Acoustic Analysis - Science topic

Hi Melissa Vanderheyden did you manage to find a way to measure ultrasound SPL? Ages ago we used RION NA-20 meter to measure the SPL of a commerical ultrasonic device we tested on dingoes (a link to the paper: https://research-repository.griffith.edu.au/bitstream/handle/10072/18509/42858_1.pdf?sequence=1). eBay has them for about US$250 (e.g. https://www.ebay.com/itm/403041363425). However, we were restricted to using a 20Khz filter so it wasn't the best set of measurements. We also used an Anabat II to measure the charactersitics of the sounds produced (but not SPL from memory). I've just started looking at building a new, higher power ultrasonic deterrent and am keen to hear how you've progressed with you project.

All the best,

Rob

Hi Wieslaw, I get your point, and you are certainly correct to say that propagation, reflections between in the lens, and potentially lateral modes leads to a scenario (simulated by COMSOL) where plane wave propagation in the lens is not a realistic (given the wavelength in the lens). So I will retract my judgment on COMSOL results.

I guess it would be helpful for you to learn more about the theory of the impedance tube measurements and about the absorption coefficient itself - please consult your nearest standard text book on acoustics.

The key difference of the 2 and 4 mic measurement as mentioned by you here is that in the one case you want to qualify a material probe mounted in front of the rigid end of the tube, i.e. assuming that the probe itself will be used that way in a real application (or does not transmit any sound). With thr 4 mic measurement, you assume that there is a (significant) transmission through the material and that this will also be the case in the later application - and in this case, there is sound power that is neither absorbed nor reflected (and thus measured by the 2 mics). To capture this power, you need to apply the additional 2 mics on the "backside" of the sample.

So, please try to understand what you want to measure (and why), as this helps to find the best way to measure the quantity you are really looking for. Just taking some receipt, say, 2 mic method, and evaluating some data gives you some result - not necessary a meaningful one.

This will also help to understand whether a impedance tube measurement will or will not be helpful for your application (i.e. normal incidence vs. arbitrary incidence of the sound wavers, etc.).

The absorption coefficient of a panel or a membrane will according to Ver and Beranek have a transmission loss component 10^(-R/10) + a panel or mambrane dependant term. It is briefly described in this article draft with the reference:

This may be relevant eg for outdoor «stage bubbles», »bubble tennis courts», tents, or simply for light walls with somewhat low sound insulation. At low frequencies the transmission term of the »absorption» may be 0,1 or even higher.

Unfortunately not in my field of expertise. My main expertise is in Acoustics relating to assessment and prediction of environmental and building acoustics and the same for ground borne vibration. You have presumably done an online search on Google for this topic?

Good luck with your research.

Very interesting project.

I stumbled upon

There also are some other interesting findings when doing a google search for "training data animal sounds"

Paul Kinsler Thank you sir for your last reply.

This form will be union.

I have one more question . I want to calculate the transmission loss through a structure having a membrane and cavity and supported by an acrylic plate.

I am simulating using comsol multiphysics.

Over membrane I have chosen solid mechanics. The air cavity is assigned pressure acoustics but what about that backing acrylic plate.

I have to assign it as pressure acoustics or solid mechanics.

You will have to find it yourself, if it is there.

Hi, I would recommend this website for checking emotional stimuli datasets https://airtable.com/shrnVoUZrwu6riP9b/tbljKUnVvikhzaNvF/viwlo7OvlHBG2q88P?blocks=hide

"KAPODI - the searchable database of free emotional stimuli sets."

Best regards,

Diogo Branco

Differences in distress: Variance and production of American Crocodile ( Crocodylus acutus ) distress calls in Belize

Visit to that article

Hi how to work with bellhop acoustic simulator using matlab for underwater acoustic communication

Because the human can hear it.

mechanic parts has an eigenvalue lover then 200Hz (12.000 rpm) and this is normally enough for all design.

Over 3000Hz is only beep sound and it is not easy to excite.

More then 8000Hz is only rush, and for most of the People not perceptible.

Hello,

Will you conduct noise level measurements at the soundwalk stops? If so, you could ask about the perceived loudness using a scoring method and then correlate the objective and the subjective results.

see

1-Fundamentals of noise and vibration analysis for engineers

Cambridge University Press

M P Norton, D G Karczub

Year:

2003

2-Noise and Vibration Analysis: Signal Analysis and Experimental Procedures

Wiley

Anders Brandt

Year:

2011

3-Fundamentals of Noise and Vibration Analysis for Engineers

Cambridge University Press

Norton M.P., Karczub D.G.

