Slashdot Mirror
Donate Your Noise To Xiph/Mozilla's Deep-Learning Noise Suppression Project (xiph.org)
Mozilla-backed researchers are working on a real-time noise suppression algorithm using a neural network -- and they want your noise! Long-time Slashdot reader jmv writes:
The Mozilla Research RRNoise project combines classic signal processing with deep learning, but it's small and fast. No expensive GPUs required -- it runs easily on a Raspberry Pi. The result is easier to tune and sounds better than traditional noise suppression systems (been there!). And you can help!
From the site: Click on this link to let us record one minute of noise from where you are... We're interested in noise from any environment where you might communicate using voice. That can be your office, your car, on the street, or anywhere you might use your phone or computer.
They claim it already sounds better than traditional noise suppression systems, and even though the code isn't optmized yet, "it already runs about 60x faster than real-time on an x86 CPU."
From the site: Click on this link to let us record one minute of noise from where you are... We're interested in noise from any environment where you might communicate using voice. That can be your office, your car, on the street, or anywhere you might use your phone or computer.
They claim it already sounds better than traditional noise suppression systems, and even though the code isn't optmized yet, "it already runs about 60x faster than real-time on an x86 CPU."
I may have some extra my noise, I'll have to check.
You noise? Me Tarzan!
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
It says it can remove car noises, but can it remove the audience laughter from the Red Green Show? This is a problem someone needs to solve!
Remember when Mozilla made a web browser? Pepperidge Farm remembers...
Only the State obtains its revenue by coercion. - Murray Rothbard
Perhaps the poster is suffering from multiple personality disorder, so he can be Edit or Dave. When the Edit personality is dominant, everything is checked for spelling, style, punctuation and grammar. When Dave is dominant, less so.
When the copyright term is "forever minus a day", live every day like it's the last.
Donate you noise
Maybe he is like Schroedinger's Cat. He is in quantum superposition and only becomes Editor or Dave when observed.
curl slashdot.org > noise.txt
Shouldn't take more than a day or so to have all the info they will ever need.
"it already runs about 60x faster than real-time on an x86 CPU."
I'll get my 8086 based XT out of storage, should be perfect.
I suggest piping in a few tracks by SPK, in particular "Emanation Machine R.Gie 1916", the first track from their 1981 release "Information Overload Unit".
https://www.youtube.com/watch?v=G9b89PFYZ5g
When my wife first heard it, she said it was like having your head stuck inside a running vacuum cleaner. Follow it up with some Throbbing Gristle, perhaps.
You would need a supercomputing cluster to filter out what I would consider "noise" these days, which would include 95% of the pointless shit that makes up social media.
Guess I'm too old to help.
(CAPTCHA: supports)
I get the impression Slashdot works like the postal office where the shitty workers force the odd good worker to do shitty work or else they all look bad. Because for the rest of us, we have to proofread and not fuck up every single fucking time to keep our jobs. Seriously, if you're a volunteer and fucking up, I wouldn't care. But you're getting paid, right? Do you give two fucks about the shit you put out?
Dave's not here, man...
Hah!mmmmm....one second...
Aren't we all. This comment just collapsed. The writer less so.
I tried to donate noise; using a mac under 10.12.6. Mic is working fine. Safari asks if it can use the mic. The record button stays in for 60 seconds. The playback produces nothing.
I have great noise sources, and would not mind contributing.
I've fallen off your lawn, and I can't get up.
The entire project is on github.
I found this by going to the link in TFS.
I've fallen off your lawn, and I can't get up.
That link has only the source code. It does not include the training data set.
The submission link requires CC-0 attribution, which makes me hopeful that they plan to release the data freely. But I hunted all over the site and couldn't find either a link to the data or any comment about their plans for it going forward.
rage, rage against the dying of the light
Did you go off your meds again, grandpa?
Is to have enough signal in the first place to boost the SNR to the point where the noise becomes irrelevant. Good quality worn microphones close to the mouth, dual mic setups for simple background suppression. Sound engineers have known this forever.
"I bless every day that I continue to live, for every day is pure profit."
It's the first time we try this. We'll look at the quality of the data we get (yes, noise quality!) and if it's sufficiently good/useful, then we'll also make it available. It might take some time to sort out the useful samples from the ones that aren't since some already have noise suppression applied by the OS or browser.
Opus: the Swiss army knife of audio codec
Can it remove the noise from a Harley-Davidson running straight pipes?
An effective "democracy" creates the illusion the people have a say in their government.
Now, training is a little trickier because I cannot share the data.
I cannot share the current data I'm using because it's copyrighted. Hence asking for people for help getting data that I can redistribute.
So weâ(TM)re supposed to just give jmv a bunch of data with no way to know how he is using it?
Yes, because I have such a track record for keeping things private.
Opus: the Swiss army knife of audio codec
Cool story.
Millions of husbands just submitted the sound of their wife's voice.
You need to find more clever methods to grab the noise of my PC fan, however.
It's even worse for those of us who consume their news through the RSS feed. Every time a typo in a title or body gets corrected, that's a new entry in the feed. And they don't just update the article once, oh no no no! Every typo needs to be discovered independently of the others and warrants its own separate update. I see some articles fly by at least 5 times before the so-called "editors" decide it's finally good enough not to bother anymore, or too old for their readership to still give a fuck.
I can understand the occasional typo slipping through the cracks, but for fuck's sake, would it kill you to at least proofread it once before you hit "publish"?
Install windows on my workstation? You crazy? Got any idea how much I paid for the damn thing?
He did no such thing, and judging someone based on their level of sarcasm when dealing with a troll rather than their actions just shows everyone what you consider more important.
THAT's a good way of giving up any moral standing.
Nothing going on in there....
Do they not know how GUIDs work? That article shouldn't come through more than once - it should update your original local copy with the updated version.
And PS, if you use the URL as a GUID, then it had better never ever change - not change every time the headline gets fixed.