[Review] Getting animated

[Tenson]

Pete, those are pretty interesting! If you have bigger pictures you can upload them to imageshack you know. http://imageshack.us/

I can see most difference between ~100Hz and 1KHz, with the 'blue light off' photo showing lower amplitude in that range by just a smidgen (that's a technical word).

There are a few things that I notice. First is that even on the first photo with the 'blue light on' where the amplitude is supposed to be higher, there are some peak hold 'dots' that show the level had been higher previously. This might suggest slight instability in the measurement conditions. If you look at around 1K25 you can see a spike that is clearly higher in the 'off' photo, and there are a few of these in the picture.

How did you deactivate the blue light? Was it right next to you or did you have to move a bit across the room to get to it? I am wondering if you moving in the room affected the response reading. Like if you passed between the speaker and the microphones direct line of sight it could have the effect displayed.

There are a few other things I can think of that might cause this – The randomness of pink noise averaging out the peaks over a longer time. External noise changes from traffic, wife in the kitchen and so on. The microphone position changing slightly; how was it secured? My mic stand can droop just a little over time (don't giggle!).

I'm not trying to discredit the work you have put in. I'm just trying to eliminate possibilities. For me, a difference of < 0.5dB is rather too small to have much faith in when the stability measurement conditions is an unknown. A 0.5dB broad range increase in SPL is also not enough to make people say it sounds louder with the unit on. Especially after 5-10mins break of not listening.

Just my observations.

Oh and your camera works better with the blue light on!

Could we see all 6 comparisons?

 

[zanash]

The unit was next to me ...just a question of unpluging the lead.

the mic was 2m away and behind me ....

...I remained seated on the floor, in the same position through out, and had the numb bottom to go with it.

yes I spotted that ....can't account for it unless it was ambient or I created it ...but it had to be more than a transient other wise the averaging would have got rid ....

we live on a close with no through traffic ...
everything in the house was either off or in stand by ...
My wife is away ...otherwise I would not be able to play !
the mic was firmly fixed ...absolutely no movement its a uni loc photographic tripod.

I agree this just shows something happened on the 8 repeat cycles I did....the results were just about the same over a period of two hours....it was always the same freq's that showed a change and by roughly the same ammount.



its an average of 0.5db over the whole range ....some freq...had plus 3db many had plus 2db.

And oddly I'd spotted that too ..both pics were taken from the same position with the same settings [one slightly higher by a couple of centimeters but the same distance away ] I could have twitched

I'll see if I can refine things.....

 

[dean.l]

interesting results, I just can't see why the manufacturer would have not included this information in its advertising if it is that easy to measure ?

it seems the unit must do something, I think it deserves a few tests by the manufacturer to add some weight behind its claims, to stop the doubters, also has the unit been patented? I think there is a website somewhere that has all the patents. if not I would rip it apart and copy it.

could you not get the 1/2 a db just by turning the volume knob to the right a bit ?

 

[zanash]

I'm still working on this ...but this is not all its doing ...this is the first thing I've got a measure with hard numbers on ....



I don't feel like destroying it just yet !

 

[dean.l]

I was only joking

thanks for taking the time to try to explain the thing.
interesting subject.
I just want to know how it works

 

[Tenson]

Pete, I know you can't very well come all the way down here (and I don't really have the time ayway!) but if you want to post the unit and let me do some measurements then by all means do.

 

[ShinOBIWAN]

Pete you really need to forget about trying to gleam information from the DEQ. What would be much better is ARTA and run a whole suite of test from FR to CSD, to impulse.

We're dealing with rather small changes here so the results would need to be performed multiple times and averaged. ARTA will gladly average upto 9 data sets so this is a piece of piss to do.

 

[DavidF]

has the unit been patented?

good queston.

has it been?

 

[zanash]

Here is some more data....The data was recovered in the same way as before the settings of the deq and mic position were as close as possible to the previous days pics.

So having left the blue light on all night using pink noise generated by the deq , measurements were noted with the unit on, off after 5mins 10 mins 3hrs...[ie total dissipation of effect ?] the unit was removed from the room ..everything else remained the same.