Year:

2003

Dear Nitish Katiyar

I suggest "scatterer on substrate" mph file availbe both in COMSOL website and applican builder within the software.

Parinaz Sadri-Moshkenani You would need to know the correct damping in order to make a 'hack' like that, so that you can fit your structural mechanics model. Or switch to COMSOL Multiphysics.

is this a book?

You my see

1-Transform methods in applied mathematics: An introduction

Wiley-Interscience

Peter Lancaster, Kestutis Šalkauskas

Year:

1996

Language:

english

2-Integral transforms in applied mathematics

Cambridge University Press

John Miles

Year:

1971

Language:

english

Let's say I play a voice signal that will from a phone speaker and then capture it with the same phone's mic. If AEC is there, a played signal will be set as a reference signal and it will remove the captured signal. That's what I think.

Hello Sumit Bhowmick

Re the explanation of , see the following reference.

[1] Jim Woodhouse; Chapter 9 - Vibrations: tuning forks and train wheels, squealing brakes and violins; in Christine Bondi (editor); Penguin Books; 1991; pp. 202-222.

Regards,

Tom Cuff

Thank you very much sir. René Christensen

Probably not. Loudness and a parameter describing the efficiency of the singers formant may be relevant. Intonation as well.

Hi

A hand sketch or similar of the setup would help.

Impedance BC only apply for normal angle of incidence. They work on boundaries at a distance from the source, say one wavelength out or so and then, still are reflecting.

As per usual - the advice when tweaking models - start with simulating a Kundt's tube experiment first, as you there can use impedance BCs. Move to 3d from there.

/C

Hi

The rule of thumb is seven nodes per wavelength for free wave propagation and nine nodes per wavelength for reactive fields.

By the sound of it, try to use nine close to the body and less dense away from it.

/C

Towing is not advisable if you can avoid it due to costly ship/boat time, bad weather, and especially flow noise on hydrophones.

There are many commercial battery-powered buoys in wide use that can log underwater acoustic data for months. Check out the AMAR buoys at JASCO Canada for instance, which I think may include marine mammal call recognition software (https://www.jasco.com/amar-g4-specifications).

Thanks but for all that, I need needle hydrophone as it can only give sensitivity in V/kPas.

Is there any way to calibrate transducer without needle hydrophone?

And are they available in Pakistan? As import takes time and I need it urgent.

The dimensions of the thin plate, the configuration of shaker or input device, support and/or constraint of the specimen during testing, the location or distribution of area being measured for signal to test for Ao & So, as well as what transducer(s) or optical technique(s) is/are being used for those measurements, and what circuits and software are being used to condition and process the signal after acquisition.

The are many reasons that may be causing attenuation or lack of effective detection or isolation of the desired or expected signals, including that they actually may not be there at sufficient energy level for effective detection. Are you sure your estimated values for So of the increased-thickness sandwich plate are reasonable and don't actually predict the lower So values? Are you measuring 'lamb' waves on one skin of the sandwich only or of the entire plate (i.e. do you have transducers on both surfaces of the plate with careful and accurate reproduction of phase in both cases, and relative to one and other? Do you isolate or effectively filter frequency or wavelength anywhere in the process of your data collection, conditioning and/or post-acquisition processing? Are the 'lamb' waves on the skin or in the overall plate, what is the ratio of characteristic wavelength to plate thickness and how did this change by your structural alteration?).

Without knowing more of your setup, you experiment and your instrumentation, it is difficult or impossible to offer meaningful guidance.

In terms of a general vibration measurement problem, what signal are you expecting and why? Does the measured value represent something that seems impossible or not, it the result within or nearly within the expected range, and if not, is there something in the method or instrumentation or conditioning/processing that might explain any observed discrepancy? Hope this helps.... -TH

Hi Alina,

I have used many recorders throughout my career and it wasn't until I discovered the Sound Devices Model 722 that I stopped buying other field recorders. It is a two-channel digital recorder (they also make a 4-channel model), made in the US, with a wide, flat frequency response that enables you to record up to the high-frequency limit of your microphone. So if we want to record ultrasound, we use an ultrasonic microphone. If we don't need this capability, we use a standard directional mike that drops off at 18 kHz or so. The microphone signal is digitized at a sampling rate selected by the user via a menu. The digitized signal is stored as a .wav file on the internal compact flash card. When this is filled, you can easily transfer the contents of the CF card to the internal 40 GByte hard drive. This is a small, lightweight recorder that can run on rechargeabale batteries that give a recording time of 6-12 h. Take a look at their website- I think you will be pleased. By the way, I have no connection to this company!

