Any Linux programmers on here?

[HughF]

Doesn't matter what distro you use. I use mint (GTK based). If you want to program in straight C, I'd go with GTK as it looks simpler than QT (C++ only from what I gather) from my initial research. The project I was working on, our customer wrote the graphical front end in pascal using the Lazarus IDE, it looked good. If you want to use Pascal instead of C. I wrote the backend code to interface to the hardware.

 

[AdrianH]

OK I approached it slightly differently, I went through what was already installed on the PC and found most of the gcc stuff already there, so from a browser did a search of IDE for gcc and top of the list was code::blocks and there is a Suse download so i will go along that route. But I have a question to HughF on this.

My PC is 64 bit, but when I compile programs I would like them on lower power machines although I will be testing them on this 64 bit PC. Do I install the 64 bit version or the 32 bit version or is there an option from the 64 bit to compile compatible for 32 bit? Also. What happened to the really old stuff at 16 bit?


eSCHen, you mentioned server problems, the wife works in a school and supports the IT section for the simple stuff, up to now they have had to recover the servers from 4 attempts at this encryption virus/package that goes around where the disks become encrypted and you have to pay for the decryption key. The security features and virus checkers are not detecting it apparently There are some real scum around at times

ps if your post was a bit off I did not notice.
Adrian

 

[HughF]

gcc -m32 flag will build you a 32 bit binary on a 64 bit system....

 

[bump]

You'll also need the 32bit libs for cross compiling.

I'm *nix based, at a predominately Scala shop.

 

[MrFluffy]

Just a comment to throw in, if your considering putting it on a embedded device longer term, dont use a pi. I much prefer the beaglebone black or similar as I've found its a damn sight more reliable and geared towards easy gpio should you need that though it doesnt have the gfx performance of the pi, but unless your using it to watch movies in your front room...
Not sure what your trying to do, make a interactive topography map or something, but theres various projects that probably cover the requirement off somehow.
I'm not officially a programmer of late, though I do impersonate one occasionally when work demands it. I use gentoo, debian, redhat, centos etc. Whatever suits the task in hand really. Gentoo is my first go to though because I like being able to see the code and tinker/fix things for myself and sometimes it matters to just have it to hand. Steep learning curve though.

 

[Henrik]

Hmm, I don't know if Linux with sounds cards will be "real time" enough to be able to do this, and have enough resolution on the microphone.

If I understand correctly, you want to basically send a "ping", then listen for the echo coming back?

So as not to interfere with your microphone etc, you'd have to stop sending the ping before listening to the echo (otherwise you'd be basically dos'ing your own signal).
Once you've stopped sending the ping, start listening, and then determine when the echo comes back?

There are a couple of problems with this, as far as I can see (having said that, they might be overcomable).

1. The echo will be much quieter than the ambient background noise - how will you filter it out so that you get a clean signal? You could presumably do some kind of filtering to just listen to the right frequency, but I have a feeling this will be quite maths heavy (https://en.wikipedia.org/wiki/Noise_reduction).
2. At least you know what you're looking for frequency wise, and you could design your sound to be very specific, so something from here might be interesting: http://stackoverflow.com/questions/...use-for-signal-sound-one-class-classification
3. How do you deal with echos? the frequency you'll be able to generate with your sound card will be too low to be directional, i.e. it's going to bounce all sorts of stuff, not just where you're pointing the speaker/microphone.
4. What happens if your CPU starts running an interrupt whilst your echo comes back? You'll probably find the execyution too jittery IMO.

When I was working for one of the geophysical research companies, they did in effect what you're proposing there, but they had a grid of microphones listening for echoes, then set off a signal (used to be an explosive, but was later generated with a signal generator that generated signals from low hz -> high hz). When the signal was generated, the sensors would record times for when they heard the different frequencies. Masses of data generated, and then some clever people coming up with ways of dealing with echoes, echoes of echoes etc from the signals, to build up a view of the layers in the earth... The jobs used to take a lot of time to run in reasonably large compute clusters, so it's not a trivial problem (and requires a lot of maths (mainly a lot of fourier transforms if i recall)).


It might be easier to use a hybrid solution with an Arduino using a sonar sensor (these are available for a couple of quid) and a Linux computer to basically generate the display for it. You could move a range finder around with e.g. a stepper motor or a dc motor etc, and record the values in each "sweep". I suspect you'd get data like "infinity, infinity, 3m, infinity infinity" + a bit of noise overlaid. The noise is easy to deal with however with e.g. medians or just simple subtraction.


The arduino's are easy to program in C, and interface readily with linux computers over Serial. They're excellent at reading sensor values, and lots of sensors are available, but they suck at GUI's though, so use a linux computer for that

 

[AdrianH]

MrFluffy,

I am definitely not going to use a PI this is only for me at this time going to be on a PC particular a HP-T5700 series. These are used as embedded devices as they generally have flash IDE's of a few Gig, but have serial, parallel, USB, Sound card and Video, I think they cost me approx £20 each when I got them. Tey can run Linux and came with XP embedded originally.

Henrik

The echo sounder is an idea for me, not really bothered about accuracy shall we say a proof of concept, but really an ainm to learn programming of a PC, but graphics and sound card.

The dsp devices are generally powerful in a small way and can provide their own filtering if required, but how I would do it I do not know, that is the idea for me to learn.

There was a person that has done it windows based. But as I said small steps I have Code::blocks installed now I need to do the hello world and figure what is next.

Adrian

 

[HughF]

So, you are building an echo sounder, for land based navigation? Like and ultrasonic distance measurement device yes?

 

[AdrianH]

A desk top exercise I would say, it should mean by the time I get there, I should understand the sound card stuff and display stuff, I hope.

Adrian

 

[HughF]

Ok, so you want to use the sound card as the dad/Adc combo... Good plan, we actually use audacity as a rudimentary spectrum analyser at work...

The dac on the sound card might easily be capable of driving a transducer above 20kHz, you would have to try it. I know the adc's are capable of sampling higher than 20kHz...

You would need to play back the transmit pulse, a square wave at the resonant frequency of your transducer, pulse length is range dependant (how far you want to look).

Once the pulse is transmitted, you begin listening. Timing is everything as all you know in sonar is the speed of sound in the medium you are transmitting (1490m/s ish for sea water), accurate timing gives you distance.

 

[AdrianH]

Here is one I found some time ago and the concept interested my as a learning project.

http://eddiem.com/projects/chirp/chirp.htm

I have the source files from this little project as they are provided for download.

Adrian

 

[HughF]

Chirp only works if you have a broadband transducer... In the underwater imaging world, broadband transducers require composite construction instead of regular piezo, and they cost a fair bit more... The processing isn't too difficult though in a dsp...

 

[AdrianH]

Also sound is a lot slower in air, at a mere 332 metres per second, so reflected from a wall 2.5 metres away the echo should be approx 15 millisecond after the chirp, with sampling at 8 thousand times a second is a sample 1/8000 or every 0.125 ms and sample speeds should be higher I hope.

All theory, as I just do not know yet.

Adrian

Food time

 

[AdrianH]

I went and bought something, not arriving until end of week if lucky, but if it works I can always use some of the electronics or try to replicate some functions on PC.

Trying to get circuit diagram but no response from the manufacturers yet, it is Easter after all.



I am hoping that is the sensor there in the pic



Just waiting for now as I try to go through my C books.

Adrian

 

[eSCHEn]

Nowt but awesome