k is sufficient for room acoustic measurements and will save a lot of time. Make sure the default format for the output device has a 48 kHz sample rate in Audio Midi Setup.

 

[seangoodman]

The measurements were made with a very long sweep, 4M, so any clock rate difference between the replay device clock and the USB clock from the port the UMIK is connected to will cause spreading of the impulse response. There is no need to use such long sweeps, 256k is sufficient for room acoustic measurements and will save a lot of time. Make sure the default format for the output device has a 48 kHz sample rate in Audio Midi Setup.

thank you for your reply, very much appreciated. I will change the settings and do another set of measurements to see what they look like and post the new Mdat here.

do you think the dip in the 200 - 500 or so range could be due to SBIR off my front wall or the wall beside the speakers? the speakers are 83.5 cm from the wall when measured from front of baffle, I have 43cm gap from back of speaker to wall (2 rear ports one for each bass driver) Dali recommend min of 30 cm from wall so I could go back a bit say 13cm or so.

83.5cm as 1/4 would be 102hz which admittedly doesn't tie in with my right speaker nulls but 2 multiple up would, from the side wall I have 72cm which works out 119hz

am I being too hopeful that a couple of say GIK 244 panels one behind the speaker and one beside it would raise this big dip up for me?

many thanks Sean

 

[seangoodman]

The measurements were made with a very long sweep, 4M, so any clock rate difference between the replay device clock and the USB clock from the port the UMIK is connected to will cause spreading of the impulse response. There is no need to use such long sweeps, 256k is sufficient for room acoustic measurements and will save a lot of time. Make sure the default format for the output device has a 48 kHz sample rate in Audio Midi Setup.

I redid the measurements as you suggested, just on the left speaker at the moment as I didn't get time, something strange happened when I did a sweep at 256k, the top end frequencies just died on the graph, fell right off at at 2.5Kz and didn't show responses up to 20Khz.

I checked the settings at they were for a range of 20hz to 20Khz, I redid a sweep at 4m and that went right up to 20Khz as before. another thing that's very strange is that all of a sudden when I try and adjust the ETC smoothing (ms) to say 0.2 nothing happens, the graph doesn't change anymore.........no idea what I've done.....grrrrr

anyway here are the two sweeps I just did at 256 and then again at 4m perhaps you could have a quick look please and tell me what Im doing wrong

 

[John Mulcahy]

Truncated high frequencies are usually caused by excessive playback latency in the audio chain which means the end of the sweep is not captured. To fix that you can reduce the buffer size settings on the REW soundcard preferences and use the Acoustic timing reference option, which ensures REW will wait for the timing reference before it starts its capture of the sweep.

 

[sm52]

John, how to properly prepare the measurement in order to use the Acoustic timing reference? Can I use a second(left reference, right measured) loudspeaker in which all speakers together can play the full range? Or is tweeter enough? Should the reference speaker have no crossover? Or is the main thing - the same distance to the microphone of the reference and the measured speaker?

 

[John Mulcahy]

The speaker playing back the reference signal needs to be capable of reproducing a high frequency signal. It doesn't matter where it is, but to meaningfully compare separate measurements the mic should remain the same distance from the reference speaker for each measurement.

 

[sam_adams]

Truncated high frequencies are usually caused by excessive playback latency in the audio chain which means the end of the sweep is not captured. To fix that you can reduce the buffer size settings on the REW soundcard preferences...

@John Mulcahy could those i/o buffer settings values be saved in the .mdat and displayed int the info dialog for each measurement?

 

[sm52]

Thanks.

 

[larsvendel74]