Hope this helps-

Peter

Hi Vineeth,

In the paper

Narins PM (1995) Temperature dependence of auditory function in the frog. In: Advances in Hearing Research (GA Manley, GM Klump, C Köppl, H Fastl, H Oeckinghaus eds.) World Scientific Publishers, Singapore 198-206, the authors present examples of how temperature afects frog call parameters. The following is one example from this paper of how to calculate the effect of a 10 degree C temperature change (the thermal Q10) for a frequency shift of s octaves:

Hi

Please check the following link. I believe it will help you.

https://www.khanacademy.org/science/physics/linear-momentum/momentum-tutorial/a/what-are-two-dimensional-collisions

Regards

Dear Yael,

Wind turbines emit a relatively weak but characteristic noise. The noise is mainly generated by the movement of the blades through the air. This produces a swishing sound in rate with the rotation of the blades, as well as noise from the turbine machinery.

Limits should only be placed on noise over a range of wind speeds up to 12 m/s when measured at 10 m height on the wind farm site, as faster wind speeds will typically mask the noise produced by a wind farm. Separate noise limits should apply for day-time and night-time.

Turbines are audible from distances as great as 2 km only in ideal conditions; when there is very little wind where the listener is, but sufficient wind where the turbines are on a ridge-top to power them.

Peer-reviewed studies have found that living near wind turbines does not pose a risk on human health. The studies looked at a range of health effects from hearing loss, nausea, and sleep disorders to dizziness, blood pressure, tinnitus, and more.

Unlike typical noise sources at licensed sites, wind turbine noise is directly related to wind speed. Wind turbine noise guidelines in Ireland are set out by the DoEHLG guidance and are based on the principle that turbine noise should be controlled with reference to fixed limits when background noise is low, or relative to background noise itself as it increases with wind speed, whichever is the greater. A common interpretation of these limits is that turbine attributable noise should be limited to:

1. 43 dB LA90 or 5 dB above background noise, whichever is the greater at NSL for night-time· hours

2. 45 dB LA90 or 5 dB above background noise , whichever is the greater at NSL for daytime hours 35 to 40 dB LA90 or 5 dB above background, whichever is the greater, at NSL for daytime· hours where background noise is less than 30 dB LA90

In this context, background noise is defined as a function of wind speed, over the relevant period (day or night) which is quantified by measurements prior to the site being built. It is important that such background noise measurements are referenced with wind speed at the hub height of the turbine(s) proposed, as it is this which sets the level of noise generated by the turbine.

World Health Organisation Community Noise Guidelines, 1999 This guidance document recommends an external day-time limit of 55 dB LAeq to prevent serious annoyance during the daytime and evening, or 50 dB LAeq to prevent moderate annoyance, together with a night-time external noise limit of 45 dB LAeq to protect against sleep disturbance.

The World Health Organisation has recently published updated guidance on night-time noise levels designed to protect the public, based on external noise levels as averaged over a whole year xii. This recommends a target value of Lnight,outside of 40 dB where Lnight,outside is the external LAeq over a 1 year period xiii. It should be noted that this guideline limit is intended to cover noise from all sources at a specific location.

Holland Wind turbine noise is restricted to an Lden value of 47 dB and an Lnight value of 41 dB.

Germany No specific wind turbine noise guidance is available so the following generic limits apply. iN genral it is 35 to 70 dB(night) and Day 50-70 dB.

A guidance document on wind turbine noise is currently being drafted but, until such time as this is available, the following generic limits apply: 40-70dB(night) and 50-70 in day for school, hospital, mixed activity, commercial and industrial etc.

Hope content is helpful for you.

Ashish

Hi Anish,

It is interesting that vocalization frequency tends to get lower as one proceeds up an altitudinal gradient. Moreover, dominant frequency also decreases with increasing body size. For many animals, body size also increases as one proceeds up an altitudinal gradient, due to harsher conditions at higher altitudes. Harsher conditions favor larger animals with smaller surface area-to-volume ratios.

