Tuesday, 23 August 2016

Less is More?

Is the more accurate model always better?

- Once more a "techy" post -

dBsea can be quite resource-intensive to run, especially if your model demands calculation of many sources and spatially detailed outputs.
Today I'll make a case for always running simple simulations until all parameters are set as wanted, or you are ready to run your model through the night. Even though dBSea is heavily optimised with respect to resource consumption, it is easy to make a model that demands huge amounts of calculations.
A comparison of calculation time versus accuracy of output seems to be the logical way forward.

Figure 1. A comparison of a coarse solve VS a detailed one, 5 minutes VS 3 hours. Top panes are the maximum levels projected to the surface, the middle shows exclusion zones and the bottom shows the 3D version with the calculation grid superimposed.
While the above solves are by no means identical, but sound levels are within circa 10 dB (see the spectrograms in the top panes), and exclusion zones are within a factor of 2. While this might seem like big differences, when you are designing you model, and want a bit off feedback on the changes you've made, a rapid update of the results is very desirable.

Another example is the below example (Figure 2), where two ray trace solves are very close to identical. Scenario "A" took over 30 min to compute, with thousands of outgoing rays from an approximated line source. Scenario "B" was very quick to solve with no calculation of attenuation, and only 20 outgoing rays. Also no frequencies over 1 kHz were included as attenuation means these frequencies do not propagate as far.
Figure 2. Comparison of a high accuracy solve VS a quick solve. dB-levels are only for comparison.
For quick evaluation and feedback during scenario design, simplifying your solves will radically improve this phase, making sure you only need to do the "big" solve once.

Thanks for reading, please don't hesitate to comment.

Friday, 1 July 2016

We get stuck

Today I had to resort to primitive methods for retrieving our Soundtrap logging device.

We had deployed it in Carlingford Lough in Ireland, not as much to measure anything specific as to get some insight into the soundscape of our local patch. The mouth of the Lough is quite busy, with big boats passing regularly, along with the occasional fisher boat and jet ski from the friendly local life guards (RNLI).

The hydrophone was attached to an anchor but with a separate line to the buoy to avoid unnecessary noise from the mooring. It's sitting close to the shipping channel, ready to record any passing ships. I should mention that we might get dolphins or porpoises as well, as this Soundtrap supports super high sampling rate, 576,000 samples per second!

The Soundtrap is deployed, safely hovering a couple of metres above the seabed. Yes, the sticker is upside down, I'll correct that later.

Long story short, our anchor was too good, and we had to send a diver down to retrieve both hydrophone and anchor. This meant waiting for slack water, which occurs when the tide is not moving water in or out of the Lough. Because of 4-5 metres difference in high and low tide, we could not send a diver down during the tidal race!

In the end we got our kit back.
It's always good to be reminded of the forces of nature - and strong anchors!

The sublime view from Cranfield Bay over the Irish Sea and the Cooley Peninsula. 

Thanks for reading!

Tuesday, 21 June 2016

If you'd just like a little taste

dBSea Basic - simple, but free

We've decided to do a free version of dBSea, so you can now use all the nifty visualisations and get to grip with how the software feels.The basic version only uses simple logarithmic spreading loss to estimate sound levels, but for illustration purposes this works nicely. Below i have three examples.
The first picture illustrates nicely the principle of spreading loss differences between a line- and a point-source. in the middle is a rough estimation of a tug-boat (included in dBSea-basic) sailing around the Isle of Man. The last illustration is one of pile-driving in the Moray Firth in scotland. 

Basic means basic. All of the calculations could be done in a spreadsheet or on your favourite calculator, dBSea-basic gives you that, along with our graphics rendering platform, and integrated species weighting functions.

I give a brief introduction in the video below:

Have fun, and spread the word.
Thanks for reading!

Thursday, 16 June 2016

New update of dBSea - while we wait for 2.0

A smarter, more streamlined dBSea

dSBea 1.3.10 has hit our webpage. While not much has happened on the graphical output side, quite a bit of work has been done on the inner workings of the solvers, along with some productivity tweaks.

So if you've never seen dBSea before, this post might not be for you, but by all means, feel free to read on anyway! (and download a copy)

The Parabolic equation solver has got an overhaul, and now uses "Greene's Approximation" for first steps, unless the user decides to use any number of Padé terms. These two methods allow the Parabolic Equation solver to have a much broader sound emission angle from the source.
While this sounds very technical, it means that sound will now propagate in a more realistic fashion from the source - this is especially important for accurately predicting more complex scenarios. We have also improved on the energy coupling algorithm to facilitate even steeper propagation angles.

To make the solving even faster solvers now stop when they reach land (unless you choose otherwise)

We have included a slider for you to select the number of CPU-cores you want to use, so you can keep some of your processors for other tasks.

dBSeaModes modal solver is now even faster!

We have changed the way the total time of the scenario is set, so it is now harder to misunderstand exposure levels due to differences in exposure time.

