Future testing, data analysis and presentation - Page 6 - diyAudio
Go Back   Home > Forums > Commercial Sector > Manufacturers > Markaudio

Markaudio Designers and builders of audiophile grade drivers

Reply
 
Thread Tools Search this Thread
Old 1st March 2014, 01:07 AM   #51
smf is offline smf  United Kingdom
diyAudio Member
 
Join Date: Jan 2005
Location: Hong Kong
Mark - I test with REW and HolmImpulse, using an ECM8000 mike (but uncalibrated unfortunately). I'll likely move to a minidsp Umik. I have to test in my sitting room which again is not ideal but probably typical for a lot of users - especially if you live here in HK or a city and don't have a quiet outside space. I'd be happy to help beta test a bit - I live 10 mins from Tuen Mun.

On the subject of data: I'd also be interested in off axis data out beyond 30 degrees to understand how the reflections could impact the listening position response. I do every 15 degrees, but should do smaller increments near the on-axis. On the smooth v raw data question, I'd go for raw personally but understand the commercial issues on this. Would you be able to do frequency dependant windowing which approximates to how we hear? That would be similar to 1/6 octave windowing.

Stephen
  Reply With Quote
Old 1st March 2014, 01:28 AM   #52
diyAudio Member
 
markaudio's Avatar
 
Join Date: Nov 2008
Quote:
Originally Posted by smf View Post
Mark - I test with REW and HolmImpulse, using an ECM8000 mike (but uncalibrated unfortunately). I'll likely move to a minidsp Umik. I have to test in my sitting room which again is not ideal but probably typical for a lot of users - especially if you live here in HK or a city and don't have a quiet outside space. I'd be happy to help beta test a bit - I live 10 mins from Tuen Mun.

On the subject of data: I'd also be interested in off axis data out beyond 30 degrees to understand how the reflections could impact the listening position response. I do every 15 degrees, but should do smaller increments near the on-axis. On the smooth v raw data question, I'd go for raw personally but understand the commercial issues on this. Would you be able to do frequency dependant windowing which approximates to how we hear? That would be similar to 1/6 octave windowing.

Stephen
Hi Stephen,
Many thanks for your input. Along with members Dewardh, XR and others, the +30 off-axis measurement is looking potentially popular. Useful to know as we start work on the new anechoic chamber. Its internal layout can be adjusted to suit.

Glad you're in Tuen Mun, I'm near Saikung, reasonably easy to meet up re Beta testing - many thanks

Thanks
Mark.
  Reply With Quote
Old 3rd March 2014, 12:16 AM   #53
Ivo is offline Ivo  Netherlands
diyAudio Member
 
Join Date: Jan 2005
Location: Amsterdam
Beware that USB microphones are not in sync with the clock in your soundcard, which affects some measurements. I am a newb at measuring and have only a vague idea what that means and what effects it will have on your measurements, so I am hoping an expert chimes in. ARTA recommends only using sweeps with USB microphones.

I have a Minidsp UMIK-1 USB microphone, I use REW as software and use HDMI to get the signal digitally from my laptop computer to the AV-receiver. I believe it is a very simple but adequately calibrated and solid measurement chain. So far, I have mostly measured my loudspeakers (tiny fullrange driver satellites and a subwoofer) in the nearfield, the in-room response and how Audyssey MultiEQ tries to fix it. I find that measurement results are repeatable. For instance my satellite speakers measure the same on each and every session, even at 20 kHz (provided I am precise at positioning etc). Moreover, they show what I expected based on published independent measurements of the driver units and software simulations.

Before getting the long HDMI cable, I used an analogue connection. Results were basically similar, started deviating about a dB below 40 Hz. Years ago, I had a first try with the ATB PC Pro measurement system. I didn't like the interface and was distracted by work, but I also managed to measure reliably with that software. While I am sure there are serious benefits to getting expensive measurement equipment and it will certainly come across as more professional to some, I suggest that the low budget systems are already pretty good, so maybe adequate for the home/beta-testers that Markaudio wants to work with.

My wishlist for measurements would be: as much as can be made available. On the other hand, it would seem unwise to publish more than other manufacturers do. I agree people seek too much behind data that they've never seen before... More elaborate off-axis response is interesting to me, as is waterfall.

