[|Tch0rT|]

Sorry if this is the wrong forum for this post but I'm hoping people more experienced in reading REW would be more likely to see this post than in the main Speaker forum. Mods move if needed please.

I've measured my MartinLogan Motif and personally I feel like there's something off about the frequency response but when I contacted ML about it they said nothing was wrong.

Measurement I sent to ML:

ML Response:

At first glance it would not appear that there is anything wrong with the Motif. I don’t see anything to indicate a non-working or out of phase driver in the speaker. Please note the following:



1. The “On Wall” position on the back of the Motif will reduce the speakers output by 3 to 5 dB below 500 Hz. You might experiment with this switch to see what gives you the best results, regardless of the placement of the speaker.

2. This response might suggest more of an issue with surfaces in proximity to the speaker that could possibly cause some phase cancellations. ( IE, reduced output between 240 and 700 Hz.) Or it may be simply an issue with trying to get an accurate measurement in your particular space. With panel speakers sometimes you can get a more accurate reading by measuring slightly off axis ( say 20 to 30 degrees).

3. I would also check the settings in your processor or receiver. In general you would want the center set to “small”, but again experiment and see what sounds best to you.



In short, if everything sounds ok, it probably is ok. Go with what sounds best to your ears regardless of any graph that may be generated.

I dunno, that 240hz to 700hz looks wonky to me, I know a narrow dip isn't that audible but man it really drops off over a large band IMO.

So I did some more measurements:

On/Off Wall switch changes:

Off Center Measurements:

My Motif was a refurb so that adds to my doubt. I know it's different speakers and placements but I don't notice any of my other ML speakers having that behavior and for comparison my previous center, ML Encore TF:

Anyone have any ideas? Am I just over thinking it? I know it'll never be "flat" but to me it looks either defective (like I said it was a refurb so it fuels my doubt) or intentional and due to ML response it makes me believe they tuned them this way.

Thinking back on it I didn't measure with the phase reversed. I should measure it like that. Anyhow I can get the mdat if needed too.

Thanks!

 

[AudiocRaver]

I agree, it does not look right to me. That is definitely an audible deficiency, not explainable by wall proximity, and I cannot imagine anyone saying it was normal.

 

[|Tch0rT|]

Thanks for the reply. I thought it might've been a ceiling reflection at first but after changing the position and off center measurements prove that it's not. I should email them the extra info but given their first reply basically reducing it down to "In short, if everything sounds ok, it probably is ok. Go with what sounds best to your ears regardless of any graph that may be generated." doesn't really give me the best vibes about it.

 

[AudiocRaver]

A ceiling reflection, or any single reflection, will not cause a frequency response disruption that dramatic. Try a close-up (4 to 6 inches away) measurement straight on-axis for each driver. You should see the crossover frequencies clearly, and the response for each driver should be very flat over its active frequency range.

 

[|Tch0rT|]

I emailed ML the other graphs and they set up an RMA for testing. I'll have to try those measurements before sending it out. From my knowledge of audio I'm thinking the crossover might have a issue like one of the components is bad/out of spec but I'm not very knowledgeable about passive crossover design. I haven't measure a lot of speakers but the ones I have do not have a dip that broad and odd at the crossover points, usually it seems to be a sharp narrow dip. The crossover freq are 450 & 4000 Hz. My other thought about it is the problem of horizontal MTM style centers having lobing issues despite this being more of WMTMW (I suppose since the panel is so wide). Measuring them in the way you suggest should at least show me if it's a lobing issue. I guess I know what I'm doing after work in the morning LOL.

 

[|Tch0rT|]

Try a close-up (4 to 6 inches away) measurement straight on-axis for each driver. You should see the crossover frequencies clearly, and the response for each driver should be very flat over its active frequency range.

Here are those measurements, I clipped the tweeter but I'm not that interested in that measurement too much at the moment...

I averaged it just for fun...

That's ideally more or less what I guess what it "should" be. :dontknow:

mdat: https://www.dropbox.com/s/2iz5s84tz8y44jg/ML Motif on axis all drivers.mdat?dl=0

 

[AudiocRaver]

Thanks for including the .mdat file, that is helpful. Looking at the more recent measurements included in that file, they do appear much closer to what would be expected for that kind of design.

Interesting that the big drop from 200 Hz to 300 Hz does not appear in these measurements. Without it, the remaining anomalies in the earlier measurements are more explainable as cancellations from reflections. Don't know exactly what is different about the measurements, but the later ones do appear fairly normal.

 

[|Tch0rT|]

I've merged all the previous measurements other than the one from measuring the drivers individually into a single mdat and with proper notes:

https://www.dropbox.com/s/ef3atr8zfctekln/ML Motif - all recent measurements.mdat?dl=0

I can do more if you have any suggestions before I send out my center as I don't want to end up paying for shipping and finding out that it was just fine and some anomaly in my measuring. My room is about 12.5 feet wide, 20ft long, and 8ft high. The biggest two things of note is that the first and oldest measurement was done with a CSL Dayton UMM-6 measured on top of the TV about 53" off the floor and the crossover was set to 80hz. I sold the UMM-6 and got a CSL miniDSP UMIK-1 (in case I ever went with one of miniDSP's Dirac products) so all other measurements were done with the UMIK and below the TV on a stand about 10" off the floor and the crossover is set to 100hz.

 

[Wayne A. Pflughaupt]

Is the speaker inside a cabinet or entertainment center of some type?

Regards,
Wayne

 

[AudiocRaver]

I've merged all the previous measurements other than the one from measuring the drivers individually into a single mdat and with proper notes:

https://www.dropbox.com/s/ef3atr8zfctekln/ML Motif - all recent measurements.mdat?dl=0

I can do more if you have any suggestions before I send out my center as I don't want to end up paying for shipping and finding out that it was just fine and some anomaly in my measuring. My room is about 12.5 feet wide, 20ft long, and 8ft high. The biggest two things of note is that the first and oldest measurement was done with a CSL Dayton UMM-6 measured on top of the TV about 53" off the floor and the crossover was set to 80hz. I sold the UMM-6 and got a CSL miniDSP UMIK-1 (in case I ever went with one of miniDSP's Dirac products) so all other measurements were done with the UMIK and below the TV on a stand about 10" off the floor and the crossover is set to 100hz.

Thanks for the additional mdat file and all the additional detail about the measurements. With all the additional detail in mind, it does appear that the unit could be in spec and that the frequency response variations are explainable by cancellations from reflections. To experiment with seeing the difference a floor reflection can make, for instance - it is an experiment only, not at all practical for normal use, but it can help you see if your speaker is behaving - it takes several layers of thick, fluffy blanket with each layer further thickened by bunching and wrinkling it, all precisely placed to eliminate a floor reflection.

Based on your most recent info, I would probably not return the speaker, but of course that is your call.

 

[|Tch0rT|]

Is the speaker inside a cabinet or entertainment center of some type?

No, I have a shelf on top of my TV with nothing to the side or above it other than the walls and ceilings, it's maybe 3 ft from the walls and dead center between 12.5" width of the room. I've experimented with putting it in front of the TV on a little stand recently. Looks like it'll probably go back on top.

With all the additional detail in mind, it does appear that the unit could be in spec and that the frequency response variations are explainable by cancellations from reflections.

Can you show me where in the data I can see that so I know what to look for? I know from the impulse response graph that when it's placed in front of the TV I can see what I'm guessing is floor bounce. Could that be the cause of the dip at 300Hz? If I'm reading it right there's some phase issues in the dips on both top of the tv and in front of it? Do you think something like Dirac, Audiolense, or Acourate would possibly be helpful for an issue like this? I can't get away with ceiling mounting an acoustic panel since I rent. I thought about measuring it outside but it was about -12 F today, maybe in warmer weather...

Thanks again. :T

 

[AudiocRaver]

Can you show me where in the data I can see that so I know what to look for? I know from the impulse response graph that when it's placed in front of the TV I can see what I'm guessing is floor bounce. Could that be the cause of the dip at 300Hz? If I'm reading it right there's some phase issues in the dips on both top of the tv and in front of it? Do you think something like Dirac, Audiolense, or Acourate would possibly be helpful for an issue like this? I can't get away with ceiling mounting an acoustic panel since I rent. I thought about measuring it outside but it was about -12 F today, maybe in warmer weather...

Thanks again. :T

Measuring outside can really help, when the weather allows, but even then there is the ground reflection to deal with. I have a few old blankets I bunch up and pile together to deal with it.

Your most normal looking graph in your latest .mdat file is #3, "Center 1/2015 @ 3ft." With 1/48th octave smoothing, you can see the cancellation notches most clearly. The main frequencies are at 290 Hz and 420 Hz. The formula to use is

1 / ( frequency x 2 ) = delay

So the notches at 290 and 420 correspond to the reflections at 1.2 ms and 1.7 ms on the impulse diagram, measured from the beginnings of the impulses, not the peaks. See the attached photos.

That shows best what the speaker is capable of, and that it is probably not defective. The better MLP curve is #1. With 1/3 or 1/6 octave smoothing, which is more like what we actually hear, it really doesn't look that bad. Don't know how you feel about over the TV vs under, most would probably pick under TV for most natural source direction, but over might be the better frequency response choice in your case.

Room correction software can help, but if it was me and I got the #1 curve for center channel which is primarily dialogue for most movies (yes, that is an over-generalization) I would probably live with it - unless the system needed lots of correction and I had to buy a room correction solution anyway.

Those are my thoughts from balmy Nebraska (24 degrees:sn.

 

[|Tch0rT|]

Thanks again Wayne. The Impulse response have always kind of confused me somewhat since people either have one like the style of your view % Fs or the dB FS and I don't know which one to use. I thought the peaks were what to look for. Thanks for the formula. I looked at the measurement 1 from the top of the TV and the dip is about 514Hz which comes in to just under 1ms so about 1ft away causing the dip if I understand correctly but that doesn't make sense to me as the only thing near the Motif in the range is the TV it's sitting on... well maybe. The Motif is curved and it's on a base that can tilt and I tilt it down towards the MLP so maybe the curve of the speaker makes the woofers go behind the TV slightly and bounce... Hmmmm.

