Forums

basic questions on jpeg compression

Started by manishp April 12, 2012
robert bristow-johnson <rbj@audioimagination.com> wrote:

(snip, I wrote)

>> It seems to me that you could also quantize a continuous time signal, >> though the result wouldn't be contiguous (in the mathematical sense).
> hmmm. what'sa mathematically contiguous result? another hmmm > i don't understand yet.
One you can draw without lifting the pencil off the paper. -- glen
On 4/12/12 5:59 PM, glen herrmannsfeldt wrote:
> robert bristow-johnson<rbj@audioimagination.com> wrote: > > (snip, I wrote) > >>> It seems to me that you could also quantize a continuous time signal, >>> though the result wouldn't be contiguous (in the mathematical sense). > >> hmmm. what'sa mathematically contiguous result? another hmmm >> i don't understand yet. > > One you can draw without lifting the pencil off the paper.
can you move the pencil vertically (like a staircase function)? and the word i might have used is "continuous". if you're in mathland and say "continuous" as a property, i think of what you're saying about not needing to lift the pencil. but when you say "contiguous" i think of adjacent regions in a domain that touch each other with no space in between. -- r b-j rbj@audioimagination.com "Imagination is more important than knowledge."
robert bristow-johnson <rbj@audioimagination.com> wrote:
(snip, I wrote)

>>>> It seems to me that you could also quantize a continuous time signal, >>>> though the result wouldn't be contiguous (in the mathematical sense).
(snip)
> and the word i might have used is "continuous".
Oh, yes, I meant continuous. I don't know how that happened. It is right in one line, and not the next.
> if you're in mathland and say "continuous" as a property, > i think of what you're saying about not needing to lift > the pencil. but when you say "contiguous" i think of adjacent > regions in a domain that touch each other with no space in between.
And I didn't even notice when you asked that I wrote the wrong one. -- glen
On 12 Apr., 19:10, Tim Wescott <t...@seemywebsite.com> wrote:

snip
> Do this simple experiment: > > Go harvest a picture of something of aesthetic value off of the Internet > (I'm trying to get my mind out of the gutter, so I won't suggest porn -- > directly). >
snip The most famous picture ever in image processing is the Playmate from November 1972 Lena S&#2013266166;derberg ;) -Lasse
On Fri, 13 Apr 2012 13:55:20 -0700, langwadt@fonz.dk wrote:

> On 12 Apr., 19:10, Tim Wescott <t...@seemywebsite.com> wrote: > > snip >> Do this simple experiment: >> >> Go harvest a picture of something of aesthetic value off of the >> Internet (I'm trying to get my mind out of the gutter, so I won't >> suggest porn -- directly). >> > snip > > The most famous picture ever in image processing is the Playmate from > November 1972 > Lena S&ouml;derberg
I know. And then they went and cropped it at her shoulder. :( http://en.wikipedia.org/wiki/Lenna -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com
On Fri, 13 Apr 2012 16:12:41 -0500, Tim Wescott <tim@seemywebsite.com>
wrote:

>On Fri, 13 Apr 2012 13:55:20 -0700, langwadt@fonz.dk wrote: > >> On 12 Apr., 19:10, Tim Wescott <t...@seemywebsite.com> wrote: >> >> snip >>> Do this simple experiment: >>> >>> Go harvest a picture of something of aesthetic value off of the >>> Internet (I'm trying to get my mind out of the gutter, so I won't >>> suggest porn -- directly). >>> >> snip >> >> The most famous picture ever in image processing is the Playmate from >> November 1972 >> Lena S&ouml;derberg > >I know. And then they went and cropped it at her shoulder. :( > >http://en.wikipedia.org/wiki/Lenna > >-- >My liberal friends think I'm a conservative kook. >My conservative friends think I'm a liberal kook. >Why am I not happy that they have found common ground? > >Tim Wescott, Communications, Control, Circuits & Software >http://www.wescottdesign.com
My fave bit was that she was a guest at and spoke at the 50th annual ICT conference ni 1997: http://en.wikipedia.org/wiki/Lena_S%C3%B6derberg Eric Jacobsen Anchor Hill Communications www.anchorhill.com
On 12.04.2012 11:43, manishp wrote:
> Folks, > > I have few basic questions on related to compression used in JPEG. > > 1. Is the DCT applied over Y, CB and CR separately?
Yes.
> 2. Is there a specific reason why DCT is chosen over DFT?
The way I look at it this is basically because the DCT is a real to real transformation, whereas DFT creates complex (but symmetric) numbers.
> 3. Can quantization step be roughly seen as a filter since it basically > tones down high frequency components?
Not really. The quantization causes distortion as already mentioned. DCT is picked because it is for natural images close to the transformation that decorrelates the signal, i.e. it is almost the KLT. Under a certain abstraction quantization in the DCT domain is then (almost) optimal in the rate-distortion sense. (This is actually not precisely true, but it would be true if the transformed data would be Gaussian).
> 4. Another related question is, in an earlier question I had asked if > filters really need conversion from time to frequency domain before they > can operate. The response was that filters are designed to operate directly > on time domain inputs and no conversion is normally needed.
There is no need to make a precise time (or spacial) to frequency domain transformation. Rather, you want to run a transformation that makes the signal components afterwards independent (ideally) to maximize the rate-distortion performance of the overall lossy transformation. KLT creates only uncorrelated components (not the same, only for Gaussians this is the same), and the DCT is neither exactly the KLT (though it would if the image would consist of data whose correlation only depends on the pixel position difference). Greetings, Thomas
On 4/13/2012 8:43 PM, Eric Jacobsen wrote:
> On Fri, 13 Apr 2012 16:12:41 -0500, Tim Wescott<tim@seemywebsite.com> > wrote: > >> On Fri, 13 Apr 2012 13:55:20 -0700, langwadt@fonz.dk wrote: >> >>> On 12 Apr., 19:10, Tim Wescott<t...@seemywebsite.com> wrote: >>> >>> snip >>>> Do this simple experiment: >>>> >>>> Go harvest a picture of something of aesthetic value off of the >>>> Internet (I'm trying to get my mind out of the gutter, so I won't >>>> suggest porn -- directly). >>>> >>> snip >>> >>> The most famous picture ever in image processing is the Playmate from >>> November 1972 >>> Lena S&ouml;derberg >> >> I know. And then they went and cropped it at her shoulder. :( >> >> http://en.wikipedia.org/wiki/Lenna >> >> -- >> My liberal friends think I'm a conservative kook. >> My conservative friends think I'm a liberal kook. >> Why am I not happy that they have found common ground? >> >> Tim Wescott, Communications, Control, Circuits& Software >> http://www.wescottdesign.com > > My fave bit was that she was a guest at and spoke at the 50th annual > ICT conference ni 1997: > > http://en.wikipedia.org/wiki/Lena_S%C3%B6derberg
She was a grandmother by then, still strikingly beautiful. Her English is delightful. Jerry -- Engineering is the art of making what you want from things you can get. &#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;