Choosing between Epson V550 and Canon 9000F

Alan, when you open the Configuration tab in Epsonscan the Color tab has an exposure adjustment that does not grey out when you select No Color Adjustment. I've only scanned with it at recommended but wonder if moving it toward one extreme or the other will affect the scan with no color correction selected, it should with either of the adjustment options selected.

shutterfinger said:

Alan, when you open the Configuration tab in Epsonscan the Color tab has an exposure adjustment that does not grey out when you select No Color Adjustment. I've only scanned with it at recommended but wonder if moving it toward one extreme or the other will affect the scan with no color correction selected, it should with either of the adjustment options selected.

Click to expand...

I leave the gamma at 2.2 and the auto exposure in the middle because I don't know what else to do. I;ve got enough troubles with scanning without introducing more unknowns.

Maybe some else could explain gamma and the other setting. I just don't know.

Alan, if you are waiting on an answer from me, please let me get home tonight to answer. Work is blowing-up on me right now...
Cheers,

Kino said:

Yes it does make a difference; it places the data values in a different place and affects how you can subsequently manipulate the data produced. That being said, it largely depends upon the target you are scanning; will the image need drastic intervention or is it already within the dynamic range needed to produce a good image?

Click to expand...

Ok then it depends on what you mean by making a difference. For the epson scanners choosing a BP or WP has nothing to do with loosing any data or getting a better actual scan. With a 16bit number you have integers from 0-65535 to choose from, these devices use a 16bit (or even less) Analogue to Digital chip. What this means is they take the voltage from the CCD which varies in a linear manner to the amount of photons that hit the CCD in a certain period of time, to number between 0 and 65535. Now for each of those number you can move them around to your hearts content, using math, and as long as you use consistent rounding rules and your manipulations do not exceed 0 and 65535, nothing is lost.

Now if you try and squeeze those number from 0 and 65535 to 0-255 you will most certainly loose precision.

You can test yourself with a pocket calculator. Your could use following formula as example,

Output = Lift + Gain x Input Gamma

Some older high end scanners like some drum scanners use a log amplifier to adjust the voltage that comes from the CCD/PMT (sensor) before it is processed by the Analog to Digital chip(A/D). This approach was used when 16bit A/D and 16bit storage was too expensive. This means that the voltage signal that comes from the sensor is no longer linear in relationship to the light hitting it. This allows a 12bit A/D to rival the output of 16bit A/D. With these type of devices it is important to adjust the gain on the log amplifier correctly BEFORE the scan.

Similarly if you are going to use your Epson scanner to output 8bit files then you must set your BP/WP and all your corrections BEFORE you save the file.

One final point to remember HOWEVER is the software controls in epsonscan are unique to epson. They don't publish what they are or how they work, which is the same for any Adobe product BTW. That doesn't mean they are very similar or indeed the same, you may like the way the epsonscan works better...

Kino said:

Alan, if you are waiting on an answer from me, please let me get home tonight to answer. Work is blowing-up on me right now...
Cheers,

Click to expand...

Sure. Take your time. I'm not going anywhere. I'm retired.

Ted Baker said:

Ok then it depends on what you mean by making a difference. For the epson scanners choosing a BP or WP has nothing to do with loosing any data or getting a better actual scan. With a 16bit number you have integers from 0-65535 to choose from, these devices use a 16bit (or even less) Analogue to Digital chip. What this means is they take the voltage from the CCD which varies in a linear manner to the amount of photons that hit the CCD in a certain period of time, to number between 0 and 65535. Now for each of those number you can move them around to your heart content, using math, and as long as you use consistent rounding rules and your manipulations do not exceed 0 and 65535, nothing is lost.

Now if you try and squeeze those number from 0 and 65535 to 0-255 you will most certainly loose precision.

You can test yourself with a pocket calculator. Your could use following formula as example,

Output = Lift + Gain x Input Gain

Some older high end scanners like some drum scanners use a log amplifier to adjust the voltage that comes from the CCD/PMT (sensor) before it is processed by the Analog to Digital chip(A/D). This approach was used when 16bit A/D and 16bit storage was too expensive. This means that the voltage signal that comes from the sensor is no longer linear in relationship to the light hitting it. This allows a 12bit A/D to rival the output of 16bit A/D. With these type of devices it is important to adjust the gain on the log amplifier correctly BEFORE the scan.

Similarly if you are going to use your Epson scanner to output 8bit files then you must set your BP/WP and all your corrections BEFORE you save the file.

One final point to remember HOWEVER is the software controls in Epsonscan are unique to Epson. They don't publish what they are or how they work, which is the same for any Adobe product BTW. That doesn't mean they are very similar or indeed the same, you may like the way the Epsonscan works better...

Click to expand...

