Minolta 5400 and Vuescan 7.6.81

[thomas]

Hi,
First I would like to say that I appreciate Vuescan and the work done
by Ed. Second I have to admitt that still after 6 months trying I do
not know how to use it with my "new" Minolta Dimage 5400.
During the last months I have mainly had problems when using the
program with the grain dissolver on. For version 7.6.81 I do only get
good (better) results if the grain dissolver is on! Can anybody help
me with a good and stable workflow for Minolta 5400 with Vuescan. I am
totaly lost.
Regards
ThomasD

 

[Shaun Markham]

Hi,
First I would like to say that I appreciate Vuescan and the work done
by Ed. Second I have to admitt that still after 6 months trying I do
not know how to use it with my "new" Minolta Dimage 5400.
During the last months I have mainly had problems when using the
program with the grain dissolver on. For version 7.6.81 I do only get
good (better) results if the grain dissolver is on! Can anybody help
me with a good and stable workflow for Minolta 5400 with Vuescan. I am
totaly lost.
Regards
ThomasD

I can sympathise with you one this one although my experience has been
slightly different. I have had a 5400 for 7 months now and up to version
7.6.78 I could not calibrate the thing properly using IT8 calibration
targets. I found that I could not get any detail in the black regions of a
scan with grain dissolver switched off and I also found my prints were not
very good. Switching the grain dissolver on improved things a lot but my
colours were still very poor. Since upgrading to 7.6.80 and 81, suddenly
it works great with grain dissolver both off and on and my workflow is now
a lot simpler with printed colours looking the way I expect. So as far as
I can tell the density calibration problem has been fixed - thanks very
much Ed - sanity restored.

I'm not sure what your problem is though. What do you find the
specific differences are between the grain dissolver off and on?

Shaun

 

[thomas]

I can sympathise with you one this one although my experience has been
slightly different. I have had a 5400 for 7 months now and up to version
7.6.78 I could not calibrate the thing properly using IT8 calibration
targets. I found that I could not get any detail in the black regions of a
scan with grain dissolver switched off and I also found my prints were not
very good. Switching the grain dissolver on improved things a lot but my
colours were still very poor. Since upgrading to 7.6.80 and 81, suddenly
it works great with grain dissolver both off and on and my workflow is now
a lot simpler with printed colours looking the way I expect. So as far as
I can tell the density calibration problem has been fixed - thanks very
much Ed - sanity restored.

I'm not sure what your problem is though. What do you find the
specific differences are between the grain dissolver off and on?

Shaun

I refer to scanning slides mostly velvia and provia. Use it-8 targets
from Wolf. My pictures get diffused (unclear colours like a foggy day
in London) and sometimes unsharp with the grain dissolver off. In
previous versions the same thing happened when the grain dissolver was
on!

The histograms (shown in Vuescan) differ quite a bit in the two cases,
dissolver on and off. I believe something happens with the black point
even if I have it on 0 in both cases. Histograms look rather ok but
they look quite different! In the diffused (bad) case there are often
high spikes close to the black point for 0 and the rest is rather
flat. There seems to be another scale on the histogram!

In Photshop Elements I can correct most of the problem with levels.
According to the info I see in Elements the black pointer has to be
moved quite a bit to the right for getting clear and sharp colours.
I do not know if anybody understands my explanation but it was a try.
I am very confused.
Regards
ThomasD

 

[Shaun Markham]

I refer to scanning slides mostly velvia and provia. Use it-8 targets
from Wolf. My pictures get diffused (unclear colours like a foggy day
in London) and sometimes unsharp with the grain dissolver off. In
previous versions the same thing happened when the grain dissolver was
on!

The histograms (shown in Vuescan) differ quite a bit in the two cases,
dissolver on and off. I believe something happens with the black point
even if I have it on 0 in both cases. Histograms look rather ok but
they look quite different! In the diffused (bad) case there are often
high spikes close to the black point for 0 and the rest is rather
flat. There seems to be another scale on the histogram!