For quick checks of your solution, we've added 20log and 10log transmission loss lines to the cross-section plot. This plot allows a radial view of all the calculated transmission losses and is a very useful tool for investigating results.

Thanks for reading!

We are working hard on a dBSea 2.0, and rumour has it that it will feature full 3D-solvers!

(By the way, this post was all done with examples from dBSea running in linux (wine usp10.dll set to native, then built-in))

Monday, 13 June 2016

News from our R&D division

We are making dBSea even better

Our oceans are complex three-dimensional environments, and it can be daunting for land-dwellers such as our-selves to grasp a world where going over or under objects is as easy as going around them. 
However, sound propagation very much takes place in all three dimensions.

So while dBSea did take into account that sound can pass through sediment and even small islands, sound has not yet moved horizontally around corners. We are incorporating this part of sound propagation into dBSea right now, and I for one, is quite exited about the results!
Above are two scenarios from Nantucket Bay/Sound, solved with 3D-solving and 2D-solving respectively. I find it especially amazing to see the calculated sound interference patterns in the space behind land that casts an acoustic shadow.
We are currently working to optimise and validate the upgraded solvers, and are looking forward to bringing them to the market when we are ready.

Thanks for reading!

Monday, 16 May 2016

Wildlife encounter in the Shannon Estuary

A little time for the local wildlife.

I was recently in the Shannon Estuary with Irish Whale and Dolphin Group Consulting, Galway-Mayo Institute of Technology and Irwin Carr Consulting to gather information before undertaking a noise related job. We were treated to a lovely sighting of some of the very animals that we work with the industry to protect. The dolphins here are all bottlenose dolphins, a rather large dolphin, that many people will remember from the classic "Flipper" TV-show  (1964-67).
in the video you can hear Joanne and Simon from IWDG Consulting trying to get pictures of the dorsal fins, to add and compare the individuals we saw with a database on land, and this way better understand the population dynamics of the local population.

It will always be a special experience for us to see these animals were they live. Too often, we cannot see the fauna in the area, because it is small or simply because it lives underwater in a world much harder to access. 

Wednesday, 4 May 2016

Underwater Noise Course - Sharing our knowledge and passion

dBSea Underwater Noise Course at GMIT, April 2016

We had the privilege of running this course with attendees from France, Denmark, United Kingdom, Greece and Ireland. We trust that the participants learned something from the event, but I must say that we learned much as well. It is amazing to have this amount of experience in one place to discuss the issues of the new and evolving field that underwater acoustics is.
Some of the common anthropogenic noise sources in the sea.
For this course we were fortunate enough to fit into a smaller room, making conversation and discussion much easier. Much of the info was accessible on the participants on their own computers and the big monitor Martin is pointing at below could be effectively used for demonstrations and sharing of model results.
Martin points to a very important point on the results colour grid.
Even the maths nerd were well catered for as we were led through the ins-and-outs of parabolic equations, normal modes and ray tracing - all these approximate the sound field around one or multiple sources. Because they are all approximations they have strengths and weaknesses, and knowing the theory will assist you in making the correct choices.
Some steps in understanding the way the parabolic equation solver works in dBSea.
Aside from the technical stuff, we had Amanda Robinson (Associate director Marshall Day Acoustics Australia), Martin Keane (dBSea developer), Raymond Alcorn (CEO Exceedence Ltd) and Joanne O'Brien (GMIT) brought in to give us insights into the industry application of modelling software, along with present and historical industry cases. Amanda provided insight into the noise assesment involved with the salvage operation of one of the biggest ship groundings in New Zealand ever. All containers were removed, and the hull of the ship is now an addition to the natural reef below.
Pictures from the RENA salvage operation in NZ. Noise modelling was involved in minimising ecosystem damage during salvage activities.
The biologist in me was very exited about the presentation of data, on the hearing and sonic capabilities of fish and invertebrates. They rely on particle motion as well as pressure for sensing and communicating, making impact assessment all the more challenging. Particle motion is not well understood compared to sound pressure, this is a knowledge gap the community is working to close. Understanding what noises drive the displacement of fish has great economical value and ecological importance.
The sound life of fish (toadfish singing).
My own contribution to the course mostly consisted of summarising some of the main points of underwater noise management, and driving home the point that sound really behaves different underwater compared to the air-borne noise we are so used to (as evident from earlier posts on this blog!).
"Ducting" can cause sound to travel very strangely in water (see more). 
We aimed to present real data in a relevant context, believing that this furthers understanding and motivation better than detached information that has to be contextualised individually later on. This to encourage an all-round learning that we hope will develop a more intuitive understanding of underwater noise. Compare the figure below with this blog post to see why a systemic approach is necessary to understand impacts of noise.
Sample spectra of anthropogenic noise sources compared to the frequency specific attenuation profile. Notice how attenuation steeply increases over 10 kHz (>1 dB/km).

We hope to undertake more of these courses in the future, as we ourselves also benefit hugely from sharing some of our specialised knowledge with the industry, the universities, NGOs and government institutions.

Thanks for joining us for the course and thank for reading.
The dBSea/GMIT team