well to give you some useful to try i must ask if you have found or even tried to find the sweetspot in the lower frequence. Its a big must and one thing one cannot shortcut, or it can be done, but its much much more effective to start in that end. wen you know where to put the big x on the floor we can start to look at the response, i would prefer without any acoustic treatment at all actually to get a bigger picture from scratch. It takes either a lot of cash to treat a livingroom, or a lot of hours if you are the diy type as myself. Just to make a crazy big leap out here in this forum on my first time logging in, ive just started to get my studio gear together, and as late as a few days ago ive fired up my workplace for the first time. Ive probably done calculations for at least six months, what speakers, how much can i greaze my wife before driving her completely mad building a mastering studio in our livingroom, and how do i make it work as hifi as possible when i want to relax and enjoy music rather than work with it. So i made a little experiment, i got a couple of gaming speakers from my son in law, small plastic strange looking kind of hairdriers speakers, Razer, cost about 150-200 euro so its not crap but its definitley not something that i would even look at a normal day, but why not, if the science and numbers matching, who knows,,, we pretty much doesent listen to our speakers at all in a standard livingroom, we listen to the room it self and all other stuff that throws soundpreassure back at us, I really had to make this little hifi experiment of mine at least 150% more expencive so i put a 300 euro subwoofer at the floor to give my hair driers a little bottom, the drivers are about a couple of inch wide, only one per speaker as well. ii sat down adjusted the subwoofer level to the speakers run it through my interface and pc, sat down and its beyond any expentations, even if i compare them to lets say Adam A8x studio monitors for about 2000 euro for a pair, this stuff i have here at home in the sweetspot of my room sweeps the floor with pretty much anything up to 2500 euros speakers placed in the wrong environment/place/room. Well this was a really humble entre in this forum, LOL, but what the heck, this is my straight and honest opinion how to start to get an idea where to begin, So with all this massive text written, if you dont know where the sweetspot is, tear it all down and start from scratch i`ll help you with as much as i can, if you dont think im a complete idiot tehn. Cheers/lLasse

 

[seangoodman]

well to give you some useful to try i must ask if you have found or even tried to find the sweetspot in the lower frequence. Its a big must and one thing one cannot shortcut, or it can be done, but its much much more effective to start in that end. wen you know where to put the big x on the floor we can start to look at the response, i would prefer without any acoustic treatment at all actually to get a bigger picture from scratch. It takes either a lot of cash to treat a livingroom, or a lot of hours if you are the diy type as myself. Just to make a crazy big leap out here in this forum on my first time logging in, ive just started to get my studio gear together, and as late as a few days ago ive fired up my workplace for the first time. Ive probably done calculations for at least six months, what speakers, how much can i greaze my wife before driving her completely mad building a mastering studio in our livingroom, and how do i make it work as hifi as possible when i want to relax and enjoy music rather than work with it. So i made a little experiment, i got a couple of gaming speakers from my son in law, small plastic strange looking kind of hairdriers speakers, Razer, cost about 150-200 euro so its not crap but its definitley not something that i would even look at a normal day, but why not, if the science and numbers matching, who knows,,, we pretty much doesent listen to our speakers at all in a standard livingroom, we listen to the room it self and all other stuff that throws soundpreassure back at us, I really had to make this little hifi experiment of mine at least 150% more expencive so i put a 300 euro subwoofer at the floor to give my hair driers a little bottom, the drivers are about a couple of inch wide, only one per speaker as well. ii sat down adjusted the subwoofer level to the speakers run it through my interface and pc, sat down and its beyond any expentations, even if i compare them to lets say Adam A8x studio monitors for about 2000 euro for a pair, this stuff i have here at home in the sweetspot of my room sweeps the floor with pretty much anything up to 2500 euros speakers placed in the wrong environment/place/room. Well this was a really humble entre in this forum, LOL, but what the heck, this is my straight and honest opinion how to start to get an idea where to begin, So with all this massive text written, if you dont know where the sweetspot is, tear it all down and start from scratch i`ll help you with as much as i can, if you dont think im a complete idiot tehn. Cheers/lLasse

Thanks for your reply, yes i have spent alot of time working on speaker positioning, furniture positioning upmto the point of moving a speaker 10 cm left right forwarpds or backwards. Same with listening position and other furniture in the room. Each time taking measurements and doing some re listening to see what my ears were telling me.

once i was happy that i had got the most out of positioning etc i moved onto 1st reflection point treatment which gave a big jump up in test results and listening experience.

haveing got this far i now want to see if i can get a bit more of an improvement out of things in particular raising the null points that can be seen in my measurements around the mid bass say 200 to 500 ish. I cant work out what is causing these nulls and what i can do to improve them. This is where i need somebody to. Look at my mdats and room pictures and give me. Few ideas

 

[seangoodman]

I have also taken johns advice about doing sweeps at 256 and done the setting changes to use audio accoustic timing, can post new mdats at some point but not at the moment as i am on ipad now. However dip in the 200-500 is still there to be seen

 

