PDA

View Full Version : Buying Color Perfect Vs farting around with vuescan?



Atracksler
18-Jul-2015, 08:48
Im just starting scanning my C41 negatives. Ive gotten great B/W Results with vuescan, but I'm not thrilled with the color.

Should I just bite the bullet and get Color Perfect for converting the negatives? or is there a more efficient vuescan workflow?

Pali K
18-Jul-2015, 11:26
Try the free version which will add lines to your scan but at least you'll know what to expect.

I find color perfect to be the best and it is near flawless in all auto mode for ektar. I have to tweak Portra a bit afterwards but it comes really close.

Pali

richardman
18-Jul-2015, 12:26
See my post just yesterday on my experience on using Silverfast.

SergeyT
19-Jul-2015, 02:26
Everything that color perfect does can be relatively easy replicated by curves in PS. CP is just a plugin which uses the data loaded into PS.
Use it on a copy of your neg scanned as positive in demo to produce a reference if you wish and then invert your original and go for the same look with a few curves layers - one for dark point, one for white point, one for tones and one for white balance fine tuning.

SergeyT.

alkklab
19-Jul-2015, 10:28
Color Perfect plugin has significantly improved my scans. Scanning raw in Vuescan and then converting Portra 160 with it, the negatives are are almost always looking good with minimal photoshop corrections. Never had luck working with Silverfast.

Kirk Gittings
19-Jul-2015, 10:44
Everything that color perfect does can be relatively easy replicated by curves in PS. CP is just a plugin which uses the data loaded into PS.
Use it on a copy of your neg scanned as positive in demo to produce a reference if you wish and then invert your original and go for the same look with a few curves layers - one for dark point, one for white point, one for tones and one for white balance fine tuning.

SergeyT.

I tested this years ago and came to the same conclusion. Maybe it does something special now?

adelorenzo
19-Jul-2015, 12:43
Yes I have found the color conversion in Color Perfect to be really good and sometimes better than what I can manage on my own. However, I rarely use it because of what it does to my workflow.

My normal workflow looks like this:

1. Scan with Epson scan software, individual frames -> TIFF
2. Import into Lightroom
3. Edit as desired
4. Export

Using Color Perfect, it starts to look like this:

1. Scan with Vuescan, strips of frames -> Raw TIFF (Because Vuescan sucks so badly at scanning multiple frames)
2. Open with Photoshop Elements, crop each individual frame and put it into a new file
3. Run ColorPerfect on each file
4. Save each file as TIFF
5. Import into Lightroom
6. Edit as desired
7. Export

Steps 2-4 are incredibly time consuming. I hate digital editing to begin with so I rarely bother.

Preston
19-Jul-2015, 15:30
I tried Color Perfect some time ago and found it fiddly. When I do scan color neg (w/VueScan) I check each of the installed profiles to see which looks best and then save the TIFF file. Once in Photo Shop, I save the file a *.PSD.

Works for me.

--P

rbultman
19-Jul-2015, 19:27
Check out the Vuescan advanced workflow for establishing the film base color for color negative film. It works well and does a good job removing the case colour.

Vuescan does suck at scanning multiple frames even on my LS-8000. The original Nikon software was very good and could automatically find the frames in a strip of 120. Not sure why Vuescan can't. It also doesn't handle inconsistent spacing on 120. I pretty much scan one frame at a time, except for 35mm.

Alan Klein
19-Jul-2015, 19:48
What scanner do you use?

rbultman
20-Jul-2015, 03:01
I use a Nikon LS-8000.

This video (https://www.youtube.com/watch?v=_pfQ61WTbug) covers how to use the 'Lock Exposure' and 'Lock Film base Color' features of Vuescan in order to remove the orange mask. You only need to watch the first 10 minutes or less. This method is effective. A flatbed scanner is used in the video but the method works for dedicate scanners too.

Note that Vuescan is not really a photo editor. It does the scan and just gets the colors close. You still need to adjust the color balance in PS, Lightroom, or some other editor.

towolf
20-Jul-2015, 05:08
Vuescan does suck at scanning multiple frames even on my LS-8000.