In a study (refernce below) we did on hearing in tropical treefrogs found along an altitudinal gradient in Puerto Rico, we concluded:

"We suggest that the animal’s body size determines the frequency particulars of the call apparatus and the inner ear."

I hope this helps,

Peter

CATT acoustic is better for theater simulations in my experience. You may also try to use AcMus developed by a Brazilian team.

Dear Sheng,

welcome,

I think you had already solved the problem. However i would like to introduce a solution. There are two types of electrical vibration harvesters. The electrostaic MEMES converter and the piezoelectric transducer. You can use four flat converters to collect the sound from all sides. You can also put the sound line in in the focus of parabolic reflector and receiver the reflected sound by one flat converters. The second solution may be better. as it uses only one flat converter.

Best wishes

My pleasure Dear Prof Aleksey. Regards

Analyzing a sound's spectral makeup (fundamental frequency plus harmonics) is easy with commercial or free software (I often use SpectraPLUS, which has a 30-day trial). What isn't so easy to access is a sound's intensity or sound pressure level (SPL) unless the recording comes with a calibration file. Inexpensive sound level meters (SLMs) may help when making live measurements, but these SLMs often present sounds as A-weighted only, meaning the measurements filter out very low and high frequencies. Very low level sounds (i.e. very soft sounds) are not easy to measure, as microphone noise alone can obscure the measurement. So, depending on what attributes of a noise or sound you're investigating, publicly available sound snippets may not be sufficient. Eric C. (former Principle Engineer, Doppler Labs; former Chief Engineer, Countryman Associates).

Dear Sumit,

Sound of cracking crystals between two hard surface depends on couple of important metallurgical as well as mechanics factors such as increasing strains, grain distribution reaches log normal packing, coarsening of crystal grains, mis orientation angles, level of temperature distribution due to fraction between two surface with grain misalignment effect , critical grain size micro-cracking and thermal enhancement, cohesion crack nucleation, defects, cavities etc.

In general, If the size of the crystal grains is closed to the wavelength of the supersonic ray, the part will not be transparent to ultrasound, in this case the there will be no ultrasound bottom of signal. it is also seen that size of the crystal grain increases the intensity of supersonic wave decreases.

Sound products can be listen at high extrusion speeds.

Atomistic modeling is highly applicable in easy understanding of crack sound between hard metal.

Ashish

For example, if you have N data in a range of T seconds, then the steplength in frequency domain is 1/T Hz. Keep in mind, FFT will also return N data in frequency domain.

Thus, what you get is exactly [0:(N-1)] * 1/T Hz in frequency domain.

A point force can generate other waves in addition to a Rayleigh wave (which is an interface mode owing to the discontinuity of elastic properties at the aluminum boundary). I think that a simple point force can also generate P-SV waves that radiate down into the aluminum. So in your graph you may be seeing the superposition of Rayleigh waves and P-SV body waves.

Additional waves are suggested by the evanescent waves in Bernard Garnier's answer above, where he is no doubt thinking more generally of many different modes of vibration in stratified media, of which the Rayleigh wave is only one type of mode. But in a homogeneous solid half space, P-SV body waves can be attributed to the contribution of two branch line integrals -- one for P-waves, and one for S-waves -- in the complex wave-number plane for cylindrical wave analysis.

Unlike Rayleigh waves, P-SV body waves should exhibit the usual 2D decay (geometric spreading loss) with travel distance. But in your model Rayleigh waves and P-SV body waves will be superposed on each other.

The Rayleigh wave presumably dominates without decay at long distances and near the surface of the aluminum. Your graph seems to be showing this leveling off of amplitude at greater distance. If your model allows you to go out further in distance, then you should check for leveling off in amplitude at much greater distances. That is where the Rayleigh wave dominates.

But P-SV body waves may dominate at greater depth, where the Rayleigh wave motions do not penetrate owing to the exponential decay with depth. So you might check for body waves by plotting the wave amplitude at depth.

In any case, there may be more wave excitation going on than the excitation of Raleigh waves alone when the forcing is a point force. In order to generate a Rayleigh wave alone, and no other waves, you will require a rather complex distribution of forces and not just a single point force (use Huygens' principle for wavefront reconstruction, applied to elastic waves). To isolate the Rayleigh wave, it is much easier to simply sample the waves far from their source, as distances where the Rayleigh wave dominates all of the other waves generated by the source.

