BREAKING NEWS: The end of JPEG is in sight

0 views
Skip to first unread message

+/-

unread,
Sep 30, 2005, 12:14:55 AM9/30/05
to
Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
compromise technology. Young genius about to rock the DI world.

WHAT'S NEXT
Honey, I Shrunk the JPEG
By Shailaja Neelakantan, September 21, 2005
BUSINESS 2.0

If downloading digital photos stalls your PC, spare a thought for the
data networks in hospitals. A midsize hospital typically gets 60 requests
every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
enormous images can quickly cripple a network.

Enter 25-year-old Arvind Thiagarajan, co-founder of Singapore-based
startup MatrixView, who wants to revolutionize digital imaging. The
data-compression algorithm he invented shrinks images into a format called a
MatrixView Universal, or MVU, which is 15 to 300 percent smaller than a
JPEG. But unlike a JPEG, which omits details, an MVU is as precise as the
original. "Data loss is unacceptable in medical diagnosis," Thiagarajan
says. That's why the startup is focusing on health care first. A well-known
hospital in Bangalore is using the technology, and MatrixView plans to ink
deals in the coming year with several Fortune 100 health-care companies in
the United States. MatrixView is also targeting other subsets of the $9
billion U.S. digital-imaging market. Right now it's negotiating with
chipmakers to embed the technology in cameras and fit more files on storage
cards. MRIs today, vacation snaps tomorrow.

http://www.business2.com/b2/web/articles/0,17863,1106847,00.html

http://www.matrixview.com/

Download white paper:
http://matrixview.com/files/ABO%20white%20paper.pdf

Matt Ion

unread,
Sep 30, 2005, 2:16:20 AM9/30/05
to
Unless it's freely available to all developers, and doesn't include some
cockeyed protection scheme that makes it difficult for one to
backup/edit/copy one's own pictures, it'll never fly for the mass market.

+/- wrote:


---
avast! Antivirus: Outbound message clean.
Virus Database (VPS): 0539-2, 09/29/2005
Tested on: 9/29/2005 11:16:05 PM
avast! - copyright (c) 1988-2005 ALWIL Software.
http://www.avast.com

Randy Berbaum

unread,
Sep 30, 2005, 3:10:44 AM9/30/05
to
In rec.photo.digital Matt Ion <sou...@moltenimage.com> wrote:
: Unless it's freely available to all developers, and doesn't include some
: cockeyed protection scheme that makes it difficult for one to
: backup/edit/copy one's own pictures, it'll never fly for the mass market.

Also nothing is said about processing time. If a sufficiently complex
program is run, it is very possible to compress any photo much further
than any currently used photo image. For the compression of an x-ray or
MRI is "fast" if the result shows up in a few min while the dr and patient
are walking back to the dr's office. But how many of us are going to be
happy waiting 2 min between shots on our digital camera, just to save 1/3
to 1/2 the memory space. Personally I think that I would rather purchase
more memory than to have to wait several min (or even 10's of seconds)
between normal shots. JMHO

Now, if such a program were developed for archiving photos in more compact
but lossless forms, It could have a big impact.

Randy

==========
Randy Berbaum
Champaign, IL

Mike Henley

unread,
Sep 30, 2005, 3:22:03 AM9/30/05
to

Matt Ion wrote:
> Unless it's freely available to all developers, and doesn't include some
> cockeyed protection scheme that makes it difficult for one to
> backup/edit/copy one's own pictures, it'll never fly for the mass market.
>

Indeed, Jpeg is an industrywide standard by a joint ISO/IEC and ITU-T
committee.

AMD is now making 5Ghz processors. Broadband is now being offered at
24Mb. So hardware-bandwidth is no issue. Last thing the industry will
do is entrust its data formats to a proprietary one from an obscure
upstart.

Trevor

unread,
Sep 30, 2005, 3:27:59 AM9/30/05
to
<big snip>

data formats are like standards - there are so many to choose from. I woudl
rather put my trust in JPEG2000, but that's taking its time getting to the
masses - anyone up to date on the Lizardtech claims?

Ron Hunter

unread,
Sep 30, 2005, 4:14:03 AM9/30/05
to
Well, let me know with Irfanview has it, and PhotoShop adopts it, THEN I
will be impressed.
BTW, NO MENTION was made of color!


--
Ron Hunter rphu...@charter.net

Nicholas Sherlock

unread,
Sep 30, 2005, 4:14:52 AM9/30/05
to
+/- wrote:
> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
> compromise technology. Young genius about to rock the DI world.
>

My bullshit senses are tingling.


Cheers,
Nicholas Sherlock

Ron Hunter

unread,
Sep 30, 2005, 4:18:10 AM9/30/05
to
Moreover, the math in the quote leaves much to be desired. A 10
megabyte image across a gigabit ethernet connection takes less than 1
second to transmit. 60 of those an hour is hardly a significant network
load.
Then there is the aspect that MRI's are NOT COLOR. I am sure that going
back to B&W is not an option for most of us.


--
Ron Hunter rphu...@charter.net

Philip Homburg

unread,
Sep 30, 2005, 4:04:20 AM9/30/05
to
In article <NsGdnSNAL6R...@rogers.com>, +/- <u...@nu.com> wrote:
> If downloading digital photos stalls your PC, spare a thought for the
>data networks in hospitals. A midsize hospital typically gets 60 requests
>every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
>enormous images can quickly cripple a network.

10 megabytes only takes about 1 second on 100 Mbps ethernet. Is that a big
deal?