I recently used a Nikon CoolScan 9000 in an art academy, which only had SilverFast on an OSX box. *That* was horrible regarding frames and batching. So much time wasted!

VueScan scanning long rolls of 35mm on a Nikon CoolScan 4000 can be done fully batched, but with black borders left and right. The frame detection goes wrong every so often and would scan a frame with a black bar on one side. I now just scan with maximum width in the preview and crop the bars later in Adobe Camera Raw plugin. In ACR just crop them all with the proper width and then shift the crop box left and right in each one for compensate for inconsistent spacing. Goes very fast.

I’ve tried a lot of stuff but in the end I settled on the "Lock exposure" + "Lock Film Base color" zeroing method and in Output save as 48-bit TIFF (with Gamma). Then just edit the files with ACR or Lightroom. The TIFFs will be kept as "originals" and the edits are stored non-destructively. Early on I wanted to have linear files (with no Gamma) and tried "RAW TIFF" and "RAW DNG", but somehow Adobe would always interpret them wrong (too dark).

towolf
20-Jul-2015, 05:17
This video (https://www.youtube.com/watch?v=_pfQ61WTbug) covers how to use the 'Lock Exposure' and 'Lock Film base Color' features of Vuescan in order to remove the orange mask. You only need to watch the first 10 minutes or less.

Just watched the video until he gets to the "Color" tab, and what he says there about "Curve low" and "Curve high" is utterly wrong. He says to set the control points on the curve both to 0.001 to "capture the whole range". The odd thing with Vuescan is, that those control points are "zeroed" at 0.25 and 0.75 respectively. Heís essentially applying a curve right there.

rbultman
20-Jul-2015, 05:37
Just watched the video until he gets to the "Color" tab, and what he says there about "Curve low" and "Curve high" is utterly wrong. He says to set the control points on the curve both to 0.001 to "capture the whole range". The odd thing with Vuescan is, that those control points are "zeroed" at 0.25 and 0.75 respectively. Heís essentially applying a curve right there.

I can't really comment on whether or not that applies a curve, but that is exactly what I have been doing. I followed Ken Lee's scanning tips (http://www.kenleegallery.com/html/tech/scanning.php) which includes this same advice, namely setting the curve low and high to the minimum. When I did this, my scans improved dramatically. I was already doing most of the other things, so I feel that these settings were critical to improving my scans.

Here is the relevant section from the Vuescan user's manual on Curve Low and High:

Use these options to change the shape of the curve that's applied after the black/white point is applied.

Decrease the low value to see more detail in the black parts of the image, and increase the high value to see more detail in the white part.

So I infer from that two things, namely setting the points modifies the shape of the curve and that a curve is applied regardless. My take-away is that all you can do is adjust the curve that WILL be applied.

This is what Ken says about Curve Low and High:

The default VueScan Preview settings have made the image appear quite harsh: the White Point setting of 1 and the default Curve low and Curve high settings have clipped the low and high tonal values. Light gray tones appear pure white - like chalk. The dark tones have been forced to pure black.
...
Because we want to get a straight scan from VueScan, we need to adjust the Curve low and Curve high settings to 0, or as low as they can go. Having these settings at Zero will ensure that the extremes of the density curve are not clipped unless we want them to be.


This echoes my feeling about Vuescan - it is a capture program with negative-to-positive conversion, it is not an editing tool. Get as much data as you can form the negs or slides, then edit in another program. Some people do negative conversion in PS, but that makes my head hurt. Actually PS makes my head hurt. I like more idiot-oriented tools like Lightroom.

Regards,
Rob

towolf
20-Jul-2015, 05:51
The funny thing is, anything in the "Color" tab is disregarded entirely when you save as RAW. So it’s essentially cargo culting to fiddle with those settings and then to save as RAW. However, if you want to save as raw *and* have positive files, then you can tick "Raw save as film" and then the settings on the Color tab *are* applied. This stuff is highly confusing and I blame the Hamricks for that.

But I didn’t want to save as RAW anyway because I’m pretty sure Adobe (Lightroom or ACR) interprets the linear values wrongly. I think it assumes Gamma, even though there is no pre-multiplied Gamma in the RAW TIFF.

In the Color tab I just zero everything. And zeroing the curve is to keep the values for Curve at 0.25 and 0.75. You can even see it, the curve is straight.

About Lightroom, I don’t really like that I have to import the TIFFs in its database to make edits. I rather browse the output folder with Bridge and then open the TIFF in the Camera Raw dialog, which offers essentially the same editing controls minus some bells and whistles. The problem with negative files in Lightroom is, that it assumes positive files. All the controls are oriented towards positive files. Of course you can apply an "upside-down" curve, but that curve comes later in the processing chain.

Hence I prefer to save as normal 48-bit TIFF, not RAW, to get positive files that are interpreted correctly, and do the edits in ACR. And set everything on "no operation" in the Color tab in VueScan, i.e., no Color Balance, Film Base Color on [1,1,1], and Generic Color Negative.

towolf
20-Jul-2015, 06:08
I followed Ken Lee's scanning tips (http://www.kenleegallery.com/html/tech/scanning.php) which includes this same advice.

About that site, I found that quite a few tips on that site are outdated or sometimes not entirely correct.

For example, this page explains how to single out the green channel (http://www.kenleegallery.com/html/tech/scanningGreen.php), since that is ostensively sharper, by scanning in RGB and singling out the green channel.

The thing is, this is what Vuescan does by default (https://www.hamrick.com/vuescan/html/vuesc29.htm#inputmakegrayfrom):


The default, "Auto", either uses the scanner hardware to convert from the color CCD to gray, or converts in VueScan, mostly from the green channel.

rbultman
20-Jul-2015, 15:03
The funny thing is, anything in the "Color" tab is disregarded entirely when you save as RAW. So itís essentially cargo culting to fiddle with those settings and then to save as RAW. However, if you want to save as raw *and* have positive files, then you can tick "Raw save as film" and then the settings on the Color tab *are* applied. This stuff is highly confusing and I blame the Hamricks for that.

Roger that (emphasis added.)


But I didnít want to save as RAW anyway because Iím pretty sure Adobe (Lightroom or ACR) interprets the linear values wrongly. I think it assumes Gamma, even though there is no pre-multiplied Gamma in the RAW TIFF.

In the Color tab I just zero everything. And zeroing the curve is to keep the values for Curve at 0.25 and 0.75. You can even see it, the curve is straight.

Noted. (Emphasis added.) I see the curve but when the settings are 0.001 for both, it only seems to affect the highlight side of the curve. When I reset low and high by double-clicking the curve, the image I see on the screen is a little flatter and duller.


Hence I prefer to save as normal 48-bit TIFF, not RAW, to get positive files that are interpreted correctly, and do the edits in ACR. And set everything on "no operation" in the Color tab in VueScan, i.e., no Color Balance, Film Base Color on [1,1,1], and Generic Color Negative.

So, I just checked my settings and I do let Vuescan do the conversion and I save as TIFF DNG. I have "TIFF file", "TIFF DNG format", and TIFF profile" checked. This means that the curve is applied, right, because the TIFF is not RAW so it has the settings on the color tab applied?

Everything else I have is identical to yours. I can't currently check the film base color settings as I am scanning some B&W and don't have base color checked.

towolf
20-Jul-2015, 18:14
When I reset low and high by double-clicking the curve, the image I see on the screen is a little flatter and duller.

Sure, it’s dull. But didn’t you say, that you wanted to leave the edits for later? It’s dull for me too, but I can apply a more accurate curve later, where the controls are not limited to two control points with strange inputs (i.e., 0.25 and 0.75).


So, I just checked my settings and I do let Vuescan do the conversion and I save as TIFF DNG. I have "TIFF file", "TIFF DNG format", and TIFF profile" checked. This means that the curve is applied, right, because the TIFF is not RAW so it has the settings on the color tab applied?

So, Vuescan has as output image formats: JPG, TIFF, and RAW

RAW can save as RAW TIFF and as RAW DNG. RAW DNG is basically the TIFF in a (simple) DNG wrapper. The point of RAW is to save the values from the scanner without modification to file. This would mean equi-distant linear values and a negative image.

And normally linear image data would be best for edits, since many image calculations and operations "assume" linear values and are more correct when performed on linear data. The problem comes in when there is confusion in whether the data is already/still linear or not.

Now, JPEG and TIFF are for saving regular images with a traditional Gamma compression (https://en.wikipedia.org/wiki/Gamma_correction#Explanation) (i.e., raise values to the power of 1/2.2 to expand the lower tonal values). TIFF allows doing this at 16-bit (which lessens the problem of gamma compression for edits). Problem with TIFF is that I don’t think it is standardized in-how-far Gamma is applied or not, i.e., software has to guess, I think. Saving JPG and TIFF will also invert the values to make a positive, and apply anything on the Color tab.

Now, the test is the following: Write in two formats at the same time, RAW and JPG, and view them side-by-side in Lightroom or Bridge (I mathematically inverted the values on the command line, but the same is visible when you click "Raw save film" (https://www.hamrick.com/vuescan/html/vuesc33.htm#outputrawsavefilm), which writes RAW as inverted positive). They should look the same, but they don’t. The (linear) RAW is compressed on the black end. I think it is decompressed when it shouldn’t be, making the low register of the histogram "too dark". This is true for both RAW DNG and RAW TIFF.

I really wanted to use RAW DNG for a long time, but it just didn’t work as I thought it should. I just made my life a lot easier by saving to regular 16-bit per channel TIFF. It is supported in ACR and Lightroom just fine and edits are saved non-destructively. And it is supported in any other software out-of-the-box as well.

The thing with "Film base color" is that only once you tick "Lock Film base color" as well it is really locked. For batch scanning this has the advantage that no preview scan is necessary, which cuts down the time to run through the batch (This is mostly useful for long 35mm film strips on a Coolscan).

towolf
20-Jul-2015, 18:22
So, I just checked my settings and I do let Vuescan do the conversion and I save as TIFF DNG. I have "TIFF file", "TIFF DNG format", and TIFF profile" checked. This means that the curve is applied, right, because the TIFF is not RAW so it has the settings on the color tab applied?

Right, I forgot the final option, which is "regular gamma compressed TIFF, but in a DNG wrapper", which is what you say you are using. Iím no sure what benefit this DNG wrapper gives, but from my earlier experiences with DNG ("I think it should have the metadata that specifies it is linear, but why doesnít it work?") I just stopped wrapping in DNG and just went for TIFF. I see no benefit in DNG and it just makes everything more opaque.

The whole Output tab in VueScan is a complicated nightmare.

Anyhow, ACR can store its non-destructive edits inside of the TIFF file as this Adobe XMP stuff, and that is all I need.

Hereís an example from the end of a TIFF file from today:



