Audio and signal processing - unsorted stuff: Difference between revisions

From Helpful
Jump to navigation Jump to search
mNo edit summary
 
 
Line 87: Line 87:
<!--
<!--


Analysis of the spectral envelope of a digital signal,  
Linear predictive coding (LPC) is analysis of the spectral ''envelope'' of a digital signal,  
with some basic assumptions of mostly-voiced speech,
but often specifically a speech signal because it makes some basic assumptions  
and therefore useful for speech parameter estimation - and for that ''only''.
that mostly just hold for (voiced) speech.




e.g. seen in LPC vocoders, often for speech transmission.
LPC as a general algorithm is also useful in a more theoretical, 'detecting signals in noise' way,
but LPC was primarily used to parameterize speech.


LPC was created to compress speech (into parameters rather than a squished waveform).
Probably its first large application was LPC vocoders for speech transmission,
That model is roughly to use the vowel/noise, three-filter
: converting into parameters (rather than a squished waveform)  
: via a voice model (often the [[source-filter model]] of speech)




LPC is still useful because that process of compression was required to find formants,
That prediction itself is a means of further compressing these parameters.
and estimate pitch, which remain rather useful things in e.g. phonetics, speech recognition.
Being ''linear'' prediction, it is conceptually not a lot more than interpolation/extrapolation,
or linear regression.


The often-slow-changing nature of each band's parameter also makes it compressible.






 
While there is now better speech compression,  
 
LPC's estimation of pitch and formants remains useful things like
Linear Predictive coding, in the widest sense, is linear prediction applied to digital signals.
phonetics and speech recognition.
 
Where linear prediction is estimation based on a linear function - conceptually a little more than interpolation/extrapolation, or linear regression, but not by much.
 
Linear prediction itself is more often part of larger systems (e.g. as a smoother in a Kalman filter)






---


Warped LPC (WLPC)
Warped LPC (WLPC)


 
---
 
--
 
 
LPC ''in general'' is very wide,
is actually a much more general, and its origins were in general signal analyis and theory of coding, applied e.g. to detect signals in noise.
 
In the context of speech analysis and vocoders, LPC may be it
: the application of a linear filter applied to the parameters of a [https://en.wikipedia.org/wiki/Source%E2%80%93filter_model source-filter model] of speech
 
: may refer to the wider process of getting those parametters based on a waveform {{verify}}
 
 
The first amounts to sending spectral envelope information - and the often-slow-changing nature of each band's parameter also makes it compressible.
 
 
 
LPC and PSOLA seem to originate
 
 
An estimation of the spectral envelope of a digital signal, but often specifically a speech signal.
 
Used in speech analysis, speech compression.
 
 
 


The most basic decent model of human speech is probably the Harmonic + Noise model
The most basic decent model of human speech is probably the Harmonic + Noise model

Latest revision as of 16:23, 1 July 2024

The physical and human spects dealing with audio, video, and images

Vision and color perception: objectively describing color · the eyes and the brain · physics, numbers, and (non)linearity · color spaces · references, links, and unsorted stuff

Image: file formats · noise reduction · halftoning, dithering · illuminant correction · Image descriptors · Reverse image search · image feature and contour detection · OCR · Image - unsorted

Video: file format notes · video encoding notes · On display speed · Screen tearing and vsync

Simpler display types · Video display notes · Display DIY
Subtitle format notes


Audio physics and physiology: Sound physics and some human psychoacoustics · Descriptions used for sound and music

Noise stuff: Stray signals and noise · sound-related noise names · electronic non-coupled noise names · electronic coupled noise · ground loop · strategies to avoid coupled noise · Sampling, reproduction, and transmission distortions · (tape) noise reduction


Digital sound and processing: capture, storage, reproduction · on APIs (and latency) · programming and codecs · some glossary · Audio and signal processing - unsorted stuff

Music electronics: device voltage and impedance, audio and otherwise · amps and speakers · basic audio hacks · Simple ADCs and DACs · digital audio · multichannel and surround
On the stage side: microphones · studio and stage notes · Effects · sync


Electronic music:

Electronic music - musical terms
MIDI · Some history, ways of making noises · Gaming synth · microcontroller synth
Modular synth (eurorack, mostly):
sync · power supply · formats (physical, interconnects)
DAW: Ableton notes · MuLab notes · Mainstage notes


Unsorted: Visuals DIY · Signal analysis, modeling, processing (some audio, some more generic) · Music fingerprinting and identification

For more, see Category:Audio, video, images


Speech analysis and processing

Source-filter model

Vocoders

Voice coders analyses speech into parameeters, then synthesizes based on those parameters.

They were once made with the idea to parametrize speech for efficient transmission.

It was an important development in telecom, and also used in phonetics.

It was also potentially an instrument of sorts. A voder is the production part only, and sometimes is even a playable machine[1].

These days vocoders are mostly used to distort vocals and instruments in music.



See also:


Linear predictive coding (LPC) for speech; and PSOLA

STRAIGHT

Semi-sorted

Performance metrics

THD and THD+N

SINAD

Other metrics

Unsorted

Retrieved from "https://helpful.knobs-dials.com/index.php?title=Audio_and_signal_processing_-_unsorted_stuff&oldid=97765"