As Fabrizio above has correctly mentioned, you need to consider just the frequency range for human voice in your application, which is much more limited than the hearing frequency range. In addition to the acoustical response of the room you are interested in to model (a train station), an important detail is to consider the background noise as accurate as possible. This will also affect the total required sound power of your PA system in order to get an appropriate intelligibility level.

Best

Unqualified comments :-

Time weighted negotiating of a mass, collection of matter considering field effects ( say gravity, possibly friction and so on) with a directional effort is work, forced displacment sort of effects.

Rate of doign work is power (watts) , sort of time rated expense of energy. Describing power has nothing to do with underlying phenomenon.

Most important thing to understand or find or reconcile is

proxy of or displacments/field

proxy of or efforts,

and of course ratio of expended against recoved power offers clues about nature of losses , mostly thermal, and proverbial efficiency etc.

Hi Kapil

Not excatly what you are asking but perhaps anway useful ?

The below link contains links to excellent material for self study on signal analysis. The writings are old but the material is among the best available. Start from the top and continue downward.

https://qringtech.com/learnmore/recommended-reading/

Sincerely

Claes

i7 core with 16GB RAM minimum.

Please see the link below: https://pdfs.semanticscholar.org/2c7f/30a9d9a516399e07f4210a32eed8595884c6.pdf. Acoustic Analysis of Voice: A Tutorial.

Best regards

I completely agree with the answers from Tim Ziemer and Xenia Kontogianni . We did a similar investigation on the STI in a virtual room. Check out the poster and let me know if you have any questions.

What do you mean for semi-supervised learning?

Reinforcement learning?

I agree that Audacity is of no use-- it's primarily an editing program, not for measurement. Praat and SAP would both probably work for the simple measurements you described. I believe Praat (which was developed for work on speech) has some tools for measuring speech-related features such as formants, which might be useful for nonhuman primate vocalizations. Raven Pro (http://ravensoundsoftware.com/) has (I'm told) an easier learning curve than Praat and includes basic time, frequency, and relative power measurements (though no formant measurements). There's also a set of freely available tutorial videos available (see "Training" on that website).

Can you please clarify by "acoustic properties"?

A metal alloy as every solid propagates compressional waves that are "sound waves" with a sound speed of SQR(Young Modulus/Density) with an hysteretic damping generally close to 1E-3; measuring the Young Modulus is a simpler lab test than measuring the hysteretic damping, both can be made on a simple parallelepipedic sample; sound speed and sound absorption ((1-hysteretic damping)*number of wavelengthes) can be also assessed explicitly with an ultrasound probe on a rod (send a short pulse and compare with the echo).

If you manufacture a musical instrument and want to characterize the influence of the alloy on the musical quality of this instrument, this is an absolutely different issue calling for decades of research (the most popular is trying to understand the role of the varnish versus wood selection in Stradivarius violins): I would just ask a good musician to assess it by playing and ranking various instruments manufactured conventionally and by additive printing...

Maybe "reverberation" might be considered simply as "multipath" for aircraft noise in outdoor environments, and in that case you might solve it explicitly e.g using adaptative beamforming algorithms. Acoustic tracking of low flying helos is a conventional problem solved satisfactorily from many years in open field, I know some systems based on likelihood maximisation, it take a fraction of a second to get iteratively the most likely position, then your tracking algorithm will have to reject aberrant dots. A key issue is now far your aircraft noise and course is stationary or not, as rightly commented by Ronald! Of course if you deal with supersonic low flying jets, you will find the Mach wave direction, not directly the plane heading... [An issue solved in acoustic sniper detection systems by tabulating the speed (hence the Mach cone angle) for a library of ammunitions and associated signatures]

Please have a look at these useful RG links.

couldn't find any standard database but yes i guess this would be helpful https://opendata.stackexchange.com/questions/6771/required-a-audio-format-baby-crying-data-set

These links may help you

Hi, i'm an acoustics professor in Colombia and i'm really interested in the project and in general in Master Acoustics International , could you give me more information.

Best regards

Do you know this paper?