<x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="Adobe XMP Core 5.3-c007 1.136881, 2010/06/10-18:11:35 ">
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
<rdf:Description rdf:about=""
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:xmp="http://ns.adobe.com/xap/1.0/"
xmlns:xmpMM="http://ns.adobe.com/xap/1.0/mm/"
xmlns:stEvt="http://ns.adobe.com/xap/1.0/sType/ResourceEvent#"
xmlns:crs="http://ns.adobe.com/camera-raw-settings/1.0/"
dc:format="image/tiff"
xmp:CreatorTool="VueScan 9 x64 (9.5.06)"
xmp:CreateDate="2015-07-20T21:01:17"
xmp:Rating="0"
xmp:MetadataDate="2015-07-20T21:27:12+02:00"
xmpMM:DocumentID="EC8DDA722C18433F1F81DCAD21EC5817"
xmpMM:OriginalDocumentID="EC8DDA722C18433F1F81DCAD21EC5817"
xmpMM:InstanceID="xmp.iid:F49D7E52152FE511860DBD5BEE8B4EBC"
crs:RawFileName="2015-07-20-0001.tif"
crs:Version="6.7"
crs:ProcessVersion="5.7"
crs:Exposure="0.00"
crs:Shadows="10"
crs:Brightness="0"
crs:Contrast="+48"
crs:Sharpness="100"
crs:LuminanceSmoothing="0"
crs:VignetteAmount="0"
crs:FillLight="0"
crs:HighlightRecovery="0"
crs:Clarity="+75"
crs:SplitToningShadowHue="0"
crs:SplitToningShadowSaturation="0"
crs:SplitToningHighlightHue="0"
crs:SplitToningHighlightSaturation="0"
crs:SplitToningBalance="0"
crs:ParametricShadows="0"
crs:ParametricDarks="0"
crs:ParametricLights="0"
crs:ParametricHighlights="0"
crs:ParametricShadowSplit="25"
crs:ParametricMidtoneSplit="50"
crs:ParametricHighlightSplit="75"
crs:SharpenRadius="+0.7"
crs:SharpenDetail="25"
crs:SharpenEdgeMasking="0"
crs:PostCropVignetteAmount="0"
crs:GrainAmount="0"
crs:LensProfileEnable="0"
crs:LensManualDistortionAmount="0"
crs:PerspectiveVertical="0"
crs:PerspectiveHorizontal="0"
crs:PerspectiveRotate="0.0"
crs:PerspectiveScale="100"
crs:Exposure2012="-1.00"
crs:Contrast2012="+31"
crs:Highlights2012="0"
crs:Shadows2012="0"
crs:Whites2012="0"
crs:Blacks2012="-1"
crs:Clarity2012="+37"
crs:ToneCurveName="Linear"
crs:ToneCurveName2012="Custom"
crs:LensProfileSetup="LensDefaults"
crs:HasSettings="True"
crs:CropTop="0"
crs:CropLeft="0.049689"
crs:CropBottom="0.996689"
crs:CropRight="0.99931"
crs:CropAngle="0"
crs:CropConstrainToWarp="1"
crs:HasCrop="True"
crs:AlreadyApplied="False">
<dc:description>
<rdf:Alt>
<rdf:li xml:lang="x-default">scanned using CoolScan 4000ED in VueScan</rdf:li>
</rdf:Alt>
</dc:description>
<xmpMM:History>
<rdf:Seq>
<rdf:li
stEvt:action="saved"
stEvt:instanceID="xmp.iid:F49D7E52152FE511860DBD5BEE8B4EBC"
stEvt:when="2015-07-20T21:27:12+02:00"
stEvt:softwareAgent="Adobe Photoshop Camera Raw 6.7 (Windows)"
stEvt:changed="/metadata"/>
</rdf:Seq>
</xmpMM:History>
<crs:ToneCurve>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurve>
<crs:ToneCurveRed>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurveRed>
<crs:ToneCurveGreen>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurveGreen>
<crs:ToneCurveBlue>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurveBlue>
<crs:ToneCurvePV2012>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>17, 26</rdf:li>
<rdf:li>34, 44</rdf:li>
<rdf:li>51, 60</rdf:li>
<rdf:li>68, 74</rdf:li>
<rdf:li>85, 89</rdf:li>
<rdf:li>102, 105</rdf:li>
<rdf:li>119, 120</rdf:li>
<rdf:li>136, 135</rdf:li>
<rdf:li>153, 151</rdf:li>
<rdf:li>170, 167</rdf:li>
<rdf:li>187, 183</rdf:li>
<rdf:li>204, 200</rdf:li>
<rdf:li>221, 218</rdf:li>
<rdf:li>238, 236</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurvePV2012>
<crs:ToneCurvePV2012Red>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurvePV2012Red>
<crs:ToneCurvePV2012Green>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurvePV2012Green>
<crs:ToneCurvePV2012Blue>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurvePV2012Blue>
</rdf:Description>
</rdf:RDF>
</x:xmpmeta></pre>