Anyhow, a 10 megabyte jpeg is probably more than 50 Mpixels. I don't know what
kind of viewing devices they have in hospitals, but starting out with
lower resolution images and then getting high res crops from the real
image (cropping at block boundaries is cheap in jpeg) strikes me as a good
solution to reduce bandwidth.

> Enter 25-year-old Arvind Thiagarajan, co-founder of Singapore-based
>startup MatrixView, who wants to revolutionize digital imaging. The
>data-compression algorithm he invented shrinks images into a format called a
>MatrixView Universal, or MVU, which is 15 to 300 percent smaller than a
>JPEG. But unlike a JPEG, which omits details, an MVU is as precise as the
>original. "Data loss is unacceptable in medical diagnosis," Thiagarajan
>says. That's why the startup is focusing on health care first.

The usual snake-oil. Lossless compression doesn't work all that well on
images that contain noise. Any algorithm that deletes noise is also going
to delete some image detail (unless the algorithm has some much domain
specific knowledge that you can save just the 'contents' of the image
and not the pixels.)


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency

Kingdom

unread,
Sep 30, 2005, 5:54:25 AM9/30/05
to
"+/-" <u...@nu.com> wrote in news:NsGdnSNAL6R...@rogers.com:

Doubt well ever even see this format nevermind use it, if they want cash
from hospitals they real are greedy bastards and it's about 2 years too
late we now have high speed everything!

--
f=Ma well, nearly...

*-- Jinn --*

unread,
Sep 30, 2005, 6:06:39 AM9/30/05
to

"Randy Berbaum" <rber...@bluestem.prairienet.org> wrote in message news:dhiodk$81h$1...@wildfire.prairienet.org...

> In rec.photo.digital Matt Ion <sou...@moltenimage.com> wrote:
> : Unless it's freely available to all developers, and doesn't include some
> : cockeyed protection scheme that makes it difficult for one to
> : backup/edit/copy one's own pictures, it'll never fly for the mass market.
>
> Also nothing is said about processing time. If a sufficiently complex
> program is run, it is very possible to compress any photo much further
> than any currently used photo image.

ABO's feautre (according to the page) is speed. It involves nothing but
integer manipulations.

> For the compression of an x-ray or
> MRI is "fast" if the result shows up in a few min while the dr and patient
> are walking back to the dr's office. But how many of us are going to be
> happy waiting 2 min between shots on our digital camera, just to save 1/3
> to 1/2 the memory space. Personally I think that I would rather purchase
> more memory than to have to wait several min (or even 10's of seconds)
> between normal shots. JMHO
>
> Now, if such a program were developed for archiving photos in more compact
> but lossless forms, It could have a big impact.

Well, FWIW, that's what ABO seems to offer.

*-- Jinn --*

unread,
Sep 30, 2005, 6:17:31 AM9/30/05
to

"Nicholas Sherlock" <n_she...@hotmail.com> wrote in message news:dhis73$db8$1...@lust.ihug.co.nz...

> +/- wrote:
>> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no compromise technology. Young genius about to rock
>> the DI world.
>>
>
> My bullshit senses are tingling.

Why?

Perhaps the forecast of doom is premature, but better algorithms for compression
aren't technically impossible.

Not like MP3 hasn't been bested from numerous angles.

Perhaps you simply like to express negativity to new ideas /just because/?


Bruce Coryell

unread,
Sep 30, 2005, 6:47:21 AM9/30/05
to

Remember how Foveon was supposed to revolutionize digital photography
and what a zero it turned out to be... All of us poor sods using
conventional CCD's were supposed to have hopelessly obsolete equipment
by now.

Chris Brown

unread,
Sep 30, 2005, 7:03:45 AM9/30/05
to
In article <ZI8%e.161$PA1....@monger.newsread.com>,

Bruce Coryell <bcor...@chesco.com> wrote:
>
>Remember how Foveon was supposed to revolutionize digital photography
>and what a zero it turned out to be... All of us poor sods using
>conventional CCD's were supposed to have hopelessly obsolete equipment
>by now.

Sssh! You'll wake *him* up...

Charlie Self

unread,
Sep 30, 2005, 7:08:56 AM9/30/05
to

Nicholas Sherlock wrote:
> +/- wrote:
> > Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
> > compromise technology. Young genius about to rock the DI world.
> >
>
> My bullshit senses are tingling.
>

Yup. Let's stick it in the catapult and see if it flies when it reaches
the end of the shot.

Stewy

unread,
Sep 30, 2005, 7:59:32 AM9/30/05
to
In article <u7c46qfmgu80m0ee4qboc74es3@inews_id.stereo.hq.phicoh.net>,
phi...@pch.home.cs.vu.nl (Philip Homburg) wrote:

> In article <NsGdnSNAL6R...@rogers.com>, +/- <u...@nu.com> wrote:
> > If downloading digital photos stalls your PC, spare a thought for the
> >data networks in hospitals. A midsize hospital typically gets 60 requests
> >every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
> >enormous images can quickly cripple a network.
>
> 10 megabytes only takes about 1 second on 100 Mbps ethernet. Is that a big
> deal?

Depends. Exactly how fast is your broadband? I'm on a LAN rated at
10mbps but do I ever get that? Well, if everyone else on the LAN logged
off, then maybe yes. As it is, I'm lucky to get 100kbps for either music
downloads or binaries.
So data compression without loss (I'm assuming these are B&W/false color
- ie 32 or 256 colors) IS very useful. How it deals with full color
JPEGs is another matter.

Message has been deleted

none

unread,
Sep 30, 2005, 8:55:27 AM9/30/05
to
*-- Jinn --* wrote:
> Perhaps the forecast of doom is premature, but better algorithms for compression
> aren't technically impossible.

Yes, they are. Huffman coding
(http://en.wikipedia.org/wiki/Huffman_coding) provably gives the most
efficient result for lossless compression. This is the algorithm that is
used for TIFF.

If you managed to find a general method for loslessly compressing
bitmapped images to 30% the size of a JPEG, you'd get more attention
than a 150-word press release on some no-name website.

-Mike

mark

unread,
Sep 30, 2005, 9:06:41 AM9/30/05
to

did it?

Mark Roberts

unread,
Sep 30, 2005, 8:55:01 AM9/30/05
to
"Charlie Self" <charl...@aol.com> wrote:

The "15 to 300 percent smaller than a JPEG" certainly sets off alarm
bells. 15 to 300 percent smaller than what JPEG quality level? And with
what *kind* of image (in terms of content): This has an influence an how
effective JPEG compression is.

Perhaps, a "naive, uninformed reporter" detector or "overhyping press
release" detector might be a better term than "bullshit senses", but
suspicion is certainly merited.


--
Mark Roberts
Photography and writing
www.robertstech.com

Lorem Ipsum

unread,
Sep 30, 2005, 9:10:03 AM9/30/05
to
> If you managed to find a general method for loslessly compressing
> bitmapped images to 30% the size of a JPEG, you'd get more attention than
> a 150-word press release on some no-name website.

And further, methods can be patented (if they meet the prerequisites of not
being public earlier, etc. etc.) so look to the patent office. Regardless,
patents do not require that the method be proven to be better, just unique.
I can check that out from work later.

The so-called white paper is topical, not adequate to tell what the author
is really doing. References to symbolic representations look like nothing
but adaptive compression schemes in the same file. Nothing new there in the
research community.

I look forward to authoritative reviews in the journals.

Now it's time to go to the day job on T2, 100mb desktop machines, fiber
optic backbones and one fast courier who can carry a few terabytes of images
in his arms up the elevator faster than God.


Message has been deleted

Lorem Ipsum

unread,
Sep 30, 2005, 9:30:32 AM9/30/05
to
Will you people who reply with lame one-liners please SNIP THE ARTICLE?
(mark)?


Philip Homburg

unread,
Sep 30, 2005, 9:26:55 AM9/30/05
to
In article <anyone4tennis-D5E...@newssv.kcn.ne.jp>,

Stewy <anyone...@hotmail.com> wrote:
>In article <u7c46qfmgu80m0ee4qboc74es3@inews_id.stereo.hq.phicoh.net>,
> phi...@pch.home.cs.vu.nl (Philip Homburg) wrote:
>
>> In article <NsGdnSNAL6R...@rogers.com>, +/- <u...@nu.com> wrote:
>> > If downloading digital photos stalls your PC, spare a thought for the
>> >data networks in hospitals. A midsize hospital typically gets 60 requests
>> >every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
>> >enormous images can quickly cripple a network.
>>
>> 10 megabytes only takes about 1 second on 100 Mbps ethernet. Is that a big
>> deal?
>
>Depends. Exactly how fast is your broadband?

At work, I can get 100 Mbps when I need it. The backbone tends to be
fast enough.

>As it is, I'm lucky to get 100kbps for either music
>downloads or binaries.

That means that you have a completely obsolete backbone.

(At home, I have about 3Mbps down (and the ISP's network has enough capacity),
so a single 10 MByte image takes about 30 seconds.)

>So data compression without loss (I'm assuming these are B&W/false color
>- ie 32 or 256 colors) IS very useful. How it deals with full color
>JPEGs is another matter.

There is not going to be any lossless compression that works better than jpeg
(without using domain specific knowledge). If you have images that are suitable
for lossless compression, compress them with png. For grayscale images,
compressing a TIFF with bzip2 may also work.

Stephen Poley

unread,
Sep 30, 2005, 9:59:20 AM9/30/05
to
On Fri, 30 Sep 2005 08:10:03 -0500, "Lorem Ipsum" <nos...@nospam.com>
wrote:

>> If you managed to find a general method for loslessly compressing
>> bitmapped images to 30% the size of a JPEG, you'd get more attention than
>> a 150-word press release on some no-name website.
>
>And further, methods can be patented (if they meet the prerequisites of not
>being public earlier, etc. etc.) so look to the patent office. Regardless,
>patents do not require that the method be proven to be better, just unique.
>I can check that out from work later.
>
>The so-called white paper is topical, not adequate to tell what the author
>is really doing. References to symbolic representations look like nothing
>but adaptive compression schemes in the same file.

That's being kind. It looks like snake-oil to me. For example the
references to the OSI model say to me that the author neither
understands the OSI model nor wants to.

There was a similar case of somebody making ludicrous compression claims
in the Netherlands a year or so back. I don't think anyone managed to
discover whether the author was a con-man or merely deluded, but he
certainly didn't have anything workable. (He is now dead, so I guess
we'll never know.) It seems to be this decade's perpetual motion
machine.

--
Stephen Poley

Matt Ion

unread,
Sep 30, 2005, 10:46:00 AM9/30/05
to
Trevor wrote:

That's a perfect example: even the wonderful fre InfanView has only very
very limited support for JPEG2000 because the plugin must be paid for.
99.9% of users are going to have no need for the format's extra
features/capabilities that are going to be worth actually paying for the
support, especially when regular JPG is more than sufficient.


