Dayton Omnimic Discussion Thread

[Sub Sonic]

Hi All,

Having recently bought a Dayton Omnimic V2 measurement system (thanks to site sponsor Paul Spencer/The Loudspeaker Kit), I thought I'd start a discussion thread for those learning how to use it.

It certainly seems like a capable piece of kit, but I have a long way to go to become proficient in it's use. Hopefully other forum users could help those of us who are keen to learn, helpful tips and hints as well as pitfalls for the unwary.

The kit comes with the USB microphone, a very small tabletop mic stand, 6' USB lead, case, software disc and test tones disc. The test tones are tailored specifically for use with the system. There is a 60 page Help Manual on the software disc which seems pretty comprehensive.

Discussion of other measurement systems or methods are also welcome.

Regards,

SS

Edited May 24, 2014 by Sub Sonic

[Sub Sonic]

First question , I have been playing with some drivers and thought I'd also do a sweep of the lounge room. Please be aware that the speakers (home made) and positioning are far less than optimal due to current room restraints and also due to them having to be resistant to pokey little fingers.

They consist of a 6" bass driver and 25mm tweeter, but I have been able to retain the VAF SW19 subs in the front corners.

Question: how much can a simple sine wave sweep tell you about the room's acoustics, and what things do we need to be aware of?

Thanks!

SS

Edit: 1/12 octave smoothing applied to the graph above.

Edited May 24, 2014 by Sub Sonic

[henry218]

First question , I have been playing with some drivers and thought I'd also do a sweep of the lounge room. Please be aware that the speakers (home made) and positioning are far less than optimal due to current room restraints and also due to them having to be resistant to pokey little fingers.

They consist of a 6" bass driver and 25mm tweeter, but I have been able to retain the VAF SW19 subs in the front corners.

Question: how much can a simple sine wave sweep tell you about the room's acoustics, and what things do we need to be aware of?

Lounge Frequency Response.jpg

Thanks!

SS

is that for 2 speakers? tell us more of your mic placement and omnimic settings

[Sub Sonic]

Hi Henry,

The plot is of a pair of speakers, pretty much hard up against the short wall in a room approx. 3.6m x 5.5m x 2.4m high. The speakers spaced apart by only about 1.5 metres and are flanked by fabric covered single lounge chairs. There is a wall to wall to ceiling window behind the speakers which has light lace curtains across it. The floor has fairly thick carpet and underlay. Tweeters are around 800mm off the floor.he speakers are on short stands and are on top of the subs, both sub drivers are in the lower front corners of the room.

The mic was simply hand held, pointed toward the centre of the two speakers, approx. two metres away at the listening position, at tweeter level. The plot above is of both speakers together.

I am aware that this is way less than perfect :-) just experimenting and getting to know the Omnimic at this stage :-)

Regards,

SS

[henry218]

for starter, you can use the small stand on top of a chair or stacked books

then smoothing, depends on your purpose, if for room compensation, then i would think you should concentrate below 200hz with 1/6th and 1/12th then compare to non smoothing with 4.5-6msec gating and compare to ALL (no gating)

but, since you've got a powerful tool, i would suggest measure your speakers as a quasi anechoic first to see the real response.

set the mic around 90cm, tw axis then apply proper gating.

next step is to do nearfield, where you need to put the mic just in front of the dustcap with 250msec gating.

then you need to blend them

cheers

Henry

[davewantsmoore]

Question: how much can a simple sine wave sweep tell you about the room's acoustics, and what things do we need to be aware of?

 

Everything.

 

If you can view the data as impulse response, step response, energy time curve, waterfall, distortion ....  you can see everything there is to know.    Obviously these things will depend on position in space... so you may need to take many of the same measurements to analyse how they change with location and/or speaker radiation angle.

[davewantsmoore]

Lounge Frequency Response.jpg

 

 

It's difficult to make a diagnosis without seeing more (of the same, remember it contains everything) measurements from different locations.

 

It's tempting to want to think that your crossover is about 3khz, and the 10dB (enormous) hole there is because you crossover is problematic....   but we would want to see more data to see if the hole it really coming from the speaker, or if it's an acoustic effect from boundaries/placement.

 

 

is that for 2 speakers?

 

Hmm, yes.  I didn't think of that.   Only ever measure one speaker at a time.     Measuring the L and R at the same time, really only demonstrates how equal they are in response, and how well placed they are.

