All in a TIFF

Morris,

I have written an argument to your numbers but deleted it after multiplued 6 bytes x 3,000,000 bytes = 18,000,000. If I would use 15 bytes per color (32,767 steps) instead of 16 bytes per color then it will be close to 17,000,000.
Thanks for the info.
Leo
Morris
rer
Thanks for the technical support for my conclusions!

-sydney21
I'm not sure if this helps you at all, but in terms of what you are
losing with a JPEG as opposed to a TIFF, it's not a lot.
The JPEG is kind of like a ZIP file. When you open the JPEG up in
an image editor like Photoshop, it uncompresses the file. Open one
of your 6M fine JPEGs up in PS, at the bottom of the window, you'll
see that it's been uncompressed to a little over 17M in size.

-Andy
 
Ok so 6mp high is TIFF on the 602 and it comes in somewhere around
17mb in size.

6mp fine is JPEG and comes in around 2.1mb in size.

TIFF is lossless and JPEG is LOSSY. Now having stated all that
what I want to know is what is actually in a TIFF file??
100% of the original information. There are many lossless format available, most are just as large as the tif format, some even larger. Lossless is lossless, it doesn't matter which lossless format you use. Most cameras use TIF because it is widely compatible and easy on the internal processor.

If you'd like to save space, you can use the PNG format to further compress a lossless TIF without losing anyoriginal information. The downside of PNG'ing is processor time, its a very intense compression scheme. Another good comprimise is the old Amiga IFF format (supported by Photoshop), which is also lossless but also very easy on the processor.

In general, pngs will be 1/2 to 2/3rds the file size of TIFs, while IFFs will be about 2/3rds the size of a TIF. All are truly lossless formats.
Is this telling me that I'm throwing away 14.9mb of detail from my
photo? Or is there something else embedded in the TIFF file that
I'm not getting with the JPEG file?
Yes, but its exponetial, not linear. In other words, throwing away just a little detail saves a whole lot of file space, but each quality increment tossed after that reaps an exponentially lesser file size benefit.

That said, unfortunately, the 602's JPEG compression level is really too harsh even at the "finest" level. This is one reason its TIFs look so much better than its Fine JPEGs. If you JPEG one of its TIFs and compare to its own Fine JPEGs, you find that the 602 uses about a 97 JPEG compression parameter in Fine mode. This produces a relativel small file, but with a rather noticable quality loss. Once taken in JPEG mode, that quality is gone forever. It would be better if the 602's Fine mode used a 99 or 100 JPEG parameter.
If I'm throwing away almost 15mb of information about my picture
shouldn't the TIFF be way better quality wise?
Yes. Its not that its a better image per se, just that the image format itself didn't already discard quality in exchance for a smaller lower file size.
Does anyone know what all is stored in the TIFF file?
All the raw information from the orignal image, same for any lossless format. Here's one way to mentally model what the camera is doing, in film terms:

Think of using the 602's TIF format as a delivery the film negative itself, you have the ability to create/view a true original image from it. Using the a camera's JPEG format is like obtaining only a very accurate print of the original image, there is a small quality loss when it is created that you will never be able to recover. If you then retouch and resave, JPEG'ing the image again is like creating a print from a print, further degrading the original information. Resaving a TIF, again as a TIF, is like altering the negative itself, so there is no inherent quality loss.
 
I'm not sure if this helps you at all, but in terms of what you are
losing with a JPEG as opposed to a TIFF, it's not a lot.
The JPEG is kind of like a ZIP file. When you open the JPEG up in
an image editor like Photoshop, it uncompresses the file. Open one
of your 6M fine JPEGs up in PS, at the bottom of the window, you'll
see that it's been uncompressed to a little over 17M in size.
That's true, but the color information for each pixel, while when decompressed again consumes the same file size, is not a perfect representation of the original image if a JPEG was used as a basis.

