JPEGmini is 75% smaller than JPEG and not proprietary

[sharon carmel]

No justification for WebP!
http://www.jpegmini.com

[Pascal Massimino]

Hi,

On Sat, Jul 16, 2011 at 9:51 AM, sharon carmel <sharonca...@gmail.com> wrote:

No justification for WebP!
http://www.jpegmini.com

quick test with the jpg on the gallery, attached.

The forest on the right got a tad blurrier, note.

 skal

--
You received this message because you are subscribed to the Google Groups "WebP Discussion" group.
To post to this group, send email to webp-d...@webmproject.org.
To unsubscribe from this group, send email to webp-discuss...@webmproject.org.
For more options, visit this group at http://groups.google.com/a/webmproject.org/group/webp-discuss/?hl=en.

[Szabolcs Péter]

"JPEGmini is currently available as an online service."

"JPEGmini is a patent-pending photo recompression technology"

How is it non-proprietary?

SyP

[Ram Ramani]

On Sat, Jul 16, 2011 at 11:11 PM, Pascal Massimino <pascal.m...@gmail.com> wrote:

Hi,

On Sat, Jul 16, 2011 at 9:51 AM, sharon carmel <sharonca...@gmail.com> wrote:

No justification for WebP!
http://www.jpegmini.com

quick test with the jpg on the gallery, attached.

The forest on the right got a tad blurrier, note.

And where is the 75% gain? The photo reduced from 44KB to 39KB, that's ~11% if my math is right.

[Vikas Arora]

On Mon, Jul 18, 2011 at 2:25 PM, Ram Ramani <ram...@google.com> wrote:

On Sat, Jul 16, 2011 at 11:11 PM, Pascal Massimino <pascal.m...@gmail.com> wrote:

Hi,

On Sat, Jul 16, 2011 at 9:51 AM, sharon carmel <sharonca...@gmail.com> wrote:

No justification for WebP!
http://www.jpegmini.com

quick test with the jpg on the gallery, attached.

The forest on the right got a tad blurrier, note.

And where is the 75% gain? The photo reduced from 44KB to 39KB, that's ~11% if my math is right.

 

The 75% gain is not achieved for this image (taken from our Gallery). But for other images seeded in the site (http://www.jpegmini.com/main/home), the compression ratio is coming out between 4X-5X (corresponds to 75%-80% smaller Jpeg).

[Urvang Joshi]

I just compared JPEGmini and WebP image sizes for 5 images and here
are the comparative results:

Image JPEG Size (KB) JPEGmini Size (KB) WebP Size (KB)
http://www.gstatic.com/webp/gallery/1.jpg 43.84 39 29.61
http://www.gstatic.com/webp/gallery/2.jpg 86.25 78 59.18
http://www.gstatic.com/webp/gallery/3.jpg 297.05 260 198.38
http://www.gstatic.com/webp/gallery/4.jpg 251.03 231 172.82
http://www.gstatic.com/webp/gallery/5.jpg 120.78 99 80.76

Notice that WebP image sizes are much smaller than JPEGmini in all
cases.

On Jul 18, 4:36 pm, Vikas Arora <vik...@google.com> wrote:
> On Mon, Jul 18, 2011 at 2:25 PM, Ram Ramani <ram...@google.com> wrote:
> > On Sat, Jul 16, 2011 at 11:11 PM, Pascal Massimino <
> > pascal.massim...@gmail.com> wrote:
>
> >> Hi,
>
> >> On Sat, Jul 16, 2011 at 9:51 AM, sharon carmel <
> >> sharoncarmel2...@gmail.com> wrote:
>
> >>> No justification for WebP!
> >>>http://www.jpegmini.com
>
> >> quick test with the jpg on the gallery, attached.
> >> The forest on the right got a tad blurrier, note.
>
> > And where is the 75% gain? The photo reduced from 44KB to 39KB, that's ~11%
> > if my math is right.
>
> The 75% gain is not achieved for this image (taken from our Gallery). But
> for other images seeded in the site (http://www.jpegmini.com/main/home), the
> compression ratio is coming out between 4X-5X (corresponds to 75%-80%
> smaller Jpeg).
>
>
>
>
>
>
>
> >>  skal
>
> >>> --
>
> >>> You received this message because you are subscribed to the Google Groups
> >>> "WebP Discussion" group.

> >>> To post to this group, send email to webp-disc...@webmproject.org.

> >>> To unsubscribe from this group, send email to

> >>> webp-discuss+unsubscr...@webmproject.org.

> >>> For more options, visit this group at
> >>>http://groups.google.com/a/webmproject.org/group/webp-discuss/?hl=en.
>
> >>  --
> >> You received this message because you are subscribed to the Google Groups
> >> "WebP Discussion" group.

> >> To post to this group, send email to webp-disc...@webmproject.org.

> >> To unsubscribe from this group, send email to

> >> webp-discuss+unsubscr...@webmproject.org.

> >> For more options, visit this group at
> >>http://groups.google.com/a/webmproject.org/group/webp-discuss/?hl=en.
>
> >  --
> > You received this message because you are subscribed to the Google Groups
> > "WebP Discussion" group.

> > To post to this group, send email to webp-disc...@webmproject.org.

> > To unsubscribe from this group, send email to

> > webp-discuss+unsubscr...@webmproject.org.