When I first began to scan, I used auto adjusts in the Epson. Then I tried scanning flat and adjusting in PS Elements by hitting Auto Levels on the flat scanned image file. I could not tell the difference between the auto scan results and the flat scanned Elements Auto levels results. They were equal. Color and exposure looked the same for the most part. (Slide film. I didn't try color negative film)

So that's why I ask if it really helps to adjust the BP and WP on the histogram in Epsonscan? Even though the resultant image appears the same then the alternative method, are there some sort of additional bits in there that provide a better results????

It doesn't matter if the BP and WP are set in scan software or post software as long as that are at the edges of the histogram or beyond as you loose no image detail in either. Now if you set the BP and WP inside the histogram then you loose image detail.

The auto exposure control is for compensating for incorrectly exposed negatives/positives. In silverfast it says exposure and goes 3 stops either side of neutral/correct exposure. I do not know if Epson's marks are in stops or not or if the low/high are the equivalent of under exposed/over exposed or not as I have not done any critical scanning with this software. The way to find out is to take an known incorrect negative or positive, scan a recommended then scan at each mark on the scale, compare the scans and see which produces the best results. It likely will not matter with No Color Correction selected.

Ted Baker said:

Ok then it depends on what you mean by making a difference. For the epson scanners choosing a BP or WP has nothing to do with loosing any data or getting a better actual scan. With a 16bit number you have integers from 0-65535 to choose from, these devices use a 16bit (or even less) Analogue to Digital chip. What this means is they take the voltage from the CCD which varies in a linear manner to the amount of photons that hit the CCD in a certain period of time, to number between 0 and 65535. Now for each of those number you can move them around to your hearts content, using math, and as long as you use consistent rounding rules and your manipulations do not exceed 0 and 65535, nothing is lost.

Now if you try and squeeze those number from 0 and 65535 to 0-255 you will most certainly loose precision.

Click to expand...

Ted,
Sorry, I used 8 bit as shorthand, as so many books on digital imaging tend to do, to speed up the discussion; I should have noted that in my response. I was in no way advocating using 8 bits for scanning, as mach banding artifacts can occur! So if I used 0-255, you should mentally slot in the corresponding 0-65535. It's just easier to use the lower bit value and extrapolate...

Ted Baker said:

You can test yourself with a pocket calculator. Your could use following formula as example,

Output = Lift + Gain x Input Gamma

Some older high end scanners like some drum scanners use a log amplifier to adjust the voltage that comes from the CCD/PMT (sensor) before it is processed by the Analog to Digital chip(A/D). This approach was used when 16bit A/D and 16bit storage was too expensive. This means that the voltage signal that comes from the sensor is no longer linear in relationship to the light hitting it. This allows a 12bit A/D to rival the output of 16bit A/D. With these type of devices it is important to adjust the gain on the log amplifier correctly BEFORE the scan.

Click to expand...

Yes, the original Kodak Cineon file format (.CIN and later .DPX) for digital cinema was only 10 bits to avoid that very problem.

Ted Baker said:

Similarly if you are going to use your Epson scanner to output 8bit files then you must set your BP/WP and all your corrections BEFORE you save the file.

One final point to remember HOWEVER is the software controls in epsonscan are unique to epson. They don't publish what they are or how they work, which is the same for any Adobe product BTW. That doesn't mean they are very similar or indeed the same, you may like the way the epsonscan works better...

Click to expand...

Agreed, I have never attempted to confirm that the gammas I quoted are the exact actual output, but it would make perfect sense with the long history of electronic imaging science that goes back to CRTs...

Alan Edward Klein said:

So that's why I ask if it really helps to adjust the BP and WP on the histogram in Epsonscan? Even though the resultant image appears the same then the alternative method, are there some sort of additional bits in there that provide a better results????

Click to expand...

I think I gave the detailed answer earlier... If you want a simple answer, then in your specific case, i.e. scanning a transparency, saving 16bit, there are no additional bits to be gained from epsonscan. All controls in epsonscan are POST scanning controls. Vuescan and silverfast both offer controls that actually affect the scan itself. i.e. exposure time and multi exposure.

Silverfast offers a lot more POST scanning features. I just use vuescan it's cheap, and use the manual exposure feature and run the scans through my own software.

Alan Edward Klein said:

I leave the gamma at 2.2 and the auto exposure in the middle because I don't know what else to do. I;ve got enough troubles with scanning without introducing more unknowns.

Maybe some else could explain gamma and the other setting. I just don't know.

Click to expand...

OK, lets see if I can navigate the gamma minefield without making too many errors! Bear with me as I tend to ramble all over the place and I am going to generalize, so it's going to no doubt drive some people crazy with my fast and loose analogies...

The scanner sensor itself has a Gamma of "1" or unity, a lot like a piece of film with a latent image upon it. It only sees the scene it totality from brightest to darkest within it's ability to resolve.

Until you develop the film or electronically impose a LUT or Image Transform Matrix upon the Raw Scan, it will remain a "1".

Put the film in developer or impose a LUT on the raw data and you impose a Gamma upon the signal or the latent image via development and the relationships between the values will be redefined by compaction or expansion.

Traditional "normal" film development pegs a negