Feedback to the Presenters of

ALMA Fall Symposium, October 15, 2005

Metropolitan Yiking Hotel

Guang Dong, China

The first session was presented by the Vice President of ALMA and the topic was "Closing the Engineering Gap Between Asian Suppliers and Western Customers". It was not technical. Although the message was simple and does not carry much weight in terms of "essence" or "behind the word's meaning". The translation provided by the girl was terribly out of this world. Only 30% or so was properly translated where the others were completely wrong.

For example: -

Presenter = Often problems arise because the customer and supplier did not agree on a specification and when deadlines are not meet, the supplier should make sure to inform the customers as early as possible, instead of not making the deadline and not informing the customer at all.

Chinese translator (but written in English for the presenters) = Often the supplier does not agree with customer regarding specification and hence deadline is not meet. The supplier already inform the customer but nothing could be done.

Notice the lost of the meaning "problem arise" not mentioned by the translator, instead turned into "the problem"

Notice "the supplier should inform" turned into "already inform"

Notice "inform as soon as possible" turned into "nothing could be done"

There are countless mistakes, hence I will end here for now.

The second session  was regarding measurements setup importance. Perhaps due to lack of time the presenter did not managed to stress the importance of the followings: -

1) Near-field

He did not specify what is the distance required for near-field. Although he mentioned where direct source is much louder than reflections or ambient so that they cannot influence the results.

An example of the calculation of sound pressure addition should have been in place to stress the application of theory to practical uses. For example see -

2) Gating via Windowing of Time Domain

He did not stress the fundamental theory of impulse response. Although he mentioned using proper window like Blackmen, but without explaining the fundamentals, the user will follow his advice and still get things wrong.

Example - mid-range driver capable of 100Hz would have an impulse length of t(seconds)=1/f(Hz) becomes 0.01 seconds = 10 ms, but his example was cutting it just before 1st reflection, which was 5ms.

When the trailing impulses are cut and FFT performed using a rectangular window, the results would be wrong. This kind of explanation should be "stressed" as in MLSSA Window Technique.

I also didn't see a small gap in the front, i.e. if mic to speaker distance was 1 meter, than there should be around 2.8 ms of delay. But there wasn't or I didn't noticed, hence he was probably using a near-field measured impulse response as example.

This is a mistake because it did not show the 1st reflection and subsequent reflection properly. This slight error in presentation will ultimately mid-lead the lack of acoustics fundamental theories public. Again my example show a much accurate representation of good anechoic chamber without requiring gating versus measurement in office which require gating as shown in MLSSA Window Technique.

Come-on, we can tell a 1st reflection strike back distinctively. The impulse response used by the presenter did not have a "distinctive" strike back. The level comparison with the others was so little, it must have been a near-field measurement.

There are a few more, but I lost the paper where I wrote down the points to be stressed, hence I will end for session 2.

The third session was by Peter Larson of LOUDSOFT.

I wish Peter would scare the audience a bit of the true fundamentals of Finite Element, else the audience or users will not appreciate the capabilities. Without the scaring they will never realize how hard it is to make such mathematical algorithm useful to everyday engineering. Adding this topic would only take between 5 to 10 minutes. Like something like this - Finite Element Analysis on Loudspeaker

I had the opportunity to talk to Peter and asked about mesh seizes in relation to different wavelengths sizes and he quickly answers the software automatically adjust the mesh sizes. From this I am pleased to know and suspect the mesh size are split no less than 4 to at least simulation a triangle wave. I will talk to Peter more later.

Peter also explains the software automatically increase stiffness at areas near glue like dustcap contact to cone or voice coil, cone to surround etc. This again is very good and automatically address the real physics without requiring the user to input. Bad if the user did not know the software automatically adjust and added extra properties, which I would.

The verbal translation by Eric was also terrible, unfortunately. I know Eric personally and hope he does not get mad at me. However Eric is a truly qualified engineer, where he as proven time to time to me in many technical aspect through his support to Philips Sound Solution in Shen Zhen via email.

Eric's written translation, is however, extremely commendable. I will give near 100% score to his written translation.

The fourth session was by Dr.- Ing Stefen Irrgang of Klippel GmbH

Klippel is a new generation measurement system and covers the field of "non-linear" behavior in extremely detail attitude and since the loudspeaker are often played in the non-linear range hence this system and it's theory are extremely valuable to learn. Their website contains countless documents for download. They are a much read. Read it 10 times if you need to.

Since it's subject is very deep, hence I don't have anything to say, except to learn from it. Please visit their website and listen to the distortion demonstration. It is a valuable learning experience.

The verbal translation was provided by Eric, and since this subject requires more technical terms, hence the verbal translation was worst than LOUDSOFT session.

Sorry Eric, but someone has got to give feedback for the benefit of our future generations.

Dr Irrgang was very quick to ask what can we do about it?

I explain the terrible verbal translation is mostly of to lack of time and the pressure faced by the translator on a LIVE event. The solution would be to define a script and write it down in English and pass it to Eric to translate during his free time without pressure of time. Then during presentation they need to synchronize. It would be perfect as Eric's written translation is near 100%. This means he knows the subject extremely well, just that he does not have enough time to react during a LIVE conversation.

The fifth session was by Steve F. Tmme of Listen Inc

Steve stress on harmonic distortion until the 10th and examples of which harmonic will be significant at how many percentage or dB. It was an mind opening presentation. Coupled with Klippel's non-linear distortion demonstration. I finally have a starting insight to harmonic distortion.

Steve's verbal translator was worst than Eric by allot allot. Eric's verbal translation was nearly going to pass, but still failed by a little. But Steve's verbal translator was extremely weak in the fields of statistical mathematics. Hence nothing was properly translated in respect to technical and mathematical issues.

Regarding the translation problem. Why do I stress so much instead of other things. It is because the "essence" of the subject is completely lost. All of us are unfortunate that acoustics or sound cannot be seen.

Although we could hear but we cannot remember. Unlike in a "visual domain" we could really remember a face, scenery etc.  But we have much trouble in remembering sound..... example Flat Frequency Response

I consider Acoustics the hardest subject when compared to Electronics, Electrical, Computer Hardware, Computer Software, Mechanical Engineering Hardware. Because for all others we can see, we can touch, we can calculate and they never go wrong.

In acoustics we cannot see, we cannot touch, and it always go wrong (i.e. there are 99.999% of bad loudspeakers in this world and less than 0.00001% good loudspeaker) call me a perfectionist but it is true.

The other other engineering as difficult as acoustics are aerodynamic. Fluid dynamics are hard but their application are limited to huge industries, which does not affect the earth's population.

Aerodynamics only affect the military and flight companies and lastly Formula 1.

Acoustics is the only everyday human application, with the least human studying it. The essence are not known.

I shall end here and hope this feedback is useful to the presenters as well as other English users.