Shigeaki Amano, Tadahisa Kondo, and Sachiyo Kajikawa 2001. Analysis on infant speech with longitudinal recordings The Journal of the Acoustical Society of America 110, 2703 (2001);

https://doi.org/10.1121/1.4777322

Let me add that after applying any program finding just peaks probably some elimination procedure would be desired. Indeed, the example in the red box supplied by Hadjer shows that any automatic program returning all local maxima and minima will display too many points - in the same manner those which significant as the non-significant. Of course, the expected correct answer should eliminate some of them. This depends on the assumed DEFINITION of a significant point of changes. Perhaps it would be enough to implement in the code a suitable criterion like the following:

For a given fixed half-width t>0 do the following:

1. for every point of maximum (x,y) apply:

If [yn <= y whenever x-t <= xn <= x+t], then decide: [ (x,y) is a significant point of local maximum]

2. for every point of minimum (x,y) apply:

If [yn >= y whenever x-t <= xn <= x+t], then decide: [ (x,y) is a significant point of local minimum]

3. Eliminate a point if it is not unique among other points of the same type within distance not exceeding t.

The third step is just to choose one of points in cases when e.g. at two neiboring maximal points say x1 and x2 the values y1 and y2 are almost equal:) Perhaps some average of the similar x-s and y-s would be better? Or, before running the program finding max and min points, to perform some standard smoothing procedure?

Regards, Joachim

Hi, Mathan,

As additional resources for Trevor J Cox book, you may find an interesting detailed explanation in J.F. Allard and N. Atalla’s book Propagation of Sound in Porous Media (Willey, 2009). You may also find another brief explanation in Hoda S. Seddeq’s paper published in Australian Journal of Basic and Applied Sciences, 3(4): 4610-4617, 2009: Factors Influencing Acoustic Performance of Sound Absorptive Materials.  Almost similar topic provided by Tiuc et al. in their paper entitled The analysis of factors that influence the sound absorption coefficient of porous materials which are available at RG.

According to my own experience in working with felt, it is possible to increase the sound absorption of the fibrous material based sound absorber in the mid-frequency range. The simplest way is by using thin elastic membrane backed fibrous layer. When the fibrous layer coated with a thin elastic membrane, it changes the reactance of the structure and brings the possibility for absorbing the sound waves energy through combined resonant and viscous damping mechanism simultaneously. Should you need further discussion, please feel free to send an email privately.

Best regards,

Iwan.

Hi Laura,

I agree with Pavel regarding the channels and amplitude. I also don't use SoundRuler (sorry!) but I thought it might be useful to add that you need to be sure there is no background noise overlapping your calls of interest. If for example these recordings were made at a zoo, there may be visitors chatting, or in the field there could be other animals calling etc. on your recordings. If there is, then you can either need to filter it (if it does not overlap in frequency with the monkey calls), or if that's not possible, you could manually extract the frequency measures, or simply eliminate those calls from your analyses. It might be that you can use the read-out on all your "clean" calls (i.e. no background noise, one individual calling at a time) and in that case it could be just as simple as you say!

Good luck with your interesting project!

All the best,

Esther

You must try Long Term Average Spectrum

A group of parameters are important for that purpose, for example:

Pitch, Harmonics, First five formants at list, Duration, Melody, Intonation, Sonority, Transitions of phonemes, etc.

Best regards,

D. Escobedo

Dear Mad Helmi Bin Ab. Majid,

It matters which type of signal do you measure (i.e. stationary, nonstationary) which is implicated by the source of the signal. Is it a biological/medical, artificially generated periodical signal or other?

I guess, the question refers to acoustics, so the easiest and the most accurate way to measure the SNR is to do it directly by device performing the acquisition.  However, this method is frequently not available or results not transfered to data sheet.

You can use some easy approximation: divide mean value of the signal by a standard deviation of the signal measured within certain time window. If the recording device does not provide any additional pink noise or the only noise present is the white noise - the task is easy. According to theory, the expected value of white noise refers to mean value of normal distribution, and autocorrelation function should obtain value 0 (or close to it in practice). Then, the approximation above is sufficient for SNR.

You can also perform  a similar action in frequency domain, which requires firstly to calculate an amplitude spectrum (or power spectrum) of the signal in certain small/big time window (or for the whole measurement at once). Though, you need to be aware of two effects:

A) the wider time window - the better resolution of the spectrum, but the longer time of acquisition and worse time resolution of the results;

B) You need to know, which part of the spectrum your noise accupies, especially if the noise is not white.

The easiest way - you can divide the amplitude (or sum of the amplitudes present in the certain part of the spectrum) referring to your 'usable' signal, by the same referring to part occupied by noise.

For power spectrum SNR = (average signal power)/ average noise power), which in dB refers to (SNR_dB=10 log₁₀(SNR).