Imagine if you printed a 1000 x 1000 pixel image, then rescanned it into a 1000 x 1000 image. Its the same size now, but its not exactly the same image you started with--much the same saving a TIF as a JPEG, then again as a TIF.
 
sg10,

.Is not the JPG file compression ratio is equal to a size ratio of compressed to original files? Then for 6M, F and TIFF (the file sizes approx. 2M and 17M) Compression = 17M/2M = 8.5 or 1:8.5.
Leo
Ok so 6mp high is TIFF on the 602 and it comes in somewhere around
17mb in size.

6mp fine is JPEG and comes in around 2.1mb in size.

TIFF is lossless and JPEG is LOSSY. Now having stated all that
what I want to know is what is actually in a TIFF file??
100% of the original information. There are many lossless format
available, most are just as large as the tif format, some even
larger. Lossless is lossless, it doesn't matter which lossless
format you use. Most cameras use TIF because it is widely
compatible and easy on the internal processor.

If you'd like to save space, you can use the PNG format to further
compress a lossless TIF without losing anyoriginal information.
The downside of PNG'ing is processor time, its a very intense
compression scheme. Another good comprimise is the old Amiga IFF
format (supported by Photoshop), which is also lossless but also
very easy on the processor.

In general, pngs will be 1/2 to 2/3rds the file size of TIFs, while
IFFs will be about 2/3rds the size of a TIF. All are truly
lossless formats.
Is this telling me that I'm throwing away 14.9mb of detail from my
photo? Or is there something else embedded in the TIFF file that
I'm not getting with the JPEG file?
Yes, but its exponetial, not linear. In other words, throwing away
just a little detail saves a whole lot of file space, but each
quality increment tossed after that reaps an exponentially lesser
file size benefit.

That said, unfortunately, the 602's JPEG compression level is
really too harsh even at the "finest" level. This is one reason
its TIFs look so much better than its Fine JPEGs. If you JPEG one
of its TIFs and compare to its own Fine JPEGs, you find that the
602 uses about a 97 JPEG compression parameter in Fine mode. This
produces a relativel small file, but with a rather noticable
quality loss. Once taken in JPEG mode, that quality is gone
forever. It would be better if the 602's Fine mode used a 99 or
100 JPEG parameter.
If I'm throwing away almost 15mb of information about my picture
shouldn't the TIFF be way better quality wise?
Yes. Its not that its a better image per se, just that the image
format itself didn't already discard quality in exchance for a
smaller lower file size.
Does anyone know what all is stored in the TIFF file?
All the raw information from the orignal image, same for any
lossless format. Here's one way to mentally model what the camera
is doing, in film terms:

Think of using the 602's TIF format as a delivery the film negative
itself, you have the ability to create/view a true original image
from it. Using the a camera's JPEG format is like obtaining only a
very accurate print of the original image, there is a small quality
loss when it is created that you will never be able to recover. If
you then retouch and resave, JPEG'ing the image again is like
creating a print from a print, further degrading the original
information. Resaving a TIF, again as a TIF, is like altering the
negative itself, so there is no inherent quality loss.
 
Other than small overhead, yes. But thats not really lossless vs lossy. PNGs for example are compressed, but not lossy. JPEGs are compressed, and lossy. Also, TIFs aren't the largest lossless file format, just an option.

To better answer your original question, TIFs, and any lossless format, contain or at least preserve the full 24-bits of data needed to represent the exact original color of each individual pixel in the 24-bit image (i.e. each pixel is assigned 1 of 16M+ available colors by the camera).

If you do some simple math, it takes 3 bytes to represent 24-bits, so multiply by the number of pixels (6M in this case, so 18MB), to see that TIFs aren't very data compressed. A TIF is roughly the size of the raw color data required per pixel.

JPEGs are much more complex, an algorithm is applied to the image's original color data to make it look fairly close to the original, but using information that is much more densely compressed. Its easy to see how this can be done, since many images have some large areas of color. For example, a solid blue sky will be more compressed by JPEGing than a picture of a crowded stadium.

