Improving resolution with digital processing - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 23rd June 2015, 10:52 AM   #1
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Default Improving resolution with digital processing

Hi and sorry if i will say silly things.
I am very ignorant on the subject.
But i read an article saying that to be able to read for instance a car number plate from a satellite the original resolution is not enough.
But applying some sort of digital processing (i saw mentionend the term "fractal", without understanding much) it is possible to improve the actual resolution and in the end read the number
This is extremely amazing
I wonder if anything similar could be possible for instance with a digital MP3 file.
Increasing its resolution with a particular processing methodology.
Would it be possible ?
The interesting thing is that files of small size could sound maybe even better than a CD ?

Image Processing | Geoinformatics @ IITK



Thanks a lot.
Regards, ginetto

Last edited by ginetto61; 23rd June 2015 at 10:55 AM.
  Reply With Quote
Old 23rd June 2015, 11:48 AM   #2
sreten is offline sreten  United Kingdom
diyAudio Member
 
Join Date: Nov 2003
Location: Brighton UK
Hi,

If a program knows its specifically looking at a numberplate, it
can digout the most likely number from an unrecognisable mess.

However the mess must include enough information, and
the result has no more information than the original file.
The resultant number is a best guess with a certainty factor.

It cannot reconstruct what the numberplate actually looks like.
(Though based on rules it can predict what it should look like.)

With music you don't know what your reconstructing so
you cannot improve it. MP3 has a set of construction
rules that allow the reduced data rate encoding.
You can't get any more information out of the data.

rgds, sreten.
__________________
There is nothing more practical than a really good theory - Ludwig Boltzmann
When your only tool is a hammer, every problem looks like a nail - Abraham Maslow

Last edited by sreten; 23rd June 2015 at 11:53 AM.
  Reply With Quote
Old 23rd June 2015, 12:23 PM   #3
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Hi and thanks a lot for the very kind and helpful reply.
I was clearly missing something of the process.
Issue closed.
Regards, gino
  Reply With Quote
Old 23rd June 2015, 04:28 PM   #4
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Numberplates (like many things in real life) contain lots of redundant information. This allows an algorithm to predict what it can't see from what it can see. The snag with this is that you can never be certain that you got it right.

Music contains redundancy too, which is partly why data compression works. The snag is that the more predictable the music, the more boring we find it. Hence data prediction would only work well with music which is not worth listening too.
  Reply With Quote
Old 23rd June 2015, 05:21 PM   #5
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Quote:
Originally Posted by DF96 View Post
Numberplates (like many things in real life) contain lots of redundant information. This allows an algorithm to predict what it can't see from what it can see.
The snag with this is that you can never be certain that you got it right.
Music contains redundancy too, which is partly why data compression works. The snag is that the more predictable the music, the more boring we find it. Hence data prediction would only work well with music which is not worth listening too
Hi and thanks a lot for the very interesting explanation
But another question instead ... given the same compressed MP3 is possible to get different quality of sound using different players SW ?
thanks a lot again, gino
  Reply With Quote
Old 23rd June 2015, 06:07 PM   #6
diyAudio Member
 
cogitech's Avatar
 
Join Date: Jan 2011
Location: Kamloops, BC
Some of what is going on when increasing the resolution of a digital image is "deconvolution" or "deconvolving filters". This does not necessarily require the application to "know what it is looking for". There are some incredibly powerful deconvolution algorithms in use today.

https://en.wikipedia.org/?title=Deco..._other_imaging

Could this be applied to a digital audio bit stream? I don't think so...
  Reply With Quote
Old 23rd June 2015, 06:45 PM   #7
diyAudio Member
 
Join Date: Nov 2009
Location: Birmingham, UK
Number plate recognition is made easier because in the UK (and I suppose most other countries) only one font is legal.
This way a lot can be successfully predicted by just a few data points.

If a police cars automated number plate recognition system can't read your plate they will pull you over. If you had plates made in a different font you are technically driving without an MOT and consequently without insurance either. This results in a fine, some points on your licence and when it comes to insurance renewal higher premiums.
  Reply With Quote
Old 23rd June 2015, 07:20 PM   #8
diyAudio Member
 
scott wurcer's Avatar
 
Join Date: Jan 2004
Location: cambridge ma
Quote:
Originally Posted by cogitech View Post
Could this be applied to a digital audio bit stream? I don't think so...
The mp3 encoding is not linear or time invariant and what is left over is not uncorrelated noise, all things that would trip up the usual algorithms.
__________________
"The question of who is right and who is wrong has seemed to me always too small to be worth a moment's thought, while the question of what is right and what is wrong has seemed all-important."
  Reply With Quote
Old 23rd June 2015, 07:59 PM   #9
diyAudio Member
 
Join Date: Sep 2004
Location: New Zealand
Hi
Why not start with a better codec like Ogg if small file size is needed or Flac
arguably the best of the bunch ?

If you read up on Mastering audio which should be about trying to provide you the listener
with as close as possible to the original sound,you will find good engineers
always insist on the best original source which would not/never include MP3
as it uses lossy audio data compression. A different if not better approach
is used by ECM where the recording is the key to success, rather than
post alteration. https://en.wikipedia.org/wiki/Jan_Erik_Kongshaug .

DBX and Dolby have since the 1970's generally improved the
inability of reel to reel recorders to capture full dynamic range, and
more recently the ability of digital to overload see:
ftp://ftp.dbxpro.com/pub/pdfs/WhitePapers/Type%20IV.pdf

A Type 1 DBX 150x when used with a hard disc recorder like a yamaha
cdrHD1500 is a reasonably cost effective way of experiencing the
benefits of higher dynamic range, and works exceptionally well with CD
providing dare I say an almost analogue sound to that medium, and
certainly far more acceptable. You need though to spend about $40.00
upgrading its dual and single op amps to LM4562 and LME49710.

Cheers / Chris

Last edited by Chris Daly; 23rd June 2015 at 08:16 PM.
  Reply With Quote
Old 23rd June 2015, 08:36 PM   #10
sreten is offline sreten  United Kingdom
diyAudio Member
 
Join Date: Nov 2003
Location: Brighton UK
Quote:
Originally Posted by ginetto61 View Post
But another question instead ... given the same compressed MP3 is
possible to get different quality of sound using different players SW ?
thanks a lot again, gino
Hi,

AFAIUT the effectiveness of MP3 decoding can and does vary.
As does the encoding you use when ripping files to MP3's.

Its all very complicated and I don't pretend to understand
the minor differences, just that it generally works quite well.

rgds, sreten.
__________________
There is nothing more practical than a really good theory - Ludwig Boltzmann
When your only tool is a hammer, every problem looks like a nail - Abraham Maslow
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
(Digital) Signal Processing wakibaki Everything Else 6 22nd December 2010 10:11 AM
High resolution image of Levinson digital output board Dr.H Digital Source 11 31st October 2010 03:08 AM
High Resolution Multi-Channel Digital Interface Brian Brown Digital Source 34 15th January 2008 06:48 PM


New To Site? Need Help?

All times are GMT. The time now is 10:34 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright 1999-2015 diyAudio

Content Relevant URLs by vBSEO 3.3.2