Please, search some information about the type of the signal you use and the possible noise influencing your results. It might be, that in the spectrum you will see a strong influence of a low freaquency or only high frequency noise (if there is no additional source of low frequency noise added).

This is the fast, simple answer. If you need more help, please don't hesitate to write.

PS> I hope the publication attached would be helpful:

Regards,

Richard Gomolka

Hi

Here is another opensource software, Code Tympan, that uses raytracing. It might not yet be exactly what you want but may still be worth your taking a look at as anew solver is in the works.

http://code-tympan.org/index.php 

It seems that Code Tympan currently is slower than SoundPlan. 

Sincerely

Claes

I did some experiments on noisy speaker recognition using speech enhancement with non-negative matrix factorization (check FASST toolkit)

The approach works well, but only if you apply it for both noisy train and test data. To my knowledge, any speech separation algorithm leaves some residual noise, and consequently, there is still a mismatch between clean speech and the enhanced one. If your training data are clen, I'd suggest to apply artificial noising techniques.

It might also depend on your classifier. For me the best was obtained with i-vector extractor and plda classifier trained on noisy data enhanced with NMF approach

Hi

For muffler design, all dimensions matter, though some decide the base performance more than other. These tend to decide the system fundamental resonances. 

I believe it is worth your while taking a look here

and to take a look at what Prof. Åbom & Prof Boden have been doing for quite some time. 

For design, perhaps some of my ramblings can be of interest though they only adress Test, Simulation, QA and Optimization in general rather than muffler design in particular?

For what it is worth. It is worthwhile to simulate muffler not only regarding acoustics as they should be able to withstand also fatigue from pressure pulsation loading, and if the muffler is extended (large with respect to system dimensions), you must identify also good suspension positions, mechanical resonance should not lie at bad motor RPMs or at typical cruise speed, fit within a limited space envelope, etc. 

Sincerely

Claes 

Best way I know of is what Sina said with the CPSD or the coherence. To do this with MATLAB's built in functions, use the "cpsd" for the cross-power spectral density or mscohere function for the coherence.

If they're noisy signals as you say, you'll may want to use a Hanning window with some overlap for the window edge effects. We usually use 75% overlap for aeroacoustic broadband signals.

The Bird Audio Detection project may be of interest, although the aim is detection rather than identification. The dataset has 17000 recordings tagged as either containing or not containing birds. The survey paper reviews the most common methodologies.

Thank you so much!!!

If :played back "means sounding all at once in the same place, your question devolves mathematically into asking how many numbers can sum to, say, 3.  Mathematicians are inclined to invoke such ideas as infinity in such cases, though the answer is clearly larger than we can name.

If you at what point does adding yet another sound to a large mix become irrelevant, the answer is also clearly large, but far from infinite.  Having sung in large choruses, I opine that after about a hundred or two voice, the addition (or subtraction) one more (or less) hardly matters in any musical sense.  (The gospel choir at UCSD once numbered over 500, but their major problem was finding a space in which they could rehearse).

The interesting part of your question lies in between 0 and infinity.  Modern psycholofy and cognitive science both suggest that  7 plus or minus 2 is about the limit of our "attention".  But they also suggest we don't multitask, but scan, i.e., shift our focused attention rapidly among competing messages.  Any instrument-rated pilot can tell you that simultaneously scanning (meaning observing and understanding) 7 or so instruments (as well as looking out the window) stretches human capabilities.  Not impossible, but difficult, especially if your life depends on it.

So even though I can't truly answer your question, my best guess would be the well-established psychological number associated with short-term memory: seven plus or minus two.  Seems our brain - evolved over million of years - have figured that before that number, we stand a chance to figure things out.  After that number, run!

Hi Anvesh, you can contact the GTAC service to take your questions. It is an excellent tool (http://www.plm.automation.siemens.com/en_in/support/gtac/index.shtml)

Please start  GOOGLE SCHOLAR    and not GOOGLE .

give your question in the mask (e.g.What is the Acoustic Impedance of pentacene ?)

and you will obtain  991 hits. Read them and extract the infos related to your research.

Good luck

JRG

I use the method described by Claused et al. for fitting.

[1]    Clauset, A., Shalizi, C., Newman, M., 2009. Power-law distributions in empirical data. SIAM Rev. 51 (4), 661–703.

There are practical Matlab functions and  Python packages to do the job: