VS: raw.tif, compressed raw.tif and regular ole tif

[Billman]

As I'm becoming more adept at understanding and using Vuescan, I find
that my objectives are shifting. I'm scanning a boatload of negatives
and I think I would like to start scanning/storing the raw scans
rather than the jpgs I've been producing. My question is this: Does
the raw file saved by viewscan with the .tif extension have *any*
processing done on it at all?

I've seen that there is a recommendation to follow the advanced
workflow (lock exposure/lock film color base) prior scanning to a raw
file, which suggests that there is *some* processing going on, but I'm
not sure.

A secondary question is: If I do scan to a raw file and use the LZW
compressed option, what impact does this have? Can I still use VS to
"scan from file" when processing these images? Can I open this
compressed tif files in Photoshop? (I supposed the question is really
whether photoshop can interpret the LZW compression.)

Thanks -- Bill

 

[Wilson]

I've been contemplating scanning raw as well. But aren't the raw
files decipherable only to vuescan? If so, this would not be a good
choice for long term archive since, heaven forbid, eventually vuescan
will be no more?

W

 

[Mac McDougald]

A secondary question is: If I do scan to a raw file and use the LZW
compressed option, what impact does this have?

I just did one, selecting only RAW and Tiff Compression: On.
It did not do LZW on the RAW Tiff file.

Mac

 

[Mac McDougald]

I've been contemplating scanning raw as well. But aren't the raw
files decipherable only to vuescan? If so, this would not be a good
choice for long term archive since, heaven forbid, eventually vuescan
will be no more?

W

Vuescan RAW are actually TIFF format.
Actually far less proprietary than most camera RAW formats.

Mac

 

[Mac McDougald]

I just did one, selecting only RAW and Tiff Compression: On.
It did not do LZW on the RAW Tiff file.

Mac

Oops. Revise that.
I had RAW and TIFF selected.
If RAW output selected only, TIFF Compression is not an option period.

M

 

[Ed Hamrick]

As I'm becoming more adept at understanding and using Vuescan, I find
that my objectives are shifting. I'm scanning a boatload of negatives
and I think I would like to start scanning/storing the raw scans
rather than the jpgs I've been producing.

Yes, this is what I plan to do myself when I scan a lot of negatives.

My question is this: Does
the raw file saved by viewscan with the .tif extension have *any*
processing done on it at all?

No, there's no processing done at all. There's no color correction,
no scaling, and the pixels come straight from the CCD.

I've seen that there is a recommendation to follow the advanced
workflow (lock exposure/lock film color base) prior scanning to a raw
file, which suggests that there is *some* processing going on, but I'm
not sure.

Locking exposure is only useful if you want to later use the advanced
workflow. The idea is that you can't really lock the film base color
later if the exposure wasn't locked when the raw file was produced.

A secondary question is: If I do scan to a raw file and use the LZW
compressed option, what impact does this have?

This will result in smaller files if you're saving 24-bit RGB raw
files. For 48-bit RGB or 64-bit RGBI raw files, it doesn't usually
help much.

Can I still use VS to
"scan from file" when processing these images?
Yes.

Can I open this
compressed tif files in Photoshop?

Yes, but there were some older versions of Photoshop with a bug
that prevented LZW compressed raw files with 16 bit samples from
being read properly.

Regards,
Ed Hamrick

 

[Billman]

Thanks, Ed --- That's just the information I was looking for!

Bill

 

[Billman]

Thanks, Mac --

whoosh -- if compression isn't an option with raw files, that's a lot
of disc space!

Bill

 

[Erik Krause]

whoosh -- if compression isn't an option with raw files, that's a lot
of disc space!

You probably misunderstod. There is a possibility to switch on Raw
compression if 'Raw file' is selected. Available options: On, Off, Auto

 

[Mac McDougald]

You probably misunderstod. There is a possibility to switch on Raw
compression if 'Raw file' is selected. Available options: On, Off, Auto

Actually, *I'm* the one that misunderstood.
I just now see the Raw Compression dialog.
My bad.

M

 

[Mendel Leisk]

As I'm becoming more adept at understanding and using Vuescan, I find
that my objectives are shifting. I'm scanning a boatload of negatives
and I think I would like to start scanning/storing the raw scans
rather than the jpgs I've been producing. My question is this: Does
the raw file saved by viewscan with the .tif extension have *any*
processing done on it at all?

I've seen that there is a recommendation to follow the advanced
workflow (lock exposure/lock film color base) prior scanning to a raw
file, which suggests that there is *some* processing going on, but I'm
not sure.

A secondary question is: If I do scan to a raw file and use the LZW
compressed option, what impact does this have? Can I still use VS to
"scan from file" when processing these images? Can I open this
compressed tif files in Photoshop? (I supposed the question is really
whether photoshop can interpret the LZW compression.)

Thanks -- Bill

Regarding "advanced workflow" affecting Vuescan raw file, I do see a
difference. With advanced workflow on, my raw file output of Tri-X
scans was quite a bit darker. I did scan-from-disk from both version,
and could not see a difference in the finished image, so haven't
bothered with advanced workflow, when outputting raws. I do use awf
downstream though, when scanning from disk. Vuescan raw file is also
(definitely) affected by setting media type to "color negative". The
red/green/blue balance is shifted.

I've also experimented with outputting greyscale and 8 bit Vuescan Raw
Files, when scanning Tri-X. Bottom line, 48 bit rgb is superior.

A suggestion: if you're wondering something like this, why not just
try it both ways and observe the results. Then you see first-hand,
specific to your scanner results.

Regarding LZW compression of raw files, I do that all the time, and
have no problems, either in Vuescan or Photoshop. Well, except: I
clean my Vuescan raw files of dust and scratches, in Photoshop, and
while PS has no problems opening or saving the files, the compression
ratio dwindles to nil, or worse. Sometimes the re saved files are even
bigger than uncompressed. The initial Vuescan output raw file usually
has 20~30% compression. My solution is to do a Vuescan scan-from-disk
after completing the PS cleaning of each roll, outputting just raw
file again, with LZW set to "on". This regains the compression ratio
lost by PS editing.

IT IS VERY IMPORTANT TO OUTPUT THESE TO A SEPARATE DIRECTORY, and
worth setting "warn on overwrite". If you write to same directory with
same name it overwrites the file you're using for scan-from-disk, with
a very small replacement. Akin to sawing off the branch your sitting
on, I guess...

I think you're on the right track saving the Vuescan raw files. In my
workflow, they are the digital negatives, and when cleaned,
invaluable. To help cleaning them, I set up a custom PS viewing mode
with the gamma cranked right up. I will use this as proof viewing mode
when inspecting the darker areas of the raw file, but toggle back to
regular rgb for the brighter areas.

Also, you might want to check out Ian Lyon's ComputerDarkroom web
site. He has a tutorial on dust and scratch cleanup that I found very
useful. It uses the PS history brush in conjunction with "Dust and
Scratch" and "Noise Added" snapshot. When cleaning the darker areas of
raw files, I'd suggested avoiding the "add noise" step he suggests.
Using two snapshots, before and after adding noise, will allow you to
toggle the history brush source.

 

[Ed Hamrick]

Regarding "advanced workflow" affecting Vuescan raw file, I do see a
difference. With advanced workflow on, my raw file output of Tri-X
scans was quite a bit darker.

The only thing in Advanced Workflow that affects raw scan files is
the CCD exposure time.

You probably locked the exposure to a lower value, which results in
darker raw scan files.

Regards,
Ed Hamrick

 

[Billman]

Hello, Mendel -- thanks for the response on Vuescan raw files. I
tried sending you this message to you directly at (e-mail address removed), but
that bounced. I wanted to ask you a followup question:

Regarding "advanced workflow" affecting Vuescan raw file, I do see a difference.
With advanced workflow on, my raw file output of Tri-X scans was

quite a
bit darker.

I did scan-from-disk from both version, and could not see a

difference in
the finished

image, so haven't bothered with advanced workflow, when outputting

raws. I
do use

awf downstream though, when scanning from disk. Vuescan raw file is also
(definitely)
affected by setting media type to "color negative". The

red/green/blue
balance is shifted.


Four questions, actually:

(1) What is the other option besides setting the media type to "color
negative"? I know there is "image" but I haven't tried that yet.

(2) The raw scan tif files came out "negative" and I had to invert
them in
PSE. Is this typical of raw scans, or is this related to the media
type
question?

(3) I, too, notice a color shift in some images in the raw scans (when
I
inverted them). Can you tell me more about the shift you're seeing?

(4) With regard the compression, I'm experiecing the "advertised"
compression with 24 bit raw files, but not with the 48 bit raw files.
There is almost no difference in file size for 48 bit scans for me;
your
message suggested that your experience is different. Can you tell me
more
about your compression ratios right out of VS?

And thanks again for your informative response.

Bill

 

[Mendel Leisk]

Hello, Mendel -- thanks for the response on Vuescan raw files. I
tried sending you this message to you directly at (e-mail address removed), but
that bounced.

I innocently posted that email address (without the nospam) and was
deluged with garbage. I've revised my email, and it's a case of once
burned, twice shy. Sorry for the bounce. I really do not understand
why:

1. They need your email in plain text in each message.

2. They do not revise the set-up process to clearly warn of the
potential for spam, and make suggestions for setting up a garbage
email, or something like that.

3. Spammers shouldn't be "fill in the blank".

I wanted to ask you a followup question:

quite a
bit darker.
difference in
the finished
raws. I
do use
red/green/blue
balance is shifted.


Four questions, actually:

(1) What is the other option besides setting the media type to "color
negative"? I know there is "image" but I haven't tried that yet.

The other option is b/w negative. Of the 4 settings (image/slide/b&w
neg./color neg.), only color negative seems to shift the color balance
of the raw files, I guess to make an initial, wholesale compensation
for the mask.

(2) The raw scan tif files came out "negative" and I had to invert
them in
PSE. Is this typical of raw scans, or is this related to the media
type
question?

I'm not sure why you would want to invert raw files, atleast if you
intend to use them as Vuescan raw files, doing scan-from-disk. The
Vuescan raw file will never be inverted. It's described as a gamma 1
(dark, compared to pc gamma 2.2~) dump, direct from the scanner.

(3) I, too, notice a color shift in some images in the raw scans (when
I
inverted them). Can you tell me more about the shift you're seeing?

I tried awf when outputting raws, and found the ones with awf a bit
darker, though about the same color balance, I think. This was Tri-X
I'm scanning, so not a lot of "color" to begin with. Both raws seemed
to output about the same when used in scan-from-disk, so I gave up on
awf at this stage. Ed has commented on this, back a ways in this
thread.

(4) With regard the compression, I'm experiecing the "advertised"
compression with 24 bit raw files, but not with the 48 bit raw files.
There is almost no difference in file size for 48 bit scans for me;
your
message suggested that your experience is different. Can you tell me
more
about your compression ratios right out of VS?

I get files reducing down to 70~80% of what they were to begin with
(reducing in size by about 1/4). Vuescan seems to do this quite
effectively and consistantly, while Photoshop saving of these files
virtually negates the compression.

 

[Mac McDougald]

I get files reducing down to 70~80% of what they were to begin with
(reducing in size by about 1/4). Vuescan seems to do this quite
effectively and consistantly, while Photoshop saving of these files
virtually negates the compression.

- I used to think that LZW was "standard", but have found that different
programs produce quite drastically different results. The freebie
Irfanview is quite poor, and I wrote the author once, who replied he was
using a free LZW library generator of some kind and admitted it was no
where as good as the one in Photoshop/Elements. Don't know how Ed
implements.

- Also, if you are still talking about LZW of 48 bit files, this is yet
another variable. Photoshop generally does about the same ratio on these
as on 24 bit TIFFs, but once in a while, I guess due to image makeup or
something, it will produce an LZW of a 48bit file that is LARGER than the
uncompressed 48 bit file.

- I've experimented, and Corel PhotoPaint and Photoshop are fairly close
in LZW file sizes, but still significantly different, while IrfanView is
generally at least 20% larger (and doesn't open 48 bit period). I don't
generally do LZW, haven't compared VueScan's sizes, the above was all
just matter of curiosity at some point.

So, point is, LZW is not implemented the same from program to program.
And of course, image makeup determines much in degree of compression one
will get (large expanses of same color compress more, of course - a
bitmap, meaning pure b/w, will compress to 99%, etc).

Mac

 

[Chris Cox]

- Also, if you are still talking about LZW of 48 bit files, this is yet
another variable. Photoshop generally does about the same ratio on these
as on 24 bit TIFFs, but once in a while, I guess due to image makeup or
something, it will produce an LZW of a 48bit file that is LARGER than the
uncompressed 48 bit file.

Yes, LZW does have that problem (with any noisy data), and because of
compatibility issues I can't use a predictor with LZW compressed 16 bit
data (Sigh. I hate being held back by third party programs, much less
ancient versions of my own program).

- I've experimented, and Corel PhotoPaint and Photoshop are fairly close
in LZW file sizes, but still significantly different, while IrfanView is
generally at least 20% larger (and doesn't open 48 bit period). I don't
generally do LZW, haven't compared VueScan's sizes, the above was all
just matter of curiosity at some point.

So, point is, LZW is not implemented the same from program to program.

It's all the same algorithm, and all the implementations are compatible
(with a few exceptions due to bugs and predictors).

Chris