---
avast! Antivirus: Outbound message clean.
Virus Database (VPS): 0539-2, 09/29/2005

Tested on: 9/30/2005 7:45:44 AM

toby

unread,
Sep 30, 2005, 11:02:13 AM9/30/05
to

*-- Jinn --* wrote:
> "Nicholas Sherlock" <n_she...@hotmail.com> wrote in message news:dhis73$db8$1...@lust.ihug.co.nz...
> > +/- wrote:
> >> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no compromise technology. Young genius about to rock
> >> the DI world.
> >>
> >
> > My bullshit senses are tingling.
>
> Why?

Press releases that begin in that vein tend to end in this one:
http://www.smh.com.au/news/Business/Inventor-faces-his-angry-creditors/2004/12/14/1102787085195.html

toby

unread,
Sep 30, 2005, 11:09:04 AM9/30/05
to

toby wrote:
> *-- Jinn --* wrote:
> > "Nicholas Sherlock" <n_she...@hotmail.com> wrote in message news:dhis73$db8$1...@lust.ihug.co.nz...
> > > +/- wrote:
> > >> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no compromise technology. Young genius about to rock
> > >> the DI world.
> > >>
> > >
> > > My bullshit senses are tingling.
> >
> > Why?
>
> Press releases that begin in that vein tend to end in this one:
> http://www.smh.com.au/news/Business/Inventor-faces-his-angry-creditors/2004/12/14/1102787085195.html

Better article @
http://www.theage.com.au/articles/2004/09/20/1095651251085.html?from=storylhs

Gerrit 't Hart

unread,
Sep 30, 2005, 11:25:13 AM9/30/05
to

"Mark Roberts" <ma...@robertstech.com> wrote in message
news:dhjce...@news3.newsguy.com...

> The "15 to 300 percent smaller than a JPEG" certainly sets off alarm
> bells.

BTW how can anything be more than 100% smaller?
100% smaller = 0

Gerrit


Barry Pearson

unread,
Sep 30, 2005, 12:14:34 PM9/30/05
to
Thomas T. Veldhouse wrote:

> In rec.photo.digital Randy Berbaum <rber...@bluestem.prairienet.org> wrote:
> >
> > Now, if such a program were developed for archiving photos in more compact
> > but lossless forms, It could have a big impact.
>
> Adobe DNG does a pretty good job with RAW file storage. It shrinks my
> NEF files from my D70 by about 25% [rough estimate].

DNG uses lossless JPEG compression. That indicates that it can't be
nearly as much as a lossy JPEG compression. So it is unlikely to be
able to compete with this new form, assuming the statement about it is
accurate. But I am sceptical about whether this is really a lossless
compression.

--
Barry Pearson
http://www.barry.pearson.name/photography/
http://www.birdsandanimals.info/

Richard Kettlewell

unread,
Sep 30, 2005, 12:22:20 PM9/30/05
to
none <no...@test.com> writes:
> *-- Jinn --* wrote:

>> Perhaps the forecast of doom is premature, but better algorithms
>> for compression aren't technically impossible.
>
> Yes, they are. Huffman coding
> (http://en.wikipedia.org/wiki/Huffman_coding) provably gives the
> most efficient result for lossless compression. This is the
> algorithm that is used for TIFF.

Err, sort of; it gives the most efficient result if you are
constrained to map each input symbol to the same output bit pattern.
That's hardly a universal constraint.

> If you managed to find a general method for loslessly compressing
> bitmapped images to 30% the size of a JPEG, you'd get more attention
> than a 150-word press release on some no-name website.

This, however, is true.

--
http://www.greenend.org.uk/rjk/

Roger Whitehead

unread,
Sep 30, 2005, 12:26:39 PM9/30/05
to
Which means you get a negative image. Quite common in photography. 8-)

--

Roger

Dave Martindale

unread,
Sep 30, 2005, 12:31:18 PM9/30/05
to
Stewy <anyone...@hotmail.com> writes:

>> 10 megabytes only takes about 1 second on 100 Mbps ethernet. Is that a big
>> deal?

>Depends. Exactly how fast is your broadband? I'm on a LAN rated at
>10mbps but do I ever get that? Well, if everyone else on the LAN logged
>off, then maybe yes. As it is, I'm lucky to get 100kbps for either music
>downloads or binaries.

The port coming out of your cable modem is 10 Mbps, but the cable's
maximum useful bandwidth (and the modem's maximum capability) is a
fraction of that. And you do have to share it with your neighbours
because there's only one wire.

Any hospital installing bargain-basement equipment today would get at
least 100 Mbps hardware and switches not hubs. That can actually
sustain at least 50 Mbps of data transfer, and many transfers can be in
progress at the same time because of the switches as long as they use
different paths. 10 MB images are not a problem.

Dave

Peter Twydell

unread,
Sep 30, 2005, 1:50:41 PM9/30/05
to
In message
<433d58d5$0$6560$5a62...@per-qv1-newsreader-01.iinet.net.au>, Gerrit 't
Hart <gth...@sad.au> writes
You beat me to it. This is sheer nonsense, along the same lines as
journalists writing "three times smaller" when they mean (I think)
"one-third as big", or "300% bigger" when there's only a 200% increase.
And another thing... (rant, mutter, mumble)
--
Peter

Ying tong iddle-i po!

Ken Weitzel

unread,
Sep 30, 2005, 2:11:46 PM9/30/05
to


I'm giving 110% effort here, but still confused :)

Ken

PcB

unread,
Sep 30, 2005, 2:18:04 PM9/30/05
to
<<The data-compression algorithm he invented shrinks images into a format
called a MatrixView Universal, or MVU, which is 15 to 300 percent smaller
than a JPEG.
>>

Er, how can you make something 300% smaller? 100% is all of it ....