 

The mic was simply hand held, pointed toward the centre of the two speakers, approx. two metres away at the listening position, at tweeter level. The plot above is of both speakers together.

 

Ah, right.    That's my usual trick of not reading to the end of the thread before jumping in.

 

Get a mike stand (or improvise one) ... and just work on one speaker   (until you really know you're interested in both together --  which you rarely will be) ....   Put the mic one or two meters away from the speaker, straight on.... height = listening axis .... keep everything (mic and speaker) as well away from boundaries as you can.    Don't worry about moving things out of your regular listening positions for now, assuming you want to see your speaker rather than your room.

... but like you say.  Just play.   You'll get a whole host of different responses depending on where you place things.... and you'll ask "why?!"

[Sub Sonic]

Thanks Henry/ DWM.

Does the gating set a specific time frame that the system will "accept" a response? Is this to limit room interaction, reflections etc from affecting the measurement?

I'll have to take the speakers outside and get a decent baseline to work from, then try it indoors with appropriate gating and compare the graphs.

Davewantsmoore, you're right, I'll need to take notes on setup so that any external influences in measurements will be predictable and can be taken into account.

Regards,

SS

[henry218]

the gating enables you to measure in your living room, but i prefer to raise up the speakers.

perhaps put some thick blanket on the floor between speakers and mic.

look at impulse response. but typical living room should be about 4.5-6msec.

another thing is, put your speakers at the middle of your living room, so you will minimized reflections.

[davewantsmoore]

Does the gating set a specific time frame that the system will "accept" a response? Is this to limit room interaction, reflections etc from affecting the measurement?

 

Yes, that's right.

 

By setting the length of time for which the measurement is captured, we ignore all reflected sound which arrives outside this window..... also, we lose all frequencies which are of longer length than the time window.    This is why measurements of 'bass' in a room is not a measurement of the speaker, but really a measurement of the room.    The time window required to capture one cycle, is so long that the sound has bounced many times around the room before one cycle is measured  (hence you can't avoid the effects of the room).

 

 

To determine where to set the "gate"  (window length) .... Look at the impulse response.   You'll see the peak from the direct sound.... and then some milliseconds later more small peaks (reflected sound) will arrive.   You want to set the window shorter than when the first reflection arrives  (which will show you the speaker uncontaminated by the room).    By moving speaker and mic away from boundaries, you can increase the amount of time before that first reflection arrives, which allows you to increase the time window for the measurement.

[davewantsmoore]

Davewantsmoore, you're right, I'll need to take notes on setup so that any external influences in measurements will be predictable and can be taken into account.

 

A good way to tinker and learn, is to first learn how to take a measurement of your speaker which represents it uncontaminated by the environment (room).      Ideally far from boundaries, and using the correct time window to remove the rooms contribution.

 

Once you have this graph of your speaker.... you can go forth and take all manner of measurements in your room.   Anywhere you like.... any gating you like.... just play.    Each measurement you get, will be different in some way from the speaker-only data (ie.  speaker without the room).    Your task will be to understand the cause of why each one is different.    The goal here is to give you an understanding of the two main things affecting sound.     For the LF.  Boundary effects / reflections / room modes  (ie.  what the room does)    ... and for the HF.   Speaker directivity / polar response / radiation pattern  (ie.  the fact that speakers do not radiate sound equally in all directions).

[Sub Sonic]

Thanks Gentlemen,

I'll have some fun playing with it :-) The instructions are pretty good but getting help from others is just as important :-)

Cheers!

SS

[Sub Sonic]

Here's a FR graph of a speaker I put together years ago, Seas 4.5" bass mid, Audax TW010F1 tweeter, small 1.5l sealed enclosure.

 

It was enlightening having a play in real time, moving the speaker around and watching the FR. The tweeter was 900mm above the carpet, mic was 900mm distant on tweeter axis, 1/12 octave smoothing, gating was 5ms (from memory). It needs some work, as soon as it goes off axis, a hole around 4k appeared, getting rapidly larger the further off axis it was. You can see the start of the dip even on axis. I will replace the tweeter with a 25mm soft dome tweeter and bring the crossover point down a bit. The bass mid driver has response which starts falling off around 3k, and I think there is too large a gap which the current tweeter cannot really bridge.

 

 

Any suggestions welcome.

 

Regards,

 

SS

Edited May 26, 2014 by Sub Sonic

[henry218]

Here's a FR graph of a speaker I put together years ago, Seas 4.5" bass mid, Audax TW010F1 tweeter, small 1.5l sealed enclosure.

 

It was enlightening having a play in real time, moving the speaker around and watching the FR. The tweeter was 900mm above the carpet, mic was 900mm distant on tweeter axis, 1/12 octave smoothing, gating was 5ms (from memory). It needs some work, as soon as it goes off axis, a hole around 4k appeared, getting rapidly larger the further off axis it was. You can see the start of the dip even on axis. I will replace the tweeter with a 25mm soft dome tweeter and bring the crossover point down a bit. The bass mid driver has response which starts falling off around 3k, and I think there is too large a gap which the current tweeter cannot really bridge.

 

Natth, Audax, Felt.jpg

 

Any suggestions welcome.

 

Regards,

 

SS

what's the gating?

 

sorry, missed the word gating on your post .

 

something not right definitely , post the near field measurements for the mid woofer and 90cm for the tweeters

Edited May 26, 2014 by henry218

[Sub Sonic]

Definitely something wrong, it was done fairly quickly by ear 10-15 years ago LOL! I'm beginning to appreciate how useful a decent measurement system is, long way to go though, lots to learn.

The gear is packed up for the night, but will post the results when I do them. I could have left it set up if it was in the shed but noise from the rain would have interfered with the measurements.

BTW DWM, I now have a mic stand, one less variable in the equation :-)

Regards,

SS

[davewantsmoore]

If you can post the tweeter, woofer, and combined responses on the same chart, that would help to see what is up.

[svenr]

Get a copy of Joseph D'Appolito's "Testing Loudspeakers".  That will tell you all the important aspects and saves you many hours of forum discussions.

[Sub Sonic]

Just been having more of a play with the Omnimic. I thought I'd measure a speaker with a better response and play with a few of the variables to see what difference it made. The speaker was a Duntech PCL 10, measured approx. 1300mm above the ground, at 700mm distance, between the woofer and tweeter, 4ms gating. It would have been nicer to have a greater distance but proximity of other objects would have caused more discrepancies/errors. I also tried different smoothing.

 

First pic is with 1/12 octave smoothing, the second pic is with zero smoothing. I'd be very happy to be able to design a speaker like this

 

 

[henry218]

whats the total height of the speakers? you will need to have at least as far as the height of the speakers or at least 2x of the baffle width.

[Sub Sonic]

Hi henry218, the speakers themselves are 275mm high but were placed on top of other speakers around 1100mm high. No drivers in the other speaker which were foam filled and faced away, so hopefully there was not too much interaction. I did play with placement on the other speaker and it seemed to make almost no difference.

 

Regards,

 

SS

[henry218]

Hi henry218, the speakers themselves are 275mm high but were placed on top of other speakers around 1100mm high. No drivers in the other speaker which were foam filled and faced away, so hopefully there was not too much interaction. I did play with placement on the other speaker and it seemed to make almost no difference.

Regards,

SS

when i read duntech i was expecting large floorstanders , didnt realise this is bbc style monitor.

nice speakers , is the midrange slightly forward?

Edited June 2, 2014 by henry218

[Sub Sonic]

They are some of the most natural speakers I have heard. It would be interesting doing an anechoic measurement as they are time aligned/designed for a listening distance of 3.5 metres. They also have some serious felt treatment around the drivers which could affect the combined response at close range. I'll have to try some measurements with the felt removed (non destructive, it is pushed/velcroed on).

[davewantsmoore]

That's a very nice direct response.

 

The next essential step in characterising the speaker is to take measurements at the same distance, but different angles of radiation.

 

The direct response is important, but it doesn't tell us as much as people often think ......   Once you have more data, you can start to judge how the speaker will sound in a room, and how well it's crossover is designed  (the sum/cancel between drivers is complex and changes with listener angle/location).

[Sub Sonic]

What is the best way of measuring in room subwoofer response? I've been playing with the Omnimic again and getting impressive looking graphs as far as low end response goes, but not sure which settings to use to end up with a meaningful result.

I've just been using the normal frequency response graphs with the bass sweep, with the mic at the listening position, any suggestions or corrections welcome :-)

Thanks,

SS

[davewantsmoore]

The things to understand is that you are also measuring your room as well when measuring subwoofer frequencies.   Peaks and dips in the response are being caused by your room/placement ...  This is fine, because it's what you are interested in when setting up a subwoofer in your room.

 

If you were designing / tweaking the subwoofer.... then to look at the subwoofer true response, you would need to measure outside and/or very close to the woofer.