But to be perfectly honest, I read the German magazines and your drivers are usually tested there in standardized ways. I get good info from those graphs which, combined with the information you share here about your design choices, is already a lot of information.
__________________
We want to be happy, bold and free
  Reply With Quote
Old 3rd March 2014, 01:37 AM   #54
diyAudio Member
 
markaudio's Avatar
 
Join Date: Nov 2008
Quote:
Originally Posted by Ivo View Post

While I am sure there are serious benefits to getting expensive measurement equipment and it will certainly come across as more professional to some, I suggest that the low budget systems are already pretty good, so maybe adequate for the home/beta-testers that Markaudio wants to work with.

My wishlist for measurements would be: as much as can be made available. On the other hand, it would seem unwise to publish more than other manufacturers do. I agree people seek too much behind data that they've never seen before... More elaborate off-axis response is interesting to me, as is waterfall.

But to be perfectly honest, I read the German magazines and your drivers are usually tested there in standardized ways. I get good info from those graphs which, combined with the information you share here about your design choices, is already a lot of information.
Hello Ivo, (Guys)
Many thanks for posting. For a "newbie" you've illustrated a great deal thoughtful and interesting insight into testing issues.

Hopefully, there will be more posts made on home-based testing equipment (hardware/software). I'm trying to get a more comprehensive picture of what "home testing" means for members and exactly how they go about checking drivers/loudspeaker systems.

The divide between most driver/loudspeaker system manufacturers and the Diy audio community (end users) is sadly too wide in my opinion. I've returned from a audio-driver manufacturing event having spent time with makers who will be familiar to many Diyers. I asked several professional colleagues at that meeting to read and comment on the similar thread in the Full-Range forum thats just been closed. Their reaction to that thread only served to maintain their opinions that home-testing is (quoting 3 colleagues) a "flawed" exercise. Lack on anechoic isolation, inadequate mics and so on. They were critical of the some of the posts on technical grounds, especially the driver operational knowledge on display. Sadly that thread appears to have re-inforced their negative opinions that the Diyaudio side of the market is too variable in practice and knowledge. That it will remain small and therefore doesn't warrant additional time and support.

Such thinking from the bulk of manufacturing side I think is mistaken. Equally, they think I'm "crazy (quoting) maintaining hands-on connections with the Diyaudio community, that I should devote more of my time to OEM production. Sadly, recent events on Diyaudio have done little to reduce this divide. Its not likely that the bigger driver makers will become actively engaged with the Diyaudio community anytime soon.

Which brings me nicely back to the latter of your comments. Hopefully, our finances will stretch to enable the purchase of Linear X500 or Clio (etc) as I'd like to provide more data. The bulk of our testing is component specific, especially cone testing, done on a multi-laser head rig in an isolated room. The benefits of all this work is the enjoyed by end-users in the form of very consistent close tolerances on all driver production.

In the meantime, I hope more posting will be done by members on the issue of semi-anechoic isolation. I'd much like to know how home testers go about isolating their equipment and test drivers/loudpeakers from room/environmental influences.

Later in the month, I will ask the same professional colleagues to read this thread. I think there's a possibility to influence (change) their current opinions as the comments in this thread are looking more mature in quality and perspective.

Thanks
Mark.

Last edited by markaudio; 3rd March 2014 at 02:23 AM.
  Reply With Quote
Old 3rd March 2014, 06:14 AM   #55
xrk971 is online now xrk971  United States
diyAudio Member
 
xrk971's Avatar
 
Join Date: Oct 2012
Location: Virginia
Quote:
Their reaction to that thread only served to maintain their opinions that home-testing is (quoting 3 colleagues) a "flawed" exercise. Lack on anechoic isolation, inadequate mics and so on. They were critical of the some of the posts on technical grounds, especially the driver operational knowledge on display.
Mark,
We can all agree that home testing is not going to be the same or as good as a dedicated professionally built and operated anechoic chamber. I think industry professionals who dismiss home testing as "flawed" are equating it to not having much value, and I think that they would be mistaken in thinking this. It is of value to the user who finds a response grossly different than published specs! And it is of negative value to the manufacturer when that user publishes that finding. I think it is great that you started this thread to encourage and also to provide users interested in measuring, a common framework and best practices to do so. I hope your colleagues don't dissuade you and that you continue to see the value in it because when users are armed with so much powerful equipment and knowledge via the web, manufacturers can't put the genie back in the bottle.

In the diy speaker building threads that I post, I have been providing both measurements and high quality sound clips of my speakers. Sure, some say that it is a flawed or pointless exercise to provide sound clips because it is not the same as being in the same room listening with your own ears, and that the sound coming out of the mp3 clip to computer speakers or headphones what have you... is not the same. My point is to provide as much info as possible and let people make up their own mind if they think it sounds good and they want to build it. I think that a good speaker may end up sounding bad if not recorded properly and played back on bad speakers, however, a bad speaker cannot be made to sound good via a good recording. Analogous to this, with the exception to room gain increasing bass extension, a bad driver cannot be made to measure better than it is physically able to perform (frequency response wise, impulse response, spectral decay, etc.). With gating or outdoor measurements, much of this can be negated.
  Reply With Quote
Old 3rd March 2014, 11:05 PM   #56
diyAudio Member
 
markaudio's Avatar
 
Join Date: Nov 2008
Quote:
Originally Posted by xrk971 View Post
Mark,
We can all agree that home testing is not going to be the same or as good as a dedicated professionally built and operated anechoic chamber. I think industry professionals who dismiss home testing as "flawed" are equating it to not having much value, and I think that they would be mistaken in thinking this. It is of value to the user who finds a response grossly different than published specs! And it is of negative value to the manufacturer when that user publishes that finding. I think it is great that you started this thread to encourage and also to provide users interested in measuring, a common framework and best practices to do so. I hope your colleagues don't dissuade you and that you continue to see the value in it because when users are armed with so much powerful equipment and knowledge via the web, manufacturers can't put the genie back in the bottle.

In the diy speaker building threads that I post, I have been providing both measurements and high quality sound clips of my speakers. Sure, some say that it is a flawed or pointless exercise to provide sound clips because it is not the same as being in the same room listening with your own ears, and that the sound coming out of the mp3 clip to computer speakers or headphones what have you... is not the same. My point is to provide as much info as possible and let people make up their own mind if they think it sounds good and they want to build it. I think that a good speaker may end up sounding bad if not recorded properly and played back on bad speakers, however, a bad speaker cannot be made to sound good via a good recording. Analogous to this, with the exception to room gain increasing bass extension, a bad driver cannot be made to measure better than it is physically able to perform (frequency response wise, impulse response, spectral decay, etc.). With gating or outdoor measurements, much of this can be negated.
Hi XR (Guys)
Thoughtful post, thanks.
Hopefully we can begin to better appreciate the complexity of the situation. The Diyaudio community rarely speaks with one voice on testing issues. The recent 3" or 4" driver thread on the Full-Range section became chaotic with no consensus on testing and data priorities. From a makers perspective, its a nightmare scenario. Who should the maker take notice of? Which home tests look reliable? I didn't see any in-depth descriptions of testing room environments on that thread. Makers are by en large creatures of specification habit. If there's no comprehensive agreed specs, they won't take much notice. Hence my attempt to bring some consensus forward on this thread.

In practical testing terms, one of the most challenging measurements is Off-Axis testing. It requires extremely accurate alignment between driver and mic and anechoic isolation to become reliable. +2 degree variance from centring and line-angle will give wide fluctuations in frequency plots. Aligning drivers on their caps, phase plugs or whizzers adds to the complexity. To date I can't recollect seeing any Diyer's that does home off-axis testing acknowledge the difficulties, rather expounds a blind assumption that the data produced is somehow magically accurate. Its in these types of contexts that my colleagues view home testing as largely inherently flawed and view the Diy driver supply market as far too variable.

To put the supply side into context of scale, in my China province (Guangdong) we have Foster (Fostex), Tymphany Group (Peerless, etc), me and a few others along with megga factories knocking out millions of drivers every month. The dedicated full-range driver production volume represents less than 0.2% of the 300 MILLION units produced every month in Guangdong. Of the 0.2% (classed as FR's going to 20-kHz) only a fraction go to the Diy market, most end up in posh LCD TV's et al. You can now better understand why there is little/no active engagement with the diyaudio market; And why a small company like mine is always commercially vulnerable. So when the odd member comes along, gets into "attack Markaudio"mode, all they do is put at risk the survival of the only maker who believes in Diyers and diyaudio and re-enforce the existing negative views held by most other manufacturers.

Future home testing needs to mature, improve techniques and methodology, deploy anechoic isolation (even if its not total) and agree some sort of operating and presentation criteria. If this can be done, I think there's more possibility of seeing those few makers willing to get involved in Full-Range take heart, enough to feel its worth increasing support for the hobby.

So lets hope over time this thread will receive more practical contributions. I really want Diyers to work closer together on testing, evaluation and data presentation. Glad XR you're prepared to say more your home set-up, hoping others will follow.

Thanks
Mark.