--
Paul ============}
o o

// Live fast, die old //
PaulsPages are at http://homepage.ntlworld.com/pcbradley/


toby

unread,
Sep 30, 2005, 2:43:56 PM9/30/05
to
none wrote:
> *-- Jinn --* wrote:
> > Perhaps the forecast of doom is premature, but better algorithms for compression
> > aren't technically impossible.
>
> Yes, they are. Huffman coding
> (http://en.wikipedia.org/wiki/Huffman_coding) provably gives the most
> efficient result for lossless compression. This is the algorithm that is
> used for TIFF.

Huffman is not particularly effective except for bilevel (1-bit)
images. The LZW family of algorithms in particular perform better in
general. TIFF uses LZW, ZIP and varieties of RLE (such as Apple
Packbits), in addition to the CCITT Huffman-based methods defined for
faxes.

References:
TIFF standard,
http://www.digitalpreservation.gov/formats/fdd/fdd000022.shtml
LZW Explained, http://www.danbbs.dk/~dino/whirlgif/lzw.html
Intro to Data Compression,
http://www.faqs.org/faqs/compression-faq/part2/section-1.html

eawck...@yahoo.com

unread,
Sep 30, 2005, 3:18:35 PM9/30/05
to
*-- Jinn --* wrote:

>> My bullshit senses are tingling.
>
> Why?

Because there a large number of "data compression" claims that have
later been shown to be bullshit (or, usually, failure on the part of
the claimant to back up his claim). They are the perpetual motion
machines of computation.

> Perhaps you simply like to express negativity to new ideas /just because/?

More likely is that you are just ignorant of the history of these
things.

Gormless

unread,
Sep 30, 2005, 6:27:21 PM9/30/05
to

"+/-" <u...@nu.com> wrote in message news:NsGdnSNAL6R...@rogers.com...

> data networks in hospitals. A midsize hospital typically gets 60 requests
> every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
> enormous images can quickly cripple a network.
>

If 10 megabytes a minute can cripple a hospital network then I don't think
much of their networks.
And since when was a 10 Mb image 'enormous'?


Rich

unread,
Sep 30, 2005, 11:35:17 PM9/30/05
to
On Fri, 30 Sep 2005 00:14:55 -0400, "+/-" <u...@nu.com> wrote:

>Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
>compromise technology. Young genius about to rock the DI world.
>

>WHAT'S NEXT
>Honey, I Shrunk the JPEG
>By Shailaja Neelakantan, September 21, 2005
>BUSINESS 2.0
>

> If downloading digital photos stalls your PC, spare a thought for the

>data networks in hospitals. A midsize hospital typically gets 60 requests
>every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
>enormous images can quickly cripple a network.
>

> Enter 25-year-old Arvind Thiagarajan, co-founder of Singapore-based

>startup MatrixView, who wants to revolutionize digital imaging. The


>data-compression algorithm he invented shrinks images into a format called a
>MatrixView Universal, or MVU, which is 15 to 300 percent smaller than a

>JPEG. But unlike a JPEG, which omits details, an MVU is as precise as the
>original.

B.S.

Paul J Gans

unread,
Sep 30, 2005, 11:46:09 PM9/30/05
to
In rec.photo.digital Randy Berbaum <rber...@bluestem.prairienet.org> wrote:
>In rec.photo.digital Matt Ion <sou...@moltenimage.com> wrote:
>: Unless it's freely available to all developers, and doesn't include some
>: cockeyed protection scheme that makes it difficult for one to
>: backup/edit/copy one's own pictures, it'll never fly for the mass market.

>Also nothing is said about processing time. If a sufficiently complex
>program is run, it is very possible to compress any photo much further
>than any currently used photo image. For the compression of an x-ray or
>MRI is "fast" if the result shows up in a few min while the dr and patient
>are walking back to the dr's office. But how many of us are going to be
>happy waiting 2 min between shots on our digital camera, just to save 1/3
>to 1/2 the memory space. Personally I think that I would rather purchase
>more memory than to have to wait several min (or even 10's of seconds)
>between normal shots. JMHO

>Now, if such a program were developed for archiving photos in more compact
>but lossless forms, It could have a big impact.

I agree. But that is a serious major use. With an
ever larger pixel count it is getting to the point
where I will be up to my navel in DVDs with images
on them.

----- Paul J. Gans

Robert L. Haar

unread,
Oct 1, 2005, 12:16:28 AM10/1/05
to
On 2005/9/30 6:17 AM, "*-- Jinn --*" <birp...@offline.nuts> wrote:

>
> "Nicholas Sherlock" <n_she...@hotmail.com> wrote in message
> news:dhis73$db8$1...@lust.ihug.co.nz...

>> +/- wrote:
>>> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no
>>> compromise technology. Young genius about to rock
>>> the DI world.
>>>
>>

>> My bullshit senses are tingling.
>
> Why?


For one thing, the original post cannot be mathematically correct. A 300
percent reduction in size would make the file size negative.

toby

unread,
Oct 1, 2005, 1:57:05 AM10/1/05
to

Paul J Gans wrote:
> ...

>
> >Now, if such a program were developed for archiving photos in more compact
> >but lossless forms, It could have a big impact.
>
> I agree. But that is a serious major use. With an
> ever larger pixel count it is getting to the point
> where I will be up to my navel in DVDs with images
> on them.

That's odd - because my folders of 35mm/120 negatives, which carry
considerably more information than any JPEG I've seen - take up
relatively little space.