theres no easy way to do this so its a list [all db are minus figures]

freq .. .. blue on .. .. off after 3hrs.... change in db
20 ..... 67.5..... 68.1 ..... -0.6
22.5 ..... 64.8 .... 66.7..... -2.3
25 ..... 63.7 .... 66.2 .... -2.5
28 .... 62.5 .... 64 .... -1.5
31.5 .... 62.1 .... 63.2 .... -1.1
35.5 .... 60.2 .... 61.7 .... -1.5
40 .... 59.8 .... 61.3 ... -1.5
45 .... 55.2 .... 58.2 .... -3
50 ..... 53.1 .... 55.6 .... -2.6
56 .... 52.7 .... 54.7 .... -2
63 .... 54.8 .... 55.4 ... -0.6
71 .... 60.6 .... 60.1 .... +0.5*
80 .... 64.3 ..... 65.3 .... -1
89 .... 64.8 .... 66.3 .... -1.7
100 .... 62.5 .... 63.5 .... -1
112 .... 68.9 .... 68.5 .... +0.4*
125 .... 61.9 .... 62.9 .... -1
141 .... 63.1 .... 65.6 .... -2.7
160 .... 67.8 ..... 70.3 .... -2.5
180 .... 67.2 .... 68.2 .... -1
200 .... 67.3 .... 67.3 .... 0
225 ... 64.1 .... 64.1 ... 0
250 .... 63.4 .... 65 ... -1.1
280 .... 58 .... 60.4 .... -2.4
315 .... 65.3 .... 66.4 .... -1.1
355 .... 73.9 ..... 75.9 ... - 2
400 .... 75.1 .... 76.6 .... -1.5
450 ..... 66.1 .... 66.6 ... -0.4
500 .... 71.8 .... 72.4 .... -1.6
560 .... 71.3 .... 73.2 .... -0.9
630 .... 71.4 .... 71.0 .... +4*
710 ..... 69.9 .... 69.9 .... 0
800 ..... 68.5 ..... 69.5 ..... -1
900 .... 67.8 .... 69.7 .... -1.9
1000 ..... 72.1 .... 72.4 .... -0.3
1125 .... 73.1 .... 72.7 .... +0.4*
1250 .... 71.9 .... 72.5 ... -0.6
1420 .... 71.9 .... 72.5 .... 0.4
1600 .... 71.9 .... 71.5 .... +0.4*
1800 .... 70.0 .... 70.0 .... 0
2000 .... 67.2 .... 68.1 .... -0.9
2250 .... 71.3 .... 71.3 .... 0
2500 ..... 66.7 .... 66.7 .... 0
2820 .... 67.8 ..... 67.8.... 0
3150 .... 70.6 .... 70.1 .... +0.5*
3550 .... 66.6 .... 68.3 .... -1.7
4000 .... 63.8 .... 65.2 .... -1.4
4500 .... 69.0 .... 69.0 .... 0
5000 ..... 71.4 .... 71.4 .... 0
5650 ..... 71.9 .... 72.4 ... -0.5
6300 ..... 69.8 ..... 70.4 .... -0.6
7100 .... 68.3..... 67.8 ..... +0.5
8000 .... 67.4 .... 67.9 .... -0.5
9000 .... 67.1 .... 67.1 .... 0
10000 .... 67.7 .... 67.7 .... 0
11250 .... 68.7 .... 68.7 .... 0
12500 ..... 68.7 .... 68.7 .... 0
14200 ..... 69.7 .... 69.7 .... 0
16000........................................
18000........................................
20000.........................................

* denotes a rise in db

all other freqs had zero change or decrease in volume.

I'm not going to draw any infferences from this till I've had a study .....as its taken the best part of five hours to get these results !

 

[DavidF]

is ARTA and run a whole suite of test from FR to CSD, to impulse.

QUOTE]


what do these measure (or measuremnet of?)?

 

[zanash]

shinobiwan .....dead right

but as thats all I've got !

 

[zanash]

Just a few observations .....this confirms that the perceived rise in volume is shown ......

can't get my head round why certain freqs have risen ...[deq error ?]

I've just plotted a graph of db change against freq ....it doesnot look random as you can follow rises and fall in db across the graph ... if the were random they should be all over the place !

 

[ShinOBIWAN]

is ARTA and run a whole suite of test from FR to CSD, to impulse.

what do these measure (or measuremnet of?)?

FR is amplitude/frequency response showing relative sound pressure level, CSD is cumulative spectral decay showing per frequency time based and resonance behaviour and impulse is impulse response measurements used to show overall time based behviour, driver time alignment, rise time(speed of initital excitation) and equalisation time(time it takes for the sound pressure to equalise after initial excitation)

The impulse response might show some marked differences if this unit is claiming to create a more coherent volume of air for sound to pass through.

CSD might also be useful because it will show if its improving the rooms natural decay and/or resonance behaviour.

 

[ShinOBIWAN]

can't get my head round why certain freqs have risen ...[deq error ?]

That was my initial thoughts because most of the sound pressure difference between the unit being or off are well within the margin of error for the equipment your using.