Test this yourself by painting a 1 color picture and saving it as a JPEG, the file size will be truly tiny, regardless of the megapixels. The TIF format won't care, there are no file size savings induced by using fewer total colors. JPEGs however are only approximations of the original image, the actual original data is lost forever. How close it looks to the original is determined by the degree of JPEG compression used, but even the finest JPEG possible is lossy.
Ok so 6mp high is TIFF on the 602 and it comes in somewhere around
17mb in size.

6mp fine is JPEG and comes in around 2.1mb in size.

TIFF is lossless and JPEG is LOSSY. Now having stated all that
what I want to know is what is actually in a TIFF file??
100% of the original information. There are many lossless format
available, most are just as large as the tif format, some even
larger. Lossless is lossless, it doesn't matter which lossless
format you use. Most cameras use TIF because it is widely
compatible and easy on the internal processor.

If you'd like to save space, you can use the PNG format to further
compress a lossless TIF without losing anyoriginal information.
The downside of PNG'ing is processor time, its a very intense
compression scheme. Another good comprimise is the old Amiga IFF
format (supported by Photoshop), which is also lossless but also
very easy on the processor.

In general, pngs will be 1/2 to 2/3rds the file size of TIFs, while
IFFs will be about 2/3rds the size of a TIF. All are truly
lossless formats.
Is this telling me that I'm throwing away 14.9mb of detail from my
photo? Or is there something else embedded in the TIFF file that
I'm not getting with the JPEG file?
Yes, but its exponetial, not linear. In other words, throwing away
just a little detail saves a whole lot of file space, but each
quality increment tossed after that reaps an exponentially lesser
file size benefit.

That said, unfortunately, the 602's JPEG compression level is
really too harsh even at the "finest" level. This is one reason
its TIFs look so much better than its Fine JPEGs. If you JPEG one
of its TIFs and compare to its own Fine JPEGs, you find that the
602 uses about a 97 JPEG compression parameter in Fine mode. This
produces a relativel small file, but with a rather noticable
quality loss. Once taken in JPEG mode, that quality is gone
forever. It would be better if the 602's Fine mode used a 99 or
100 JPEG parameter.
If I'm throwing away almost 15mb of information about my picture
shouldn't the TIFF be way better quality wise?
Yes. Its not that its a better image per se, just that the image
format itself didn't already discard quality in exchance for a
smaller lower file size.
Does anyone know what all is stored in the TIFF file?
All the raw information from the orignal image, same for any
lossless format. Here's one way to mentally model what the camera
is doing, in film terms:

Think of using the 602's TIF format as a delivery the film negative
itself, you have the ability to create/view a true original image
from it. Using the a camera's JPEG format is like obtaining only a
very accurate print of the original image, there is a small quality
loss when it is created that you will never be able to recover. If
you then retouch and resave, JPEG'ing the image again is like
creating a print from a print, further degrading the original
information. Resaving a TIF, again as a TIF, is like altering the
negative itself, so there is no inherent quality loss.
 
Is this telling me that I'm throwing away 14.9mb of detail from my
photo? Or is there something else embedded in the TIFF file that
I'm not getting with the JPEG file?
The way I understand it, in very simplistic terms JPEG conversions look for areas of near-matching pixels. Instead of storing every single pixel in its exact colour, over and over again like a TIFF, the JPEG will describe it only once, something like "store ALL these pixels in this colour."

In practical terms, JPEG does a great job at it and you can shoot most of your shots in FINE mode. HOWEVER, JPEG files deteriorate every time they are saved, so here is what I do:
  • Shoot on JPEG at the highest quality possible (3MB FINE or 6MB FINE)
  • Download the JPEG to computer
  • Open the JPEG in PhotoShop
  • Save the JPEG as a TIFF file using SAVE AS
  • Do ALL my tweaking on the TIFF file, saving it over and over again as many times as I want
This way, I archive the raw JPEG file exactly as it came off the camera, and the tweaked copy as a TIFF file. I can always make a JPEG copy off the TIFF as needed.

Thus, my advice:
Shoot in JPEG; work in TIFF.
 
so here is what I do:
  • Shoot on JPEG at the highest quality possible (3MB FINE or 6MB FINE)
  • Download the JPEG to computer
  • Open the JPEG in PhotoShop
  • Save the JPEG as a TIFF file using SAVE AS
  • Do ALL my tweaking on the TIFF file, saving it over and over
again as many times as I want
You can also do this using the proparatory format of your manipulation software if you prefer - PS has its own format as does PSP that I use.

As I understand it, a TIFF when saved is a flattened single layer image, albeit lossless (I think PS has the option for a different kind of tiff that preserves some additional layer data) and as a working file it may be advantageous to save more data than that - so using your applications own format, you can also often save adjustment layers, masks etc. in tact so that you can go back and work more on them, remove them etc. If you need a version to post, print or e-mail, you can temporarily flatten the layers and export as a JPEG and keep the 'working' file in tact as a 'work in progress'.

Thats' how I tend to work with images that might be returned to later. I never ever re-save the original out of the camera file - I always save a copy, in an appropriate format to what I'm doing, so my originals are always in tact (and backed up).
 
As I understand it, a TIFF when saved is a flattened single layer
image, albeit lossless (I think PS has the option for a different
kind of tiff that preserves some additional layer data) and as a
working file it may be advantageous to save more data than that -
so using your applications own format, you can also often save
adjustment layers, masks etc. in tact so that you can go back and
work more on them, remove them etc. If you need a version to post,
print or e-mail, you can temporarily flatten the layers and export
as a JPEG and keep the 'working' file in tact as a 'work in
progress'.

Thats' how I tend to work with images that might be returned to
later. I never ever re-save the original out of the camera file -
I always save a copy, in an appropriate format to what I'm doing,
so my originals are always in tact (and backed up).
This is great advice especially when you use PhotoShop to add text to an image. If you tend to get crazy like me and play with drop shadows using gausian blurs, the more recent versions of PhotoShop will automatically open each new text block in a separate layer. It is best to keep your image in these layers if you ever want to go back and tweak some more.

Don't forget to flatten the image before you send it off to an on-line print service however! You can't save the file as a JPEG until you flatten it, but my print service also accepts PhotoShop files, so it is easy to forget ... especially at my usual 4AM working time.
 
As I understand it, a TIFF when saved is a flattened single layer
image, albeit lossless (I think PS has the option for a different
kind of tiff that preserves some additional layer data) and as a
working file it may be advantageous to save more data than that -
so using your applications own format, you can also often save
adjustment layers, masks etc. in tact so that you can go back and
work more on them, remove them etc. If you need a version to post,
print or e-mail, you can temporarily flatten the layers and export
as a JPEG and keep the 'working' file in tact as a 'work in
progress'.

Thats' how I tend to work with images that might be returned to
later. I never ever re-save the original out of the camera file -
I always save a copy, in an appropriate format to what I'm doing,
so my originals are always in tact (and backed up).
This is great advice especially when you use PhotoShop to add text
to an image. If you tend to get crazy like me and play with drop
shadows using gausian blurs, the more recent versions of PhotoShop
will automatically open each new text block in a separate layer. It
is best to keep your image in these layers if you ever want to go
back and tweak some more.

Don't forget to flatten the image before you send it off to an
on-line print service however! You can't save the file as a JPEG
until you flatten it, but my print service also accepts PhotoShop
files, so it is easy to forget ... especially at my usual 4AM
working time.
--I do Exactly the same as you do 'training_editor', in fact many many people i know do the same...If you happen to have a few top photos and you want to send them to a Magine for comps or printing (like Practical Photography) then what they want are Copys of your Original Jpegs compressed slightly in a Folder called Jpgs, plus the Final Finished photos as a TIFF file in a separate folder named Tiffs..All saved to a CDR along with a little Booklet With Printed images on them...Tiffs must be saved to be read on a PC and Mac...ISO960 i think..(correct me if i got that wrong Please) I don't use Photoshop (i do have it here, but to complicated for my brain)
regards,
MrScary (DennisR)
Swansea, Wales. UK

http://www.pbase.com/dennisr
http://community.webshots.com/user/mrscarecrow

Fuji S602z--2 Minolta 7000 35mm--Koni Rapid Amiga 200 6x7
 
You are absolutely correct. Most magazines (the one I write for being an exception, darn it!) use Macs. Most magazines want the original images plus your TIFF files, simply because they are much better at tweaking the shots than most of us, and would prefer to do it themselves. They also would like to see your final result too, to see what you envisoned as the final product.

