Windows 10 Noise Cancellation
Paragon os migration. How to check out Microsoft's Chrome-like Edge browser for Windows 10 by Clifford Colby If you want to test an early version of Microsoft's reworked Edge browser, here's your chance.
I've had a little thing bugging me since I got NC-7's from Sony. Now noise-cancelling headphones are usually used in travel, I understand that. But if you have a netbook or you have a really loud home environment, is there a software that can provide a noise cancelling effect for you?To my understanding, noise cancelling occurs when the incoming sound is inverted and used to cancel the noise with an antinoise signal by way of destructive interference. So why can't this effect be used by, say, a laptop? The mic could pick up the ambient sound, and then inverted and played on a headset.So is there a program like this?
Or is there something preventing from something like this being designed?. One thing that people haven't mentioned is that the ambient noise at your laptop could easily be completely different than by your ears. Part of the reason the headphones are effective is that microphone is located right next to your ear and is therefore hearing almost exactly what you would be hearing.In addition, the headphones contain dedicated hardware that is much faster than any software solution. An IC that takes a signal and invert its will do so on the order of microseconds (µs) or less, whereas a software solution would likely be in the range of milliseconds, which is orders of magnitude longer. I'm surprised no one mentioned the possibility of sound reaching your ears before it reaches the laptop, in which case it would be impossible to cancel out.But that's only a small problem. The thing is you can only count on destructive interference in a small area, you couldn't generate noise from a single source that would destructively interfere will all other noises everywhere in a room simultaneously.You could maybe get sound canceling in a corner of the room but in another, you could also have destructive interference in one place and constructive in another (amplifying the background noise).
Basically, I see it as racing against the speed of sound. Say a background noise happens, and somehow reaches the laptop mic before it reaches your ears. Then the computer would have to analyze this sound, make an inversion, the play the inversion to your headphones over whatever is already playing. And this is all assuming perfectly-performing hardware on all levels.However, this all would have to occur before the sound actually reached your ears, which is a very, very short period of time in comparison. Seems theoretically possible, but certainly not practical at this level in computing.Sources:Assuming a safe 1 meter between the computer receiving the sound and you receiving the sound, it will take 1/340 seconds to reach you.
(2.94 x 10 -3 seconds)Assuming a decent PC has a single 3.46 GHz Processor, has a good table on page 2, where playing an MP3 and MPEG2 given a minimum processor frequency would produce the resulting Playout Delay in seconds. Assuming that the recording and inverting of the noise is relatively similar to playing an MPEG2 and we can take these measurements at face value, in order to achieve 0.10 seconds of delay, we need 1.7GHz. Even taking the extremely generous assumption that doubling the processor speed (again, dedicated solely to this task) would make the delay 10x quicker (0.01 seconds), it would still be about 3x slower than the speed of sound (0.00294). This isn't entirely accurate, as far as the processing timing goes. For one, the stream wouldn't necessarily need to be encoded or compressed. PCM would likely be ideal in this situation, as it requires little processing power to demodulate.In relation to this, and related technologies allow for real-time signal routing between an application and the hardware, bypassing Windows routing. An effect to invert the audio stream wouldn't require that much processing power in addition to that.It's not impossible to achieve 2ms latency using an ASIO in combination with a few VSTs within DAW software, utilizing little CPU.EDIT: As nomel23 pointed out, I should clarify that this is still far too long for realistic noise cancellation.
Sound Cancelling Software
This post is all about the signal routing. ASIO makes low-latency possible by use of, which is a technology that's been around about as long as x86. This way, the CPU only coordinates the data transfer, and is left available for other operations.It's cheaper from CPU utilization standpoint, and frees the CPU to perform other operations, such as plugins, etc. Routing the output and bypassing the Windows sound system allows it to run streams in a format native to the soundcard; e.g. 96 kHz PCM - which is usually the output from the ADC in the first place.This keeps things simple in the realm of modulating/demodulating - no MP3 or compressed data to decode!. 2ms.Sound travels at about 340m/s.
In 2ms, it will have traveled 68cm. 2ms is one full period of a 500Hz signal. That's 10 periods of a 5kHz signal. Even if the microphone was right at each ear.the noise or whatever you were trying to block would have hit your ear looong ago.While you consider 2ms impressive for things like audio recording and transfer, it's absolutely ages for things like signal processing where you have to react in response to something. A 10kHz signal goes from its lowest to highest value in 0.05ms.And, the time delay between the sound source and your ear and the sound source and your pc mic would have to stay the same within just a few cm to be able to make sure the up wiggle of the sound wave was added to the down wiggle of your inverted signal. The pc would somehow have to know and track the position of the sound sources in the room, relative to your head.
With reflections.this is just stupid bananas. And, what happens when the sound source is behind you, where the sound will hit your headphones before it hits the mic???The delay and the fact that the mic will hear nothing close to what each individual ear hears, makes this totally practically impossible.Come on guys, this is ask science.in the guidelines right over there -Free of anecdotesFree of layman speculation.