Naive question about JPGs

5 views
Skip to first unread message

Another John

unread,
Sep 8, 2021, 5:08:45 AMSep 8
to
A (fairly) naive question about JPGs:

I'm not a "photographer", but I am a very keen "snapper": I've taken
photographs since I had my first Kodak Cresta in the 50s.

I also spent my working life in computing (as a documenter and educator,
not a coder), so I do know a fair amount.


I have the ingrained habit (since the days, when disk space was really
expensive) of editing my pictures down to a smaller size. (I never take
RAWs, only JPGs).

Mainly, I scale them (maybe 50%). (I have used the sainted
GraphicConverter since its Version2.)

BUT (the question, finally): if I simply Save the JPG without doing
anything to it, it will often go down in size very considerably (e.g.,
say 5.5MB to 3.2MB).

What is the data that is being lost?


I do know all about "JPG loss of quality", and have done since I started
using GC. I use GC's "90% quality" when saving - I can detect
virtually no difference in the pixels, even when enlarged to a great
scale, so I'm curious as to what has gone from the original JPG, for it
to have lost around 30% of the size.


My scaled down pictures, btw, always appear great on a computer screen,
on our digital display frame (nowadays the most-used medium), and even
when printed out at the local Max Spielmann.

Cheers all,
John

Jaimie Vandenbergh

unread,
Sep 8, 2021, 7:00:06 AMSep 8
to
On 8 Sep 2021 at 10:08:43 BST, "Another John" <lal...@hotmail.com>
wrote:

> BUT (the question, finally): if I simply Save the JPG without doing
> anything to it, it will often go down in size very considerably (e.g.,
> say 5.5MB to 3.2MB).
>
> What is the data that is being lost?

Nice writeup at
https://www.techradar.com/news/computing/all-you-need-to-know-about-jpeg-compression-586268/2

TL;DR: Like mp3, jpeg exploits weaknesses in the human perception
systems to toss out data that we'll never be aware of.

The size of a jpeg on open/re-save depends on the exact assumptions and
levels of "throwaway the data" that the original jpeg-producing software
used, and what the new software you're saving in uses.

Generally speaking a jpeg fresh off the camera will be intentionally
*very* high quality, in order to respond best to editing without going
bad. Same as you'd record 'master' audio tracks at a higher bitrate than
your end AAC/MP3 would be.

Cheers - Jaimie
--
"the first successful time machine will be used to retrieve
lost Doctor Who episode footage." - KKC, ugvm

J. J. Lodder

unread,
Sep 8, 2021, 7:00:54 AMSep 8
to
Another John <lal...@hotmail.com> wrote:

> A (fairly) naive question about JPGs:
>
> I'm not a "photographer", but I am a very keen "snapper": I've taken
> photographs since I had my first Kodak Cresta in the 50s.
>
> I also spent my working life in computing (as a documenter and educator,
> not a coder), so I do know a fair amount.
>
>
> I have the ingrained habit (since the days, when disk space was really
> expensive) of editing my pictures down to a smaller size. (I never take
> RAWs, only JPGs).
>
> Mainly, I scale them (maybe 50%). (I have used the sainted
> GraphicConverter since its Version2.)

I waste of time, these days, with HDs costing cents/GB.

> BUT (the question, finally): if I simply Save the JPG without doing
> anything to it, it will often go down in size very considerably (e.g.,
> say 5.5MB to 3.2MB).

There is some default setting for the compression, in that case.

> What is the data that is being lost?

Fine detail. However, it is the wrong question,
because there is data and data.
'Data' doesn't distinguish between real data, that is your image,
and just noise.
In the sense of Shannon information gets lost in compression.
In a real sense what gets lost in going from RAW to JPEG
is mostly sensor noise.
Remember that in the sense of Shannon
a picture that is nothing but noise
has the highest information content of all.

> I do know all about "JPG loss of quality", and have done since I started
> using GC. I use GC's "90% quality" when saving - I can detect
> virtually no difference in the pixels, even when enlarged to a great
> scale, so I'm curious as to what has gone from the original JPG, for it
> to have lost around 30% of the size.

You should go to great scale, then the look at the fine detail.