Many contributors send us photos that end up being rejected for a variety of technical reasons, so we would prefer to work on your raw images; at least until we get to know you.

(The most common reason we reject a lot of photos is that people seem to be so reluctant to turn the camera on its side! We could have the best photo in the world, but if it's not shot as a vertical, there is no way we can use it for our cover.)

Plus, don't forget that there is a big difference between printing images on an ink jet or photo store printer, and a commercial printing press. When we go to print, it is not the resolution of the printer that counts anymore, it is the halftone screen frequency. This is expressed in lines-per-inch and it has nothing to do with pixels.

A good rule-of-thumb that we use is that the resolution of the photo in dots-per-inch should be at least 1.5 times the line frequency, and preferably 2 times the line frequency. Our magazine (except the cover) is printed on a Linotronic at 133 lines-per-inch. Therefore, photos should be 266 dpi for the best quality (in the final printed size). Thats why 300 dpi is a good number to aim for, because that gives us a little cropping room.

Unlike photo or inkjet printers, anything over 2 times the screen frequency is totally wasted, and simply slows down the process. (The printer warns us anytime it sees more than 2.5 times the screen frequency.)
 
You are absolutely correct. Most magazines (the one I write for
being an exception, darn it!) use Macs. Most magazines want the
original images plus your TIFF files, simply because they are much
better at tweaking the shots than most of us, and would prefer to
do it themselves. They also would like to see your final result
too, to see what you envisoned as the final product.

Many contributors send us photos that end up being rejected for a
variety of technical reasons, so we would prefer to work on your
raw images; at least until we get to know you.

(The most common reason we reject a lot of photos is that people
seem to be so reluctant to turn the camera on its side! We could
have the best photo in the world, but if it's not shot as a
vertical, there is no way we can use it for our cover.)

Plus, don't forget that there is a big difference between printing
images on an ink jet or photo store printer, and a commercial
printing press. When we go to print, it is not the resolution of
the printer that counts anymore, it is the halftone screen
frequency. This is expressed in lines-per-inch and it has nothing
to do with pixels.

A good rule-of-thumb that we use is that the resolution of the
photo in dots-per-inch should be at least 1.5 times the line
frequency, and preferably 2 times the line frequency. Our magazine
(except the cover) is printed on a Linotronic at 133
lines-per-inch. Therefore, photos should be 266 dpi for the best
quality (in the final printed size). Thats why 300 dpi is a good
number to aim for, because that gives us a little cropping room.

Unlike photo or inkjet printers, anything over 2 times the screen
frequency is totally wasted, and simply slows down the process.
(The printer warns us anytime it sees more than 2.5 times the
screen frequency.)
--The Vertical Format in Digital....How many of us are Guilty of not useing it when we should do....I'm Guilty!! Since useing my S602z I seem to use the Landscape mode at least 99% of the time, why I do that I don't know..So when I crop I nearly always end up with a Square format photo...Mind you the Square format can be quite nice, but as you said, Vertical Format must be used,especialy for any subject thats Vertical.
ie,Human, Buildings,(unless large and 1 storie high) Trees n others...

I need to win the Lottery before I can buy one of them Printers, untill then, my humble Epson 740 inkjet will have to sufice..(at least till I can afford a 950,1220,or 2200/2100)....
Will the year 2003 be kind to me!
Thanks for replying...
regards,
MrScary (DennisR)
Swansea, Wales. UK

http://www.pbase.com/dennisr
http://community.webshots.com/user/mrscarecrow

Fuji S602z--2 Minolta 7000 35mm--Koni Rapid Amiga 200 6x7
 

Keyboard shortcuts

Back
Top