So much for the revolution.

>
> ----- Paul J. Gans

Nicholas Sherlock

unread,
Oct 1, 2005, 2:26:17 AM10/1/05
to
Paul J Gans wrote:
> I agree. But that is a serious major use. With an
> ever larger pixel count it is getting to the point
> where I will be up to my navel in DVDs with images
> on them.

But storage densities are increasing too. When blu-ray DVDs come out
(And more technology after that) it will be less of a problem.

Cheers,
Nicholas Sherlock

Charlie Self

unread,
Oct 1, 2005, 5:21:26 AM10/1/05
to

25 to 50 gigs. Not bad. My only fear is that by the time Blu Ray is
affordable, we'll be staring at 100 MP cameras. There goes the advance
in storage.

Nicholas Sherlock

unread,
Oct 1, 2005, 5:37:32 AM10/1/05
to

But if they increase together at the same rate, it'll never be any worse
than it is right now :).

Cheers,
Nicholas Sherlock

Andrew Haley

unread,
Oct 1, 2005, 6:20:41 AM10/1/05
to

Relative to what? Colour film is about 100kbytes/mm^3, whereas
current hard disc drives are at about 1.5 Mbytes/mm^3.

Andrew.

Lorem Ipsum

unread,
Oct 1, 2005, 10:34:01 AM10/1/05
to
"Andrew Haley" <andr...@littlepinkcloud.invalid> wrote in message
news:11jsonp...@news.supernews.com...

So the thread degenerates to minutiae. Unplug your drive. Now measure. See
how it works?


toby

unread,
Oct 1, 2005, 12:58:44 PM10/1/05
to

Yes, that's why I said it the remark was "odd" - since the density is
remarkably different. I'm not up to my navel in film negatives. But
then maybe the poster is storing the Bettmann Archive.

(Based on specs for a typical 200GB SATA drive such as
http://www.seagate.com/cda/products/discsales/marketing/detail/0,1081,599,00.html
I compute hard drive density at 513,349 bytes/mm^3. I suppose a 600GB
drive in the same form factor would equate to roughly your figure.)

At (uncompressed) 24-bit, I empirically measure film at a minimum of
17Kb x mm^2 (typically higher), and assuming 0.11mm base and 10%
packing waste, I compute a minimum density of 140Kb/mm^3 at 24 bit.
This does not take into account continuous-tone vs. 8-bit quantisation
(if we assume 16 bit samples, density would be more like 280Kb/mm^3,
which is interestingly enough, only half current hard drive density).
More scientific data at http://medfmt.8k.com/mf/filmwins.html

(Based on assessment of many drum scans, I put low-end 35mm at approx
15 Mp equivalent, 6x7cm at a minimum of 73 Mp. But that's a whole
different war. Not to mention 4x5" and 8x10"...)

>
> Andrew.

Rainer Latka

unread,
Oct 1, 2005, 1:59:30 PM10/1/05
to
Ron Hunter schrieb am Freitag, 30. September 2005 10:18:

[...]
> Moreover, the math in the quote leaves much to be desired. A 10
> megabyte image across a gigabit ethernet connection takes less than 1
> second to transmit. 60 of those an hour is hardly a significant
> network load.
> Then there is the aspect that MRI's are NOT COLOR. I am sure that
> going back to B&W is not an option for most of us.

provided the algorithm is really lossless, one could of course compress
the RGB channels individually...

KatWoman

unread,
Oct 1, 2005, 4:13:37 PM10/1/05
to
"Look out the sky is falling"
Chicken Little

"+/-" <u...@nu.com> wrote in message news:NsGdnSNAL6R...@rogers.com...

> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap,
> no compromise technology. Young genius about to rock the DI world.
>

> WHAT'S NEXT
> Honey, I Shrunk the JPEG
> By Shailaja Neelakantan, September 21, 2005
> BUSINESS 2.0
>
> If downloading digital photos stalls your PC, spare a thought for the
> data networks in hospitals. A midsize hospital typically gets 60 requests
> every hour for MRIs and echocardiograms. At 10 megabytes apiece, the
> enormous images can quickly cripple a network.
>

> snip
> cards. MRIs today, vacation snaps tomorrow.
>
> http://www.business2.com/b2/web/articles/0,17863,1106847,00.html
>
> http://www.matrixview.com/

Skip M

unread,
Oct 1, 2005, 9:16:28 AM10/1/05
to
"toby" <to...@telegraphics.com.au> wrote in message
news:1128146225.2...@o13g2000cwo.googlegroups.com...

A 300gig external hard drive is a whole lot smaller than the cases for 300
slides.

--
Skip Middleton
http://www.shadowcatcherimagery.com


toby

unread,
Oct 2, 2005, 12:55:08 AM10/2/05
to

I imagine I could fit a lot more than 300 slides into the space of 70
DVDs (which Paul said had reached his navel, and they probably would,
stacked from the floor). But possibly not enough slides to fill 300GB
(say 6000 uncompressed 35mm scans :). Bottom line is, should Paul be
archiving on hard disk?

Skip M

unread,
Oct 2, 2005, 12:58:09 AM10/2/05
to

"toby" <to...@telegraphics.com.au> wrote in message
news:1128228908....@g43g2000cwa.googlegroups.com...

I would say emphatically, "Hell Yes!" An external HD is less likely to be
affected by the vagaries of the main computer, too.

JohnR66

unread,
Oct 2, 2005, 9:13:47 AM10/2/05
to
"Robert L. Haar" <rlh...@comcast.net> wrote in message
news:BF6385DC.BE6FD%rlh...@comcast.net...
I like the idea. Put files on the drive and you get more space!