[larsvendel74]

ok i`m with you, well its very hard to just raise dips,for an example, lets say you use an eq, pushing up the dips,( one of the first things my students tried to do when they got similar tasks as newbees ), its not a good idea to add more energy in to the room using this kind of approach, just so ve can verify its not any eq settings on some of your equipment like bass treble adjustment etc, im going to take a closer look and see what i can come up with. If you could send me a better drawing on your room, with height floor to ceiling with exakt measures, im lousy at that strange way you guys measure distance over there with this said i want the distance in mm, thats how we measure length in Sweden anyway or cm, standard is mm in construction, and is it a drop/angle in the ceiling i see in the picture? if you can go through this it would help me alot, and you btw I will do some math to see if i can figure out where the problem begin so to say, meanwhile i will check your response measure. Another question, just for fun, ever heard of faber acoustic, they have a crazy expensive app, if you go for the whole thing, and thats nothing i recommend if you dont work as an engineer, well if you have a big wallet and are a super neard its a very very very good tool!! I cant live without it, if i dont remember wrong ot is free to try for a week but signalscope x will work for this use the octave analyzer in the listening spot you have, use pinknoise in this example, use the db meter in faber set it to slow and measure in dB C, thats the way our ear works in low sound pressure, Adjust the volume to about 87 dB, then use the octave analyzer see how it looks any dips etc, then you can work your way towards each speaker individually use a chair or as i do i steal the ironing board from my wife and place the ipad on it, check at what distance any problems starts to show and is the result same on both speakers for example, but nr one, forgot it, se how the curve look like about three feet or so away from each speaker if the sound really is evolved go another feet away. this will tell you the color of your speaker and to get a better grip where your problems build up in your room we have to know what our speaker do to start with, this is a little backwards and I know it may feel totally odd and wrong, you have bought your speakers for what they are how they sound, now you want to have the most or at least improve the environment, so you can hear them better,,,, so check the curve, it has an fft as well but i want you to go with the octave, see how it is close up then go three feet, four, five,, when it start to change so you can se it and maybe hear it as well lets say in the range of 3 to 6 dB difference, stop then go back so it is as " similar" or how should i put it average from closest to far distance as small unchanged values dont look blindly on the hair cross its the smoothness of the curve we want to look at and its high and dips ey will be there trust me, and you will probably see a higher curve at low end up to about 500 600 hx before it drop down and flatten out, . After this you have to download a music player with,,,, eq,,and im not drunk and i dont do drugs,,but you need this,to your iphone or pc the source that you play pinknoise from, i use my mac or iphone, easier if you have a pc but a phone will do, , i dont remember right now what the name is on the one that i use but it is a serious music player, you want a graphic eq with at least 30 bands, the one i use have floating points and its q value can be done really narrow. Now with this very mighty tool you will now try to flatten out the curve showing in the octave analyzer to as flat as possible do the work in front of one of the speakers go and check how it looks at the other one, if it is way out off the ballpark on the other one make a new eq for that one, and save them so we can see the difference. next step, do an eq in the listen position, curve as flat you can, from about 8 kHz to 20Khz the curve should drop with at least 8 dB. And here is the thing, dont try to add something in the first flatten out procedure, push down the high stuff first, start in the low finish in the high freq, you will see how it affect the analyzer as you work, and with a little luck you might begin to see a pattern, when you take away energy in a certain frequency another will arise and vice versa. if necessary you obviously need to go in there and push some levels up, this is a fact, but start with the big picture, make it look ok then you go in and start to tweak a dB here another there on a specific Hz. But it is always a very bad idea to add low frequency trying to smoothening out the curve as we see it in your picture adding more energy will just amplify the problem. Im writing so much i cant help it, i miss my work at school i guess lol. and my grammar is rusty, so you have to bare with me. What have we achieved with all this mumbo jumbo then, well more or less we have now made your warm sounding Dali speaker,( love the sound in helicon btw, and the looks, cream for the soul as we say over here,) sounding as good as it would in an controlled area in a flat response, this is flummy i know, and it is to push it pretty far saying they sound controlled as in ,,, but now we know what the speaker does in your listening area. and now you have to play the impulse signal with your calibrated curve, I know by fact that there will be some hair pulling and upset thoughts about this, specially by now when they dont stand to read more, but imagine this curve of yours dont exist, its built into your speaker custom made for you only. See how the response look like now, use this curve for some days listen to music, turn it off and on, this is for you to get a feeling of how the whole machinery work in a strange way, i want it to be educational and that you get some steps forward in finding what you actually is missing and have done wrong, because you have we all have, so im not saying this in a bad way. But this can be one way to help you in your quest searching for the holy spot where the goosebumps starts to appear on places we never know it could exist I dont know if you should see this massive amounts with text as a good thing for you or bad, im at home, sick leave and normally i wouldn´t have the time at all, so i hope you enjoy, if it is to much and you just want an easier answer just say the word, i wont be offended in any way, but whats the fun with that, i´ll give you one lead how to think, in what ways can we, if we wanted, increase or decrease a certain frequency make a standing wave for example. and what is a node, where are they, why do it change even if the numbers seems ok on paper, how does two magnets impact on each other is one very easy way to feel the force, wavelenght and octaves how do they work. that was more than one lead but think outside the box, dont look on your 500 Hz dip and mount absorbers here and there, you follow me on that last one, sound absorbers, just to say it twice, and is your basstraps in the corner really a basstrap, or is it an absorber that is placed in the corner, how does an absorber work, and in what range is even more interesting actually. Now its time for me to do some calculations at home actually, im clocking the time difference between my near field hairdryer speakers and subwoofer, and another compensation for my head speakers, the thought is to use one sub in a fixed position for both of my pairs and with different crossover, and on top off that different spl. So with that said, it wont do it by itself so im going to have fun for a day or two, and im going to find the sweetspot in the same time. im not near starting to put up absorbers and traps, not in weeks, just to mention them again.