In Photshop Elements I can correct most of the problem with levels.
According to the info I see in Elements the black pointer has to be
moved quite a bit to the right for getting clear and sharp colours.
I do not know if anybody understands my explanation but it was a try.
I am very confused.
Regards
ThomasD

I think I understand your confusion. I would suggest if you haven't
already done so, delete the vuescan.ini file. This will reset all the
settings back to default and clear anything that may be causing the
problem - especially from previous versions of Vuescan. When I scanned an
IT8 target with versions 7.6.78 and backwards, black was reached at 18 on
the grey scale. This was when the grain dissolver was off with a measured
density of around only 2.0!

I too scan mostly velvia and provia. I have just scanned my IT8 target
with the dissolver on and the histogram for me shows spikes on the extreme
right in the red and green channels for both top and bottom graphs, the
bottom graph shows a strange red spike on the extreme left - visually the
scans look about the same to me. I mainly use a Wide Gamut RGB profile for
the output and monitor colour space along with my own generated ICC
profile for the scanner colour space. Colour balance is normally on Manual
or Auto Levels with black point on 0.1, white point at 0.02 and brightness
around 1.4 gives me the results I like. Not sure what other settings may
be giving the problems you describe.

I still normally tweak the levels again as you do as the black point is a
little to the right and this gives the punchy colours for my printer.

Shaun

 

[thomas]

I think I understand your confusion. I would suggest if you haven't
already done so, delete the vuescan.ini file. This will reset all the
settings back to default and clear anything that may be causing the
problem - especially from previous versions of Vuescan. When I scanned an
IT8 target with versions 7.6.78 and backwards, black was reached at 18 on
the grey scale. This was when the grain dissolver was off with a measured
density of around only 2.0!

I too scan mostly velvia and provia. I have just scanned my IT8 target
with the dissolver on and the histogram for me shows spikes on the extreme
right in the red and green channels for both top and bottom graphs, the
bottom graph shows a strange red spike on the extreme left - visually the
scans look about the same to me. I mainly use a Wide Gamut RGB profile for
the output and monitor colour space along with my own generated ICC
profile for the scanner colour space. Colour balance is normally on Manual
or Auto Levels with black point on 0.1, white point at 0.02 and brightness
around 1.4 gives me the results I like. Not sure what other settings may
be giving the problems you describe.

I still normally tweak the levels again as you do as the black point is a
little to the right and this gives the punchy colours for my printer.

Shaun

Thanks. I had not deleted the old vuescan.ini file! Deleting it seemed
to have solved my problems! I have also made new IT8 targets with and
without dissolver. I believe it should be clearly stated by Ed that
the .ini file should be deleted before upgrading, at least I missed
it. Upon your advice I also tested auto levels and it seems to work
better for me than white balance! Further I have increased brightness
to around 1.3 as default.
Regards and once more thanks.
Thomas

 

[Bart van der Wolf]

Hi,
First I would like to say that I appreciate Vuescan and the work done
by Ed. Second I have to admitt that still after 6 months trying I do
not know how to use it with my "new" Minolta Dimage 5400.
During the last months I have mainly had problems when using the
program with the grain dissolver on. For version 7.6.81 I do only get
good (better) results if the grain dissolver is on! Can anybody help
me with a good and stable workflow for Minolta 5400 with Vuescan. I am
totaly lost.

That's correct. I just tried 7.6.81 and only with the GD switched on it
produces acceptable (Red is out of sync with Green and Blue as the density
increases) results. Whether it's due to the new firmware that came with the
latest Minolta software upgrade, or something that Ed Hamrick changed, I
don't know.

With the GD switched off, dynamic range is still screwed up with featureless
dense areas, as I have complained about for a long time now. I can't get RGB
values lower than 50 with the film base locked to RGB 251.

Bart

 

[Bart van der Wolf]

SNIP

I found that I could not get any detail in the black regions of a
scan with grain dissolver switched off and I also found my prints were
not very good. Switching the grain dissolver on improved things a lot
but my colours were still very poor.

Yes, I have had the same experience ever since the 5400 was "supported"
(http://www.xs4all.nl/~bvdwolf/main/downloads/index.htm version 7.6.64 !),
and still do with version 7.6.81.

Since upgrading to 7.6.80 and 81, suddenly it works great with grain
dissolver both off and on and my workflow is now a lot simpler with
printed colours looking the way I expect.

I only get closer to acceptable behavior with the GD switched on. Switched
off there is no detail in dense areas.

So as far as I can tell the density calibration problem has been fixed
- thanks very much Ed - sanity restored.

Unfortunately both Thomas and myself don't share the same experience as you
do.

Bart

 

[Shaun Markham]

Yes, I have had the same experience ever since the 5400 was "supported"
(http://www.xs4all.nl/~bvdwolf/main/downloads/index.htm version 7.6.64 !),
and still do with version 7.6.81.

How did you generate those graphs?

OK, I did a little experiment to double check I wasn't doing something
strange. I reloaded 7.6.78 and deleted my vuescan.ini file. I did not
alter any settings from the default apart from resolution and media type
as image 48bit RGB.

With GD off I get,
http://photos.shaun.markham.name/test_scans/crop0001.jpg
I measured a density of 1.9 on grey 16

GD on,
http://photos.shaun.markham.name/test_scans/crop0002.jpg
I measured a density of 1.2 on grey 16 and 1.5 on 22


I now loaded 7.6.81 after deleting the vuescan.ini file again.

With GD off I now get,
http://photos.shaun.markham.name/test_scans/crop0003.jpg
I get a density of 1.5 on grey 16 and 2.5 on 22

GD on,
http://photos.shaun.markham.name/test_scans/crop0004.jpg
Density is 1.5 on 16 and 2.3 on 22 - similar to above. I'm not sure why I
don't see density figures over 4.0 though but visually it's ok.

With reference to your other post, I get
RGB values of 50 and below from grey area 17 to 22.

I'm not sure what this all means but something has definitely changed
since 7.6.78! I also notice that the histograms cover a much greater range.

If it's any help, I think my scanner is Rev 1.07 if I am looking in the
right place.

Shaun

 

[Bart van der Wolf]

How did you generate those graphs?

I made 6 scans/frames of a Kodak step tablet no.3, a neutral (silver based)
gray scale with known step densities. Of these Raw scans, with locked
exposure, the 16-bit brightnesses (linear gamma) were converted to relative
(to the brightest area) Optical Densities in MS Excel where they were
plotted and saved as files.

SNIP

If it's any help, I think my scanner is Rev 1.07 if I am looking in the
right place.

In the DiMAGE Scan utility (Windows version), hover the mouse pointer over
the top right panel where I read:
DiMAGE Scan 1.1.3
DS Elite5400 ROM Version = 1.10

Bart

 

[Ricky]

That's correct. I just tried 7.6.81 and only with the GD switched on it

produces acceptable (Red is out of sync with Green and Blue as the density
increases) results. Whether it's due to the new firmware that came with the
latest Minolta software upgrade, or something that Ed Hamrick changed, I
don't know.

With the GD switched off, dynamic range is still screwed up with featureless
dense areas, as I have complained about for a long time now. I can't get RGB
values lower than 50 with the film base locked to RGB 251.

Bart

Hi everyone,

I know this thread is a few days old now but I thought I'd share my
recent experiences of the 5400 with Vuescan.

After reading this thread I downloaded 7.6.81 (I had been using
7.6.76) and scanned a Provia IT8 target. With GD OFF, I got density
values of around 2.8 for the black square. With GD ON, I got values of
around 2.5.

I then went back to 7.6.76 and scanned the IT8 target. Exactly the
same experience, so for me I can get higher density values when using
Vuescan with the GD OFF. But...visually the scans look better
(sharper) when GD is on. This is not a good situation for me because I
can't use the Minolta software, I just cannot get it to focus. No
matter where I put the focus point, the ENTIRE scan is out of focus.
No such focus issues with Vuescan however (I have downloaded the
latest version of the Minolta software with Firmware update and it
behaves exactly the same as the older version regarding focus, or
rather lack of it).

Another problem for me with 7.6.81. I scanned a fairly dense Provia
slide of a sunset 4 times - GD/IR off, IR on (no GD), GD/IR on, GD (no
IR). I recalibrated after the first two scans after ticking GD (I'm
sure I read somewhere that it makes a difference). I then loaded each
scan into Photoshop and to my horror each scan had terrible green,
blue and magenta banding. The scanner had already been on quite a
while.

I then followed exactly the same procedure with exactly the same
settings in 7.6.76 with this slide. All four scans showed NO banding
whatsoever in any part of the scan.

I have a lot of slides to scan and I'm too scared to make a start in
case Ed Hamrick resolves these problems and I have to do them again,
although it has been quite a while since these problems were pointed
out and nothing has been done yet, so maybe he never will.

Ricky.

 

[Shaun Markham]

Hi everyone,

I know this thread is a few days old now but I thought I'd share my
recent experiences of the 5400 with Vuescan.

After reading this thread I downloaded 7.6.81 (I had been using
7.6.76) and scanned a Provia IT8 target. With GD OFF, I got density
values of around 2.8 for the black square. With GD ON, I got values of
around 2.5.

I then went back to 7.6.76 and scanned the IT8 target. Exactly the
same experience, so for me I can get higher density values when using
Vuescan with the GD OFF. But...visually the scans look better
(sharper) when GD is on. This is not a good situation for me because I
can't use the Minolta software, I just cannot get it to focus. No
matter where I put the focus point, the ENTIRE scan is out of focus.
No such focus issues with Vuescan however (I have downloaded the
latest version of the Minolta software with Firmware update and it
behaves exactly the same as the older version regarding focus, or
rather lack of it).

Some interesting information there Ricky.
This confuses me still further as I don't understand why I am getting
different results to what I read here from various people.

I generated some test histograms using version 7.6.78 and 7.6.81 with gd
on and off. Before each test I powered off the scanner and deleted the
vuescan.ini file. Scanner was then calibrated before scanning with default
values. The titles should be self explanatory.

http://photos.shaun.markham.name/test_scans/78_histo_gd_off.jpg
http://photos.shaun.markham.name/test_scans/78_histo_gd_on.jpg
http://photos.shaun.markham.name/test_scans/81_histo_gd_off.jpg
http://photos.shaun.markham.name/test_scans/81_histo_gd_on.jpg

These histograms were created with the image cropped following the black
box line surrounding the main IT8 target area.

Several things are obvious from these tests:
* The red and green channel suffer badly at the black/dense end with
version 78 and gd off
* Version 78 produces a much narrower histogram with gd on
* Version 78 produces very different graphs between gd on and off
* Version 81 produces similar graphs between gd on and off

In summary, I could not get decent images produced with version 78 or
indeed any version for the previous 7 months. Version 81 produces MUCH
better images for me. Obviously as has been shown before, YMMV.

Incidentally I use Vuescan on Linux/Firewire and my scanner uses 1.07
firmware.

Another problem for me with 7.6.81. I scanned a fairly dense Provia
slide of a sunset 4 times - GD/IR off, IR on (no GD), GD/IR on, GD (no
IR). I recalibrated after the first two scans after ticking GD (I'm
sure I read somewhere that it makes a difference). I then loaded each
scan into Photoshop and to my horror each scan had terrible green,
blue and magenta banding. The scanner had already been on quite a
while.

I then followed exactly the same procedure with exactly the same
settings in 7.6.76 with this slide. All four scans showed NO banding
whatsoever in any part of the scan.

Sounds a bit nasty. I can't comment on this as I have not experienced
banding on any version I have used to date.
Maybe I'm the lucky one for a change! It will be interesting to see what
7.6.82 brings....

Shaun

 

[Olaf Ulrich]

... and my scanner uses 1.07 firmware.

Move ahead to 1.10 firmware. Free download from
Minolta Europe's site www.konicaminoltaeurope.com.
It's hidden within the DiMAGE Scan Utility 1.1.3.

Olaf

 

[Bart van der Wolf]

SNIP

After reading this thread I downloaded 7.6.81 (I had been using
7.6.76) and scanned a Provia IT8 target. With GD OFF, I got
density values of around 2.8 for the black square. With GD ON,
I got values of around 2.5.

This is WAY too low. That is assuming the exposure was correct (non-clipped
whites), and based on my target's reference data (the file that accompanies
the target), but they should be similar to yours.

The XYZ_Y coordinate for patch GS23 of my Provia target tells me that its
Optical Density is Log10(100/0.10)=3.00, and the OD of patch GS0 is
Log10(100/77.36)=0.11 .
This means that the Relative Density difference from the VueScan Density
readings of those patches should be close to 3.00-0.11=2.89 .
On my target I get an average relative Density difference of 1.47 (!) with
the GD off, and 2.40 with the GD on. That too low Dmax is similar to my
experiences scanning a non-dye, but silver, density stepwedge. This does not
happen with the Minolta software.

One can plot the input densities (calculated OD from the XYZ_Y coordinate)
against the measured VueScan densities in an XY graph in e.g. MS Excel.
Ideally there should be a straight line relation between input and output.
With VueScan there still isn't :-(

You can copy the data from my scanner/target below, and paste it in Excel
(decimal point, comma separator), then produce an XY type Graph like the one
here:
http://www.xs4all.nl/~bvdwolf/main/downloads/SE5GD-VS7681.gif

Patch, OD, Dred, Dgreen, Dblue
GS0, 0.1115, 0.02, 0.05, 0.06
GS1, 0.2242, 0.14, 0.16, 0.16
GS2, 0.2779, 0.19, 0.22, 0.21
GS3, 0.3353, 0.24, 0.28, 0.27
GS4, 0.3959, 0.29, 0.34, 0.33
GS5, 0.4599, 0.35, 0.41, 0.39
GS6, 0.5196, 0.41, 0.47, 0.45
GS7, 0.5914, 0.47, 0.55, 0.52
GS8, 0.6643, 0.54, 0.62, 0.59
GS9, 0.7373, 0.61, 0.70, 0.66
GS10, 0.8392, 0.71, 0.81, 0.76
GS11, 0.9003, 0.76, 0.87, 0.81
GS12, 1.0088, 0.86, 0.98, 0.93
GS13, 1.0969, 0.94, 1.08, 1.00
GS14, 1.2090, 1.05, 1.18, 1.10
GS15, 1.3279, 1.14, 1.30, 1.23
GS16, 1.4622, 1.27, 1.44, 1.35
GS17, 1.6091, 1.41, 1.58, 1.48
GS18, 1.7570, 1.51, 1.71, 1.60
GS19, 1.9508, 1.70, 1.90, 1.77
GS20, 2.2076, 1.92, 2.10, 1.96
GS21, 2.3872, 2.07, 2.23, 2.09
GS22, 2.7447, 2.26, 2.41, 2.28
GS23, 3.0000, 2.40, 2.43, 2.34

Bart

 

[Shaun Markham]

One can plot the input densities (calculated OD from the XYZ_Y coordinate)
against the measured VueScan densities in an XY graph in e.g. MS Excel.
Ideally there should be a straight line relation between input and output.
With VueScan there still isn't :-(

FWIW here is my contribution with a Velvia IT8 target produced in the way
you describe.

http://photos.shaun.markham.name/test_scans/velvia_it8_gd_off.jpg
http://photos.shaun.markham.name/test_scans/velvia_it8_gd_on.jpg

There is room for improvement.

Anyone know what firmware V1.07 to V1.10 fixes/changes in the 5400?