Rainer Latka

unread,
Oct 2, 2005, 1:05:37 PM10/2/05
to
Skip M schrieb am Sonntag, 2. Oktober 2005 06:58:
>
> "toby" <to...@telegraphics.com.au> wrote in message
[...]

>> I imagine I could fit a lot more than 300 slides into the space of 70
>> DVDs (which Paul said had reached his navel, and they probably would,
>> stacked from the floor). But possibly not enough slides to fill 300GB
>> (say 6000 uncompressed 35mm scans :). Bottom line is, should Paul be
>> archiving on hard disk?
>>
>>>
>
> I would say emphatically, "Hell Yes!" An external HD is less likely
> to be affected by the vagaries of the main computer, too.

how come? You'll have to connect it to read/write on it, so any
malicious SW will reach it. And when disconnected, the risk of being
dropped is certainly higher than with a built-in disk

toby

unread,
Oct 2, 2005, 1:44:21 PM10/2/05
to

Rainer Latka wrote:
> Skip M schrieb am Sonntag, 2. Oktober 2005 06:58:
> >
> > "toby" <to...@telegraphics.com.au> wrote in message
> [...]
> >> I imagine I could fit a lot more than 300 slides into the space of 70
> >> DVDs (which Paul said had reached his navel, and they probably would,
> >> stacked from the floor). But possibly not enough slides to fill 300GB
> >> (say 6000 uncompressed 35mm scans :). Bottom line is, should Paul be
> >> archiving on hard disk?
> >>
> >>>
> >
> > I would say emphatically, "Hell Yes!" An external HD is less likely
> > to be affected by the vagaries of the main computer, too.
>
> how come? You'll have to connect it to read/write on it, so any
> malicious SW will reach it.

It can be made read-only easily enough.

> And when disconnected, the risk of being
> dropped is certainly higher than with a built-in disk

I could drop a folder of negatives in the bath, too. DVD-Rs are very
fragile media, I'd sooner use a hermetic metal case.

Lorem Ipsum

unread,
Oct 2, 2005, 2:12:04 PM10/2/05
to
While getting into the car yesterday, I dropped 4 gig of data onto the
concrete drive. It was wrapped only in an acetate envelope. Zero damage.
(8x10" negative).


ggull

unread,
Oct 2, 2005, 4:39:59 PM10/2/05
to
"Robert L. Haar" < wrote in ...
> On 2005/9/30 6:17 AM, "*-- Jinn --*" > wrote:
>> "Nicholas Sherlock" wrote ...

>>> +/- wrote:
>>>> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum
>>>> leap, no
>>>> compromise technology. Young genius about to rock the DI world.

>>> My bullshit senses are tingling.
>> Why?
>
> For one thing, the original post cannot be mathematically correct. A 300
> percent reduction in size would make the file size negative.

Also, anyone who (or whose press release) describes themselves as a "genius"
raises the bs flag too. Especially an unknown genius.


ka...@sonic.net

unread,
Oct 2, 2005, 7:55:22 PM10/2/05
to
On Fri, 30 Sep 2005 10:17:31 GMT, "*-- Jinn --*"
<birp...@offline.nuts> wrote:

>
>"Nicholas Sherlock" <n_she...@hotmail.com> wrote in message news:dhis73$db8$1...@lust.ihug.co.nz...

>> +/- wrote:
>>> Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap, no compromise technology. Young genius about to rock
>>> the DI world.
>>>
>>
>> My bullshit senses are tingling.
>
>Why?
>

>Perhaps the forecast of doom is premature, but better algorithms for compression
>aren't technically impossible.
>

>Not like MP3 hasn't been bested from numerous angles.


>
>Perhaps you simply like to express negativity to new ideas /just because/?
>

Perhaps you like to psychoanalyze people on the basis of a
single usenet posting?

Is your license to practice current?

Alturas

unread,
Oct 3, 2005, 1:09:46 AM10/3/05
to
On Fri, 30 Sep 2005 09:27:59 +0200, "Trevor" <tr...@a.c.z> wrote:

><big snip>

>data formats are like standards - there are so many to choose from. I woudl
>rather put my trust in JPEG2000, but that's taking its time getting to the
>masses - anyone up to date on the Lizardtech claims?

Yes, what is the deal with JPEG2000? Write times too slow or
something? It's nearly 2006 and we still don't see it in digicams.

Alturas

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

Mike Jacoubowsky

unread,
Oct 3, 2005, 2:28:57 AM10/3/05
to
> Moreover, the math in the quote leaves much to be desired. A 10 megabyte
> image across a gigabit ethernet connection takes less than 1 second to
> transmit. 60 of those an hour is hardly a significant network load.
> Then there is the aspect that MRI's are NOT COLOR. I am sure that going
> back to B&W is not an option for most of us.

While it may be true that MRIs are not in color, I've been around hospital
imaging equipment enough over the last couple of years to have noticed quite
a bit of use of color on the various scanning equipment. It just ain't a
black & white world anymore (although I believe the "color" is artificially
added as a visual aid, as I suspect most of the sensors are probably just
recording essentially shades of gray).

--Mike-- Chain Reaction Bicycles
www.ChainReactionBicycles.com


"Ron Hunter" <rphu...@charter.net> wrote in message
news:6x6%e.1891$np2...@fe02.lga...


