I have two PCM waveforms that are 99% identical but offset by X amount (typically 1-2 seconds)...

I have two PCM waveforms that are 99% identical but offset by X amount (typically 1-2 seconds). Is there an application available that can compare the two waveforms and output the offset should it detect a strong match?

Attached: PCM.png (500x375, 8K)

Other urls found in this thread:

dsp.stackexchange.com/questions/33485/phase-shift-between-two-signals
en.m.wikipedia.org/wiki/Cross-correlation
twitter.com/NSFWRedditGif

turn a waveform into a function? then you're just doing simple math from that point forward

I'm looking for something readily available but I will create a signal comparison function in Matlab should there be no alternative.

EAC has a compare 2 wav files feature

Something something Fourier transform

Sum them in audacity

what does it return? just boolean or an r value?

It's not a literal wave, it's a file

I'm not aware of anything, but if you knew how to make it, it'd be a pretty simple job. If you want the easy way, open them both up in audacity and compare the offset.

Least squares

why did I chuckle at this

Same channel or different channels?
Anyway, if they're on the same channel you can multiply by sine and cosine, do some linear algebray and decompose it to base frequencies with different phase shifts.

It's in the time domain user

Why are you even asking? If you know matlab already then writing a program to do it is a 20 minute task

Doesn't work unfortunately, "different samples" error. Both are 44.1KHz but not from the same source.

What do you mean exactly? Is this an automated function or something that needs to be performed manually?

Yeah, that's my last resort. It's just more time consuming so I'm trying to find better ways to automate the process.

hey y'think if i MANUALLY DREW a waveform by simply observing the completed waveform of an existing song, it would sound the same?
embarking on that adventure now... gna be a long term project @ this rate . . .

Attached: REDRAAAAAAAW.png (781x633, 15K)

it would if you were 100% accurate

yes.

You're looking for "cross correlation" .
You can easily do cross correlation in matlab.

Just do a correlation, what the fuck?

Literally just do an fft. You can check the similarities with the real terms and Use the complex terms to determine the phase difference.

>waveform analysis
Nope, you're the first person to ever think of this. I bet you have nice eyes too.

It took 10+ replies for someone to say this. The poor state of Jow Forums...

Something something z-transform

Since this is a DSP question, why didn't you ask /sci/?

I'm a brainlet and even I knew the answer. What the fuck, OP.

This is actually correct but vague. See: dsp.stackexchange.com/questions/33485/phase-shift-between-two-signals

I just realized you said offset by 1-2 seconds... I don't think this will be as easy as just running FFT. I forget the autismal intricacies of complex numbers, but I'm fairly certain once we go past 360 degrees (aka one cycle of your waveform, so 1-2 seconds delay corresponds to 1 or .5 hz respectively), you can no longer reliably compare signals using this approach. Look into this instead: en.m.wikipedia.org/wiki/Cross-correlation

I read two posts before ctrl+Fing "corr". This place is filled with fucking retards good god.

Bruh just find the zero crossings