PS - Tymphany have a job opening for a Mechanical Design Engineer. Sharpen up your Mandarin, show them you really know all about speakers, have +5 years design experience and have a Bachelor Science Mechanical Engineering degree, you might be what they're looking for! Welcome to the big wide audio world fellas.

Last edited by markaudio; 4th March 2014 at 12:16 PM.
  Reply With Quote
Old 4th March 2014, 01:41 AM   #57
xrk971 is online now xrk971  United States
diyAudio Member
 
xrk971's Avatar
 
Join Date: Oct 2012
Location: Virginia
Mark,
The off axis testing was not what I would call hard - just a PITA. I used a protractor to draw a big angle guide on the floor at measurement positions (0, 15, 30,45,60,75,90 deg) I used a piece of string measured at 1 m to set the distance from mic to front plane of driver at bezel. I think the distance is accurate to 5 mm at each point and angle is accurate to 2 deg. I think the distance is more critical than the angle. That is why I used piece of string to set distance. I moved the mic around and left speaker stationary in room rather than speaker on a carousel with fixed mic. If I did this I would have put driver front plane at rotation centerline - but this only works in anechoic chamber. The test used the same drive level of course and this was set to provide about the same listening volumes I use. I did not reference to 2.83v - maybe next time.
Cheers,
Xrk971
  Reply With Quote
Old 4th March 2014, 02:02 AM   #58
diyAudio Member
 
markaudio's Avatar
 
Join Date: Nov 2008
Quote:
Originally Posted by xrk971 View Post
Mark,
The off axis testing was not what I would call hard - just a PITA. I used a protractor to draw a big angle guide on the floor at measurement positions (0, 15, 30,45,60,75,90 deg) I used a piece of string measured at 1 m to set the distance from mic to front plane of driver at bezel. I think the distance is accurate to 5 mm at each point and angle is accurate to 2 deg. I think the distance is more critical than the angle. That is why I used piece of string to set distance. I moved the mic around and left speaker stationary in room rather than speaker on a carousel with fixed mic. If I did this I would have put driver front plane at rotation centerline - but this only works in anechoic chamber. The test used the same drive level of course and this was set to provide about the same listening volumes I use. I did not reference to 2.83v - maybe next time.
Cheers,
Xrk971
Hi XR,
Glad you're measuring but you're missing the essential points or perhaps I've not adequately explained.

Alignment angles (and horizontal plane) are critical, not the distance. That can vary be a one or two cm without causing a large change in outcome. For off-axis, just a millimetre or so off the driver's centre can affect outcomes, especially on shallow profile wide dispersion cones designs.

We have a laser alignment system that allows us to accurately track alignment and positioning. The error rate on the wider axis increases dramatically. Without these types of alignment tools, its not possible to be accurate. You might want to try using a laser pen. I helped a few local testers improve their testing by using a pen that neatly slotted into place of the mic. But (BIG BUT) you have to be megga careful not to move any part of the stand when swopping back to the mic and the position of pen and mic in the stand mount MUST remain the same.

Thanks
Mark.

Last edited by markaudio; 4th March 2014 at 02:19 AM.
  Reply With Quote
Old 4th March 2014, 05:01 AM   #59
smf is offline smf  United Kingdom
diyAudio Member
 
Join Date: Jan 2005
Location: Hong Kong
Hi Mark - I can well see why many manufacturers wouldn't be interested in the DIY community. There just isn't the money. It would be a shame for them to dismiss the community on the basis that there's a lack of understanding/knowledge though, as there are some people who clearly do know what they are talking about, but sometimes identifying who these are isn't as easy as it might be!

On the subject of accuracy, I agree with you that setting some standard approach that home testers would have to stick to rigorously would be the only way of ensuring that data was meaningfully comparable. While it doesn't matter to me for my own personal use whether my 15 deg measurement is actually for 17, 18 or 14 degrees, it wouldn't be good enough for me to report to others that the 15 degree measurement of ABC speaker is X, when it isn't actually the 15 degree measurement to begin with. I do think though that super-accuracy is probably only called for over the first 30 degrees, with a small margin of error acceptable beyond that.

Near-anechoic conditions at home are going to be a bit of a struggle for most of us. But perhaps you simply mean gated measurements? I try to get a 5ms gate before the first reflection (the floor/ceiling in my case). I typically measure at either 1m or 1.5m, with 20cm of cushions/pillows on the floor between the speaker and the mike. I aim for c.75dB at the mike position. Mike is pointed directly at the driver under test. (When I'm testing multiple drivers as part of a speaker, I always test along the intended listening axis - normally the mid/tweeter axis.)

