sanity check: this isn't right, no? (faulty speaker content)

[minusnine]

Hi,

I've just come across this forum on my return to having something approaching decent hifi and have hit a problem that I'm hoping the assembled minds here might be able to help me identify.

I recently bought a pair of kef x300a active speakers that are hooked up via usb to a Mac mini. The stereo image seemed shifted to the right speaker, so I just installed REW following the tutorial here and recorded three frequency sweeps at my listening position (black) and (carefully measure) 10cm away from each speaker in turn (red right and green left). I used a tiny Samson go usb mic I have for lecture recording. The graph below shows my results.

Never having made these measurements before I was after a sanity check - there's something wrong with the left speaker, yes? I've only just got these - new open box, but with a 5 year warranty, so I'm not expecting any issues with getting them fixed or replaced if necessary, but I just wanted to check I wasn't missing anything. I have checked that balance on the Mac and on the back of the speakers were both set to the centre.

Thanks for any help,
Andy

 

[RayDunzl]

They would appear to have a matching problem from the above.

Operate all the controls a few times...

Operate the speakers with different gain and balance settings...

See what you can find out.

Does the balance change with the gain control?

Can the balance sound/measure correctly when not centered?

Cables connected well?

Do they need a reboot?

 

[minusnine]

Many Thanks Ray, I’ve just been doing as you suggested (switched off and on again, changed master level on back of speaker, changed balance) and no change unfortunately. I can move the balance to the right place if I change it using the dial on the speaker, but the freq response of the left speaker never reaches the level or shape of the right.

 

[RayDunzl]

I've only just got these - new open box, but with a 5 year warranty

Good luck on that.

 

[Sal1950]

Have you switched the speaker position L & R?
Does the imbalance follow the speak change or remain constant?

 

[Sal1950]

sanity check

 

[minusnine]

Hi Sal, each speaker gives the same response whether it is on the right or left. playing with the balance on the back of the speaker I can match levels only at the extreme of the range but the curve shapes look very different for L vs R, and the sound is subjectively shifted entirely to the left. Best subjective balance is with a slight shift to the left but the measurements still show a big (+/- 20 dB) mismatch.

I haven't mentioned the room , which is small (9x8 ft), the speakers are roughly symmetrical in the room either side of a desk, listening position is near the rear wall. Right hand side of the room is floor to ceiling bookshelves, left hand side is stud wall with pictures on it and there's a window directly behind the desk.

I guess my question resolves to how bothered should I be by the measurements? Will a small shift of the balance control be causing issues with the fidelity the speakers can produce, such as it is, or should I ...

Hold on, what's this sneaky little option here in the REW measure window? 'Default output'? probably that should be set to 'L+R', and not 'R' as I had it, right? And if I select L+R, the left speaker gives the same response as the right, and I've simply been measuring the output of the R speaker at the L speaker (which explains why the L speaker measurement looks a lot like the listening position measurement)

You may all queue up to call me a idiot now, clearly I failed the sanity check. thanks for your suggestions Ray and Sal.

 

[solderdude]

Common sense is a flower that doesn't grow in everyone's garden.

Sometimes it takes some time to explore the many functions and options in the excellent REW program.

 

[RayDunzl]

I recently bought a pair of kef x300a active speakers that are hooked up via usb to a Mac mini. The stereo image seemed shifted to the right speaker, so I just installed REW

What became of the initially reported problem?

 

[minusnine]

Well, it's still there, I just now know it's not a levels mismatch in the speakers causing it. Whether it's the room (fairly symmetrical position of speakers and listening position) or having bookshelves on one side and bare wall/pictures on the other? I'll play with my listening position again, but any other ideas welcome.

If they involve more REW measurements, just remember that apparently I'm sent here to stress-test the adage "there's no such thing as a stupid question"

 

[RayDunzl]

In REW, do a sweep measurement with both speakers active simultaneously, and the microphone well-centered at the listening position. Maybe put it at the tip of your nose.