Lots of things influence measurements and you really need to be especially vigilent when doing these in-room. Ambient noise level is the biggest problem then consistant mic positioning and be sure to ignore low frequency information.

You really need to treat the experiment with the upmost respect and approach it as a scientific experiment rather than a curiosity if you wish the data to have some level of value. This means better measuring equipment and a good knowledge of measurement techniques and how to interprete them within the constraints in which they were taken.

What you'd need to do is perform many measurements of the unit off and on then average each data set. Using Tone bursts such as MLS and swept sine each have their own benefits and downsides to combatting ambient noise levels vs accuracy vs measurement time.

There's no one correct way to do this and the whole subjectivist vs. objectivist debate runs strong also. This brings into question the whole validity of any measured data set and nearly always causes a contentious nature to the discussion when trying to see the tree's rather than the forest.

 

[zanash]

Ok ...as were only looking at change error in the initial, intervening and final readings should show some variation created by the error ....[I bet tenson can tell you what the error is on the unit] . By the way what is the error on the deq ?

I've seven sets of data ....for each freq taken over a period of 5 hours
one example ie

9000 hz....67.1 .... 67.1....67.1.....67.1.....67.1....67.1.....67.1....

shows no drift ..but may display the same read error for each...
but as were only looking for change not an absolute db value [the error can be ignored if there is one] other readings do vary but thats the whole point of the excersise.

please read my previous post ..tenson asked the same questions.

Why do you feel the data has no value?
lets face it its the only data you have !

This was not meant as any form of scientific experiment just a look to see if anything showed up ....

you know you just can't win .....people ask you to try things then you get slapped down for doing just that !

All I was trying to do was to see if an effects could be measured.... and thats all I've provided.

 

[ShinOBIWAN]

Ok ...as were only looking at change error in the initial, intervening and final readings should show some variation created by the error ....[I bet tenson can tell you what the error is on the unit] . By the way what is the error on the deq ?

The error on the DEQ ADC for the mic input will be tiny (hundreths of a percent). Where errors creep in is the mic and by far ambient noise level and measuring setup and techniques.

SNR is king when taking measurements.

Why do you feel the data has no value?
lets face it its the only data you have !

This was not meant as any form of scientific experiment just a look to see if anything showed up ....

you know you just can't win .....people ask you to try things then you get slapped down for doing just that !

All I was trying to do was to see if an effects could be measured.... and thats all I've provided.

Why do I feel the data will have little value as it stands? - anecdotal evidence gleamed from hours discussing, watching discussions and investigating such things. You've obviously become jaded by this whole episode and you've construed my post as something it isn't.

Its also quite clear to me your not providing nearly enough data nor using the right software, techniques and hardware to investigate such small changes accurately. Based on that I don't believe you've done anything to 'win' this debate.

If you want something more thorough done to measure the effects of this unit then drop by me in Chesterfield or bring the thing to the DIY meet and have it done there.

 

[DavidF]

FR is amplitude/frequency response showing relative sound pressure level, CSD is cumulative spectral decay showing per frequency time based and resonance behaviour and impulse is impulse response measurements used to show overall time based behviour, driver time alignment, rise time(speed of initital excitation) and equalisation time(time it takes for the sound pressure to equalise after initial excitation)

The impulse response might show some marked differences if this unit is claiming to create a more coherent volume of air for sound to pass through.

CSD might also be useful because it will show if its improving the rooms natural decay and/or resonance behaviour.

thanks.....sounds interesting!

 

[DavidF]

II don't believe you've done anything to 'win' this debate.

I dont know.......hasn't pete made a start here?

If you want something more thorough done to measure the effects of this unit then drop by me in Chesterfield or bring the thing to the DIY meet and have it done there.

Well there you are , Pete, an offer you can't refuse??

 

[zanash]

I've already filled the form in .......!!!!

but yes would be good to meet anyway...yes would love to pop up with it pm me when would be convenient.


I agree ....but its all I 've got so you use what you have .....otherwise you would not have these numbers to look at.

ambient noise is immaterial if constant ......if its transient the unit will average it out [but it will overly the readings]

Win the debate......?? thats done ....ask the 31 people who have heard it. The only debate is between those that have heard it and those that have not ! Theres no debate between those with first hand experience [well on the fact they can hear a change, how it works though is something else].

 

[Tenson]

To be honest, the fact the measurements for all 7 of the 9KHz readings are exactly the same to within 0.1dB makes me even more dubious about the accuracy of the results.

BUT, taking it to the DIY meet is a brilliant idea! Though, I wonder how it is planned to get meaningful measurements somewhere where there will be music all the time and lots of talking.., Ant?