> Randy Berbaum wrote:
>> In rec.photo.digital Matt Ion <sou...@moltenimage.com> wrote:
>> : Unless it's freely available to all developers, and doesn't include
>> some : cockeyed protection scheme that makes it difficult for one to :
>> backup/edit/copy one's own pictures, it'll never fly for the mass market.
>>
>> Also nothing is said about processing time. If a sufficiently complex
>> program is run, it is very possible to compress any photo much further
>> than any currently used photo image. For the compression of an x-ray or
>> MRI is "fast" if the result shows up in a few min while the dr and
>> patient are walking back to the dr's office. But how many of us are going
>> to be happy waiting 2 min between shots on our digital camera, just to
>> save 1/3 to 1/2 the memory space. Personally I think that I would rather
>> purchase more memory than to have to wait several min (or even 10's of
>> seconds) between normal shots. JMHO
>>

>> Now, if such a program were developed for archiving photos in more
>> compact but lossless forms, It could have a big impact.
>>

>> Randy
>>
>> ==========
>> Randy Berbaum
>> Champaign, IL


>>
> Moreover, the math in the quote leaves much to be desired. A 10 megabyte
> image across a gigabit ethernet connection takes less than 1 second to
> transmit. 60 of those an hour is hardly a significant network load.
> Then there is the aspect that MRI's are NOT COLOR. I am sure that going
> back to B&W is not an option for most of us.
>
>

> --
> Ron Hunter rphu...@charter.net
>


Keith Sheppard

unread,
Oct 3, 2005, 5:31:13 AM10/3/05
to
>>Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap,
>>no compromise technology
The notion that any technological improvement necessarily heralds the end
for that which went before it is usually flawed because it ignores the
influence of marketing and commercial pressures.

Many years ago in the UK (and maybe elsewhere) we had two formats for video
recordings - VHS and Betamax. Over the course of a few years, VHS
eventually displaced Betamax and the latter sank without trace. Was this
because VHS was better? No. In fact many people in the know say that
Betamax was actually the technically superior. It was a marketing victory.
The VHS camp managed to attract more manufacturers to their cause and
eventually it became the de-facto standard.

A similar thing happened with sattelite TV broadcasting standards about ten
years later.

In the case of compression algorithms IMHO it is already too late to make a
significant dent in the grip which JPEG has on the market. The new
compression algorithm may be the best thing since sliced bread but that
alone won't help it. There is an enormous amount of hardware and software
out there which is tied to JPEG and JPEG is cemented in the semi-technical
person's mind as "the" standard for photographic image compression.

It will take more than a good algorithm to displace it. It will take
commercial pressure and probably a not inconsiderable amount of money before
it can make any headway.

Regards
Keith

Message has been deleted
Message has been deleted

Chris Brown

unread,
Oct 3, 2005, 6:57:01 AM10/3/05
to
In article <BT60f.2982$rp1....@newsfe4-win.ntli.net>,

Keith Sheppard <keith.s...@tesco.net> wrote:
>>>Finally, JPEG is doomed, algorithm geeks unite! This is the quantum leap,
>>>no compromise technology
>The notion that any technological improvement necessarily heralds the end
>for that which went before it is usually flawed because it ignores the
>influence of marketing and commercial pressures.
>
>Many years ago in the UK (and maybe elsewhere) we had two formats for video
>recordings - VHS and Betamax. Over the course of a few years, VHS
>eventually displaced Betamax and the latter sank without trace. Was this
>because VHS was better? No. In fact many people in the know say that
>Betamax was actually the technically superior.

Not this old chestnut. In general you're right about superior technology not
necessarilly leading to marketing success, but Beta vs VHS is a really bad
example of that. VHS was superior where it mattered - it had the ability to
store a full length movie at a time when Beta didn't.

Arguably the best technology was actually the third standard, the one nobody
ever mentions, Philips V2000.

All Things Mopar

unread,
Oct 3, 2005, 7:49:21 AM10/3/05
to
Today Keith Sheppard spoke these views with conviction for
everyone's edification:

> Many years ago in the UK (and maybe elsewhere) we had two
> formats for video recordings - VHS and Betamax. Over the
> course of a few years, VHS eventually displaced Betamax and
> the latter sank without trace. Was this because VHS was
> better? No. In fact many people in the know say that
> Betamax was actually the technically superior. It was a
> marketing victory. The VHS camp managed to attract more
> manufacturers to their cause and eventually it became the
> de-facto standard.

VHS vs. Beta was world-wide. What really killed Beta was
Sony's inability to get enough tape into the smaller cartridge
to compete with VHS SLP 6-hour recordings. The first Betamax
units could only record for an hour, and I think the longest
time I ever used was still around 4 or so.

I've also heard the quality issue, but personally never saw
better quality with my many Betamax recorders before I was
forced to switch to VHS. I've still got one of the last Sony
units, which I used to occasionally look at some 100 old tapes
of movies off cable TV.

> It will take more than a good algorithm to displace it. It
> will take commercial pressure and probably a not
> inconsiderable amount of money before it can make any
> headway.

I agree. Another good example is MP3 for audio. I lost track
of the standard when it went past MP15, but there's no music
nor any players that I know of that will recognize the newer
formats, and I also don't know or understand what is better
about them.

There's an old advertising saying that claims "nobody gets it
until everybody wants it". That said, your thesis is right on
the mark. It'll be really tough for even a clearly superior
compression algorithm to break through the tremendous
installed base of graphics apps that wouldn't be able to read
it for a couple of versions, if the company were even still in
business to update legacy software.

--
ATM, aka Jerry

ggull

unread,
Oct 3, 2005, 8:47:04 AM10/3/05