 

[larsvendel74]

One more, when you know where the low frequency sounds best, most detailed in your room, same level on different bass notes in music, and sinus waves, that´s the x, where the work begins, that is the place where you will sit and listen, from there you make the triangle. And you will never find it starting to place the speakers first and then start to tweak them around in the area where they stand in most common living room, they will be placed at that wall you decide off course but now im not saying any more, if the tv wont fit between, to bad,,, and off course it all depends how far one is prepared to go. think structure bound sound as well, speaker cabinet when does the radiation or at what Hz do they stop spreading out sound like the sun spreads light,, Lars

 

[seangoodman]

Thanks again for your reply, there is a photo in my first post of a room sketch, dimension are in cm. the ceiling is 2.75m high and flat.
No eq or bass controls are set during measurements.

regarding the dips in the 200-500 range i dont expect or want to eq them out, i would like to understand what is causing them and treat the room accordingly to improve

most of my listening us done vinyl so i want the room to be as good as possible first as i can perform any eq in my analogue system. Not even bass or treble adjustments.

once i have achieved this i will concider eq for the digital music but first and format is natural improvements

 

[seangoodman]

here are my latest Mdat files one for left and one for right. I was testing out at Isover pad behind the speakers the pad was 140 x 50 x 10cm think. there is a measurement with and without pad for each side. the idea was to do a mock up to see if something like a GIK 244 behind the speakers on the from wall would offer something regarding SBIR, these measurements are taken following Johns advice earlier in this thread

 

[EarlK]

If you think it's SBIR, then the obvious course of action here ( at least to me ) is to take 3 of your 6 existing 242 baffles, put them into the corners behind & beside the left speaker and then remeasure to see if there's an improvement in response ( iow; see if there's less low-mid suck-out ) .

Direct action short-circuits all speculation here .

 

[seangoodman]

If you think it's SBIR, then the obvious course of action here ( at least to me ) is to take 3 of your 6 existing 242 baffles, put them into the corners behind & beside the left speaker and then remeasure to see if there's an improvement in response ( iow; see if there's less low-mid suck-out ) .

Direct action short-circuits all speculation here .

Yeah i think thats a good idea, why do you say 3 of them though? One on the side and double up with two behind?
as you say got to be worth trying, in the mdats i just posted i did a before and after just using a lump of isover insulation i had kicking around. 10cm thick. To me the measurement regarding FR, waterfall and indeen spechtograph didnt look much improved i did however thing the impulses looked better with the pad

any body got any thought
s on the difference between with and without the home made pad? Will do your suggested test maybe tomorrow

 

[sm52]

the impulses looked better with the pad

Yes.
The pit on the middle frequencies on the left, it seems to me, is due to the chair on the left. It is best to remove it from its place.

 

[seangoodman]

Yes.
The pit on the middle frequencies on the left, it seems to me, is due to the chair on the left. It is best to remove it from its place.

sadly doesnt have any effect at all, i tested with it out of the room. Same for tests on right side, removing the chair had not effect

 

[sm52]

Following. I would check if the right and left loudspeaker are the same. To do this, swap the positions of the left and right loudspeakers. If possible. If the hole would remains on the left, then the room is to blame.

 

[seangoodman]

Following. I would check if the right and left loudspeaker are the same. To do this, swap the positions of the left and right loudspeakers. If possible. If the hole would remains on the left, then the room is to blame.

There is the same whole on both left and right chanels, i am guessing when i find out why its happening on the left then i will have a good idea for the other side. Dali speakers are handed as well, matched pair labeled as right and left speaker

 

[sm52]

Then move any to the center of the room and measure.