I don't know whether you've looked a lot at some of the home-testing software, but REW and HolmImpulse seem pretty good to me and fairly straightforward to learn (although understanding what everything means is more challenging for some of us). Maybe you'd want to specify particular settings to use too (again in the intertest of comparing like with like).

Stephen
  Reply With Quote
Old 4th March 2014, 12:49 PM   #60
diyAudio Member
 
markaudio's Avatar
 
Join Date: Nov 2008
Quote:
Originally Posted by smf View Post
Hi Mark - I can well see why many manufacturers wouldn't be interested in the DIY community. There just isn't the money. It would be a shame for them to dismiss the community on the basis that there's a lack of understanding/knowledge though, as there are some people who clearly do know what they are talking about, but sometimes identifying who these are isn't as easy as it might be!

On the subject of accuracy, I agree with you that setting some standard approach that home testers would have to stick to rigorously would be the only way of ensuring that data was meaningfully comparable. While it doesn't matter to me for my own personal use whether my 15 deg measurement is actually for 17, 18 or 14 degrees, it wouldn't be good enough for me to report to others that the 15 degree measurement of ABC speaker is X, when it isn't actually the 15 degree measurement to begin with. I do think though that super-accuracy is probably only called for over the first 30 degrees, with a small margin of error acceptable beyond that.

Near-anechoic conditions at home are going to be a bit of a struggle for most of us. But perhaps you simply mean gated measurements? I try to get a 5ms gate before the first reflection (the floor/ceiling in my case). I typically measure at either 1m or 1.5m, with 20cm of cushions/pillows on the floor between the speaker and the mike. I aim for c.75dB at the mike position. Mike is pointed directly at the driver under test. (When I'm testing multiple drivers as part of a speaker, I always test along the intended listening axis - normally the mid/tweeter axis.)

I don't know whether you've looked a lot at some of the home-testing software, but REW and HolmImpulse seem pretty good to me and fairly straightforward to learn (although understanding what everything means is more challenging for some of us). Maybe you'd want to specify particular settings to use too (again in the intertest of comparing like with like).

Stephen
Hi Stephen,
Thanks for your post.

I take your points you raise in your first paragraph. Those few makers involved in making Full-Range, I hope would take note of end users. Again its very complex. At the high end of the Full-Range driver market, Lowther and Feasterex don't normally publish frequency and impedance data. Some makers add qualifications like: Phy publish some data (no off axis) and state the following:

"These values are indicative only and are representative of typical measurements in our anechoic chamber. One should always keep in mind that the naturalness of reproduction, balance and musicality cannot be described by those measurements and only partly derive from them."

Then there's a whole bunch in-between (Fostex, TB, Markaudio, Supervox) et al. who supply data more typical of that found in mainstream commercial applications.

I could be the first from the Full-Range makers side to say "apologies" fellas, as we (makers) don't operate a consensus on testing methods and data presentation. I can appreciate the end-user scepticism that has arisen on maker's driver data given this situation. I hope for cool heads on all sides, that a mature sensible constructive engagement can be done.

Re gated, agreed it can be useful, especially in rooms that a highly reflective. Bear in mind it isn't a complete cure as frequencies above 300Hz could be a mix of both driver output variance (Full-Rangers are rarely flat response) and room influence. I still recommend guys to do some room damping. Large sound absorbing foam sheets placed around the test driver/loudspeaker would help.

Here's one enterprising young fella with a few practical sound deadening ideas. Might be worth a look:
http://www.youtube.com/watch?v=bPK264QToYU

Similar home-based sound proofing videos of interest:
http://www.youtube.com/watch?v=cp56A6TcL1E
http://www.youtube.com/watch?v=I3q5kQjMAbc
http://www.youtube.com/watch?v=qFeM2uWuMZI

Agreed there's some very interesting and useful home test gear available today. I'm hoping some time later in this thread, a few members can get together to make a "recommenders test gear list".

Cheers
Mark.

Last edited by markaudio; 4th March 2014 at 01:39 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
The future for driver data Electron Markaudio 15 3rd March 2010 09:34 PM
New testing data ucla88 Multi-Way 7 5th June 2008 12:58 AM
The future of Internet speed and data transfer Netlist Everything Else 3 3rd January 2007 12:42 AM
graphical data analysis software bocka Everything Else 3 19th July 2005 04:56 PM
Software Analysis Testing RobPhill33 Everything Else 9 24th February 2003 09:37 AM


New To Site? Need Help?

All times are GMT. The time now is 10:55 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2