I'm not sure which compromise on position I like better. On top of the TV the FR does look better and I think it sounds better but looks somewhat odd with a big center on my big TV (73"), it reduces the effectiveness of my bias light a tad but not too bad, and panning between center and L & R's you can definitely hear the change in height. In front of the TV the vocals sound slightly weird but it panning between the center and L & R's is precise (I rarely hear this transition in most material though) and it doesn't reduce the effectiveness of my bias light. I'll probably put it back on top of the TV and do some more experimenting.

I don't think my system needs a lot of correction but I could be wrong, I should probably make a post about it LOL. My home theater is my living room and my wife requires the common L shape with the couches and love seat (that's actually about the only WAF thing I have to deal with in this room) and I know that's throwing off some things and makes it so I can't really add panels for the reflections (the opposite side are windows ugh). The room correction is just icing on the cake or the cherry on top I suppose, at least until I can get into a dedicated room and do room treatments.

 

[AudiocRaver]

Thanks again Wayne. The Impulse response have always kind of confused me somewhat since people either have one like the style of your view % Fs or the dB FS and I don't know which one to use. I thought the peaks were what to look for. Thanks for the formula. I looked at the measurement 1 from the top of the TV and the dip is about 514Hz which comes in to just under 1ms so about 1ft away causing the dip if I understand correctly but that doesn't make sense to me as the only thing near the Motif in the range is the TV it's sitting on... well maybe. The Motif is curved and it's on a base that can tilt and I tilt it down towards the MLP so maybe the curve of the speaker makes the woofers go behind the TV slightly and bounce... Hmmmm

A 1 mS delay equates to about 1 ft longer path length (1.1 ft is more precise) than the direct path length. So if the direct path from speaker to mic is 8 ft, the delayed path length will be 9.1 ft long.

The reflecting surface will also have the correct "angle of incidence" to support a reflection in that direction. We often forget that changing the angle of a surface can do all we need by diverting a reflection, and can sometimes be a lot easier than deadening.

I also keep forgetting to mention... and this could account for a lot of what you are seeing... Often reflections with 1 ms to 1.5 ms delays are from the listening chair. When we sit in the chair, we completely change its reflective nature, so for measurements I like to take it out of the equation. A thick quilt with a furry side is perfect for this if you do it right.

• Bunch it up so it is thick and soft with as much fluffiness and air as possible.
• Completely cover the seat, arms, and back of the chair.
• Shape it so there is no reflective "angle of incidence" surface. This can be easier with a pillow under it, positioned like a person sitting in the chair and slumped down to a 60 degree angle.
• This one sounds weird, but it really works great... Using a mic with 90 degree angle calibration file, hang the mic by its cable at the MLP "center of head" point, with the tip of the mic just touching the surface if the fluffed out blanket.

Sounds like overkill, but it really is not. Even plushy blankets reflect sound, you need all that thickness and air and the disruption of angle and the mic touching the blanket to completely clean up those chair reflections.

Get rid of all chair reflections and see what is left.

 

[Tonto]

Deleting this post as I misunderstood what I read, so it does not apply. :dontknow: Sorry & glad you worked it out!

 

[|Tch0rT|]

A 1 mS delay equates to about 1 ft longer path length (1.1 ft is more precise) than the direct path length. So if the direct path from speaker to mic is 8 ft, the delayed path length will be 9.1 ft long.

Ah that's where I went wrong, I didn't add in the delayed path so I was looking in the wrong areas. Doh!

I also keep forgetting to mention... and this could account for a lot of what you are seeing... Often reflections with 1 ms to 1.5 ms delays are from the listening chair. When we sit in the chair, we completely change its reflective nature, so for measurements I like to take it out of the equation. A thick quilt with a furry side is perfect for this if you do it right.

Ya know my initial thought was it that it was reflections from my leather love seat but when I did the math wrong I focused on ceiling reflections then floor reflections and forgot about it possibly being my loveseat. Then I moved to another position before doing more measurements from the initial location. I put the Motif back on top of the TV and I did a few more measurements, some with blankets but they didn't make much of a difference but anyway here's proof it's fine and it's just the loveseat ugh:

mdat:
https://www.dropbox.com/s/4nuxmh5egne77v2/Motif 2-21-2015.mdat?dl=0

Thanks again for bearing with me on this. I should've figured it out on my own if I would've just measured more.

At least I know my Motif isn't defective now. BTW the miniDSP nanoAVR-HD I won from here totally makes measuring my 7.1 system with REW easy peasy due to routing feature and verifying the EQ work. :bigsmile:

 

[AudiocRaver]

Love that nano series.

Sorry I didn't mention chair effects earlier. I do that stuff so automatically, I almost forget the effect it can have.