Shaun.

 

[Robert Feinman]

One can plot the input densities (calculated OD from the XYZ_Y coordinate)
against the measured VueScan densities in an XY graph in e.g. MS Excel.
Ideally there should be a straight line relation between input and output.
With VueScan there still isn't :-(

I'm not sure why a non-straight curve is a problem. As long as all the
densities in the original are captured, all the data is available.
Characterising the scanner by making a profile for it is just what is
supposed to compensate for nonlinearities in both density and color
representation.
In my way of working I just make sure that whatever scanning software
I'm using does not clip values at either end and fix all the scanning
parameters to the values used for making the profile for all subsequent
scans.
I think too much is being asked of scanner software, use your image
editor.

 

[Bart van der Wolf]

I'm not sure why a non-straight curve is a problem. As long as all the
densities in the original are captured, all the data is available.

Dull shadows in slides (highlights in negatives) and loss of density range
because all available densitity differences get squashed in a smaller range.
A typical S-curve adjustment to spice up the midtone contrast will further
degrade those high density areas.

Characterising the scanner by making a profile for it is just what is
supposed to compensate for nonlinearities in both density and color
representation.

What's lost is lost, what can be restored is less accurate, more noisy.

In my way of working I just make sure that whatever scanning software
I'm using does not clip values at either end and fix all the scanning
parameters to the values used for making the profile for all subsequent
scans.
I think too much is being asked of scanner software, use your image
editor.

Yet my other scanners have no difficulty maintaining the original film
contrast (already compressed at the higher densities), which is part of its
characteristic. The Minolta Scan utility used to have no problem, VueScan
has lots of trouble but only (?) with the DSE5400.

Maybe the firmware upgrade to 1.10 that accompanied the Scan utility 1.1.3
changed something in that respect, because it seems that I now also have
more difficulty in getting the type of response I used to get:
http://www.xs4all.nl/~bvdwolf/main/downloads/MDSU.htm .

Bart

 

[Robert Feinman]

Maybe the firmware upgrade to 1.10 that accompanied the Scan utility 1.1.3
changed something in that respect, because it seems that I now also have
more difficulty in getting the type of response I used to get:
http://www.xs4all.nl/~bvdwolf/main/downloads/MDSU.htm .

Bart

Where did you get the firmware upgade? When I installed
the 1.1.3 from Minolta site (Europe) yesterday the scanner reported
firmware of 1.08 after the upgrade.

 

[Dierk Haasis]

Where did you get the firmware upgade? When I installed
the 1.1.3 from Minolta site (Europe) yesterday the scanner reported
firmware of 1.08 after the upgrade.

Try Minolta Europe, the firmware update is hidden within the software
upgrade, and it looks like it gets installed upon either the first
start of the new sw or it gets loaded whenever the sw is started. I've
not made up my mind, yet.

 

[Bart van der Wolf]

SNIP

Where did you get the firmware upgade? When I installed
the 1.1.3 from Minolta site (Europe) yesterday the scanner reported
firmware of 1.08 after the upgrade.

I think I accessed it through the Dutch Konica Minolta site, but it
eventually leads to an English language software download area, where I
selected the English version (Ds113e.exe for Windows XP). I had no idea it
updated the firmware, because I wasn't warned and I didn't notice any
special delay. However, hovering the mouse pointer in the top right area of
the utility reveals:
DiMAGE Scan 1.1.3
DS Elite5400 ROM version = 1.10

It is important to install with the scanner switched off, so maybe that's
what was different at your end?

Bart

 

[Fernando]

This is WAY too low. That is assuming the exposure was correct (non-clipped
whites), and based on my target's reference data (the file that accompanies
the target), but they should be similar to yours.

Any good news with the latest 7.6.83?
Was in the market for an Elite 5400, but Vuescan is a must for me
(Linux-based)... and getting good shadow dynamic range, too!
Did Ed Hamrick said something about this specific issue?

Best regards,

Fernando