[Vikas Arora]

Urvang -

Thanks for running analysis. JpegMini may have some tunnable parameters (not exposed via this service) for controlling Size/Quality balance. But with default  (JpegMini) parameters, WebP images are approximately 25% compacter.

- Vikas

To post to this group, send email to webp-d...@webmproject.org.
To unsubscribe from this group, send email to webp-discuss...@webmproject.org.

[Pascal Massimino]

Hi,

so far i can tell, jpegmini is:

  a) removing comments and useless chunks

  b) optimizing the Huffman tables for those who forgot the '-optimize' flag on cjpeg.

  c) raising the quantization matrices a little:

Example, on '3.jpg':

before:

|    - Quantization matrix #0 -    |    - Quantization matrix #1 -    

|   6   4   5   6   5   4   6   6  |   7   7   7  10   8  10  19  10  

|   5   6   7   7   6   8  10  16  |  10  19  40  26  22  26  40  40  

|  10  10   9   9  10  20  14  15  |  40  40  40  40  40  40  40  40  

|  12  16  23  20  24  24  23  20  |  40  40  40  40  40  40  40  40  

|  22  22  26  29  37  31  26  27  |  40  40  40  40  40  40  40  40  

|  35  28  22  22  32  44  32  35  |  40  40  40  40  40  40  40  40  

|  38  39  41  42  41  25  31  45  |  40  40  40  40  40  40  40  40  

|  48  45  40  48  37  40  41  40  |  40  40  40  40  40  40  40  40  

|  Estimated quality factor:  81.  |  Estimated quality factor:  81.  

after:

|    - Quantization matrix #0 -    |    - Quantization matrix #1 -    

|  13   9   5   6   5   4  13   6  |   7   7   7  11  17  11  21  21  

|   5   6   7   7  13   8  11  17  |  21  21  44  53  45  53  44  44  

|  21  11  10  10  11  22  15  16  |  44  44  44  44  44  44  44  44  

|  13  17  47  41  49  26  25  41  |  44  44  44  44  44  44  44  44  

|  45  24  28  32  41  63  28  55  |  44  44  44  44  44  44  44  44  

|  38  31  24  45  35  48  65  38  |  44  44  44  44  44  44  44  44  

|  42  79  83  46  83  27  34  91  |  44  44  44  44  44  44  44  44  

|  97  50  44  97  41  81  83  44  |  44  44  44  44  44  44  44  44  

|  Estimated quality factor:  73.  |  Estimated quality factor:  78.  

[sharon carmel]

All, JPEGmini can reduce the size lower than WebP, but WebP is
changing the photos! With 5.jpg WebP changes all the reds! with 1.jpg
WebP smoothen the entire photo.
JPEGmini compresses photos to the visibility threshold, i.e it does
not change the photo, you can compress more, but what is the use of it
if you change the photo?

On Jul 18, 3:14 pm, Urvang Joshi <urv...@google.com> wrote:
> I just compared JPEGmini and WebP image sizes for 5 images and here
> are the comparative results:
>

> Image                                           JPEG Size (KB)  JPEGmini Size (KB)      WebP Size (KB)http://www.gstatic.com/webp/gallery/1.jpg 43.84           39                      29.61http://www.gstatic.com/webp/gallery/2.jpg 86.25           78                      59.18http://www.gstatic.com/webp/gallery/3.jpg 297.05          260                     198.38http://www.gstatic.com/webp/gallery/4.jpg 251.03          231                     172.82http://www.gstatic.com/webp/gallery/5.jpg 120.78          99                      80.76

[Monty Montgomery]

On Mon, Jul 18, 2011 at 3:18 PM, sharon carmel
<sharonca...@gmail.com> wrote:
> All, JPEGmini can reduce the size lower than WebP, but WebP is
> changing the photos! With 5.jpg WebP changes all the reds!

You may be seeing an incorrect colorspace conversion, ie, a bug and
not a real difference.

> with 1.jpg
> WebP smoothen the entire photo.
> JPEGmini compresses photos to the visibility threshold

'visibility threshold' as determined by what? There's generally no
such thing as a hard perception threshold in any human sense. Both
WebP and JPEGmini are losing information over the original.

>, i.e it does
> not change the photo

Yes it does.

> you can compress more, but what is the use of it
> if you change the photo?

Both are changing the photo. This is inescapable in multi-generation
lossy compression.

I have my own reservations about WebP's justification-- but if I have
quibbles with WebP, then JPEGmini fills my head with snake-oil alarm
bells.

Anyone else here remember what a 'Krinnish gnome' is?

Monty

[sharon carmel]

>> You may be seeing an incorrect colorspace conversion, ie, a bug and
>> not a real difference.

I think it is not a bug, WebP changing the photo and you cant avoid
it.
It is a real difference.

>> 'visibility threshold' as determined by what? There's generally no
>> such thing as a hard perception threshold in any human sense. Both
>> WebP and JPEGmini are losing information over the original.

Determined by a perceptual quality measure, try http://www.jpegmini.com,
you might change your mind.

JPEGmini is serious, better than WebP, and produce standard JPEGs.

Sharon

On Jul 18, 10:42 pm, Monty Montgomery <xiphm...@gmail.com> wrote:
> On Mon, Jul 18, 2011 at 3:18 PM, sharon carmel
>