> My scaled down pictures, btw, always appear great on a computer screen,
> on our digital display frame (nowadays the most-used medium), and even
> when printed out at the local Max Spielmann.

This 'loss of information in compression' matra
is no more than a red herring that survives from the days
when pictures were measure in kilobytes,

Jan



J. J. Lodder

unread,
Sep 11, 2021, 4:53:03 PMSep 11
to
Jaimie Vandenbergh <jai...@usually.sessile.org> wrote:

> On 8 Sep 2021 at 10:08:43 BST, "Another John" <lal...@hotmail.com>
> wrote:
>
> > BUT (the question, finally): if I simply Save the JPG without doing
> > anything to it, it will often go down in size very considerably (e.g.,
> > say 5.5MB to 3.2MB).
> >
> > What is the data that is being lost?
>
> Nice writeup at
> https://www.techradar.com/news/computing/all-you-need-to-know-about-jpeg-compr
ession-586268/2
>
> TL;DR: Like mp3, jpeg exploits weaknesses in the human perception
> systems to toss out data that we'll never be aware of.
>
> The size of a jpeg on open/re-save depends on the exact assumptions and
> levels of "throwaway the data" that the original jpeg-producing software
> used, and what the new software you're saving in uses.
>
> Generally speaking a jpeg fresh off the camera will be intentionally
> *very* high quality, in order to respond best to editing without going
> bad. Same as you'd record 'master' audio tracks at a higher bitrate than
> your end AAC/MP3 would be.

Some people do have strange ideas about it.
The dislike JPEG (instead of RAW) because of the "throwaway the data".
OTOH they may be willing to 'de-noise' their pictures,
even to the point of paying for the software that does it,
because of the of the "throwaway the data" that it accomplishes.

BTW, taking up your audio example: 'remastered' CDs from old material
also achieve some their new quality by throwing away 'data'
from the original audio tapes,

Jan

--
"Some informations are more informative than others"

Stefen - fretwizzen

unread,
Sep 11, 2021, 5:39:10 PMSep 11
to
I usually go post by post unless someone has a past of really outrageous
trolling. With Peeler, I already understand what his method is, all he hungers
for is to harm Mike Easter and, of course, he'll do _anything_ to get it.
His favorite routine is to play 'shill' but reality shows it is all of Usenet
who are his targets. Peeler gets excited that Mike Easter is on the other
end. Imagine if Peeler walked up to a chimpanzee and delivered the punch
line. It wouldn't be funny.

Peeler continues to reference the doxing allegation he wrote about me, even
tho Mike Easter directly wrote that he lied about it. Why do you think Mike
Easter is trying to babysit Peeler? Even he sees his BS. He can't comment
on *any* of the stuff I quote because he sees it's true. I am happy the person
you are attacking is in my shit list.

--
Get Rich Slow!
https://duckduckgo.com/?q=dustin+cook+the+functionally+illiterate+fraud
Dustin Cook: Functionally Illiterate Fraud

STALKING_TARGET_58

unread,
Sep 11, 2021, 9:24:56 PMSep 11
to
I have known infants who argue better than Snit sock Jeremy does. Hint:
You will not go into a meet and greet, knock back all the moonshine, bang
all the goats, misappropriate the food and blow chunks in the foyer without
being ridiculed. Who *doesn't* know that this kind of bull is Snit sock
Jeremy's MO, not the approach of Carroll? It's all just nonsense... the
flooding, the forging, the writing of nonsense, the anger... the false
bravado and willingness to show how flustered he is over being banned from
the sandbox for crapping in it again... Snit sock Jeremy can not help himself
;)

Bing for 'functionally illiterate fraud' and Dustin Cook is found: <https://www.google.com/search?q=Dustin+Cook+the+functionally+illiterate+fraud>.
Fool! Socks that Snit sock Jeremy can not prove are mine. Why would Carroll,
or anyone, create socks to expose what we all regularly say about Snit
sock Jeremy?

--
One Smart Penny!
https://www.bing.com/search?q=Dustin+Cook+the+functional+illiterate+fraud
108 Warrior Dr. Kingsport, TN 37663
Reply all
Reply to author
Forward
0 new messages