Post a picture of the Impulse Response, with maybe a 10ms time range, using the "%" selection in the little window top left of the graph when the mouse is in the graph window.

---

It's possible there is some latency in the processing between the two speakers. Delay (in my opinion) is as effective (if not more so) than volume mismatch when studying the stereo imaging.

should look something like this:

and not this - has 0.2ms delay added to one channel:

The speaker whose soundwave arrives first will "sound" louder, pulling the image in that direction.

 

[minusnine]

Thanks Ray, the impulse signal showed it took about 200ms to return to baseline, but here is the first 10ms (from listening position, L+R output). measurements taken in front of either speaker look similar but with lower baseline. (edited to show %, though none of the signals get close to 100%)

 

[RayDunzl]

Nope, see above.

Use the % setting, not dbfs, pops up in the top left of window.

Although, I note two peaks in your dbfs display already (maybe).

Just change the setting on the window, don't have to remeasure.

Zoom in a little more, too.

 

[minusnine]

changed and zoomed in...

 

[RayDunzl]

Under the Controls button, select Plot Responses Normalized (to get the 100% reading).

There's a slight delay noted in the first graph. May or may not be noticeable in practice - your head or mic position will easily be that far off center.

I assume the red and green are single speaker?

Pull the "quieter" speaker an inch or two closer to you.

Is this a desktop or room-type setup?

 

[minusnine]

Thanks Ray, my room is a study (9ft x 8ft-ish), with the speakers 4ft apart either side of a desk but for listening I'm in a chair 5ft from each speaker. I've saved a series of plots as I have moved my chair (and mic) to the right (from as far left as the chair can go without blocking the door). Initial issue was balance feeling shifted to the right.
At 74, 82 and 89cm from the left wall I get the following impulse signals:

so it seems like I can create and lose that 2nd peak by moving my listening position. I will try listening in the last position (which is the furthest right).

 

[RayDunzl]

so it seems like I can create and lose that 2nd peak by moving my listening position.

You can lose it by moving a speaker, too, if that's more convenient (or possible). Pull the left one closer if the image is shifted to the right (or move the right one further away).

If you have some per-channel delay option in your playback software, that's another useful tool. When listening from my desk in the right back corner, I can set a delay to the closer speaker (other end of the room behind me) to get the stereo aligned to my way off-center location, basically directly in front of the right speaker but 9 feet away.

This is from a miniDSP in the signal chain. Sounds like proper stereo with, more like mono without, at the desk.

3.212m to the rightspeaker, 4.144m to the left, 120cm difference.

 

[solderdude]

Well, it's still there, I just now know it's not a levels mismatch in the speakers causing it. Whether it's the room (fairly symmetrical position of speakers and listening position) or having bookshelves on one side and bare wall/pictures on the other? I'll play with my listening position again, but any other ideas welcome.

Swap left and right speakers ?
Don't use stereo music (make it mono)
If the imbalance stays on the same side its the room.
When the imbalance changes to the other side its the speaker.

 

[minusnine]

Thanks Solderdude and Ray,

Having played with swapping speakers (no change) and changing toe-in (I initially had the speakers pointing straight at me, now I have them facing straight down the room and I will keep playing with positioning to see if there's a particular best amount of toe-in), I think I can more precisely describe the issue that I am sometimes hearing. Sometimes high treble content on the right of the soundstage sounds like it is coming directly from the right hand speaker, whereas I don't get the impression on the left. This might be an artefact of the recording as in other material imaging is good (as in I can close my eyes and point to where the voice or instrument is coming from, whether it is at the front or back of the stage). Subjectively this is one of the things I wanted/expected from a decent hifi system so I am maybe focusing on it more than I should.

Ray, my next project will be to get a MiniDSP like yours and then find a second hand 10" subwoofer to try to get a full range output and use the REW EQ to do some room correction. Can't/won't spend big bucks but it looks like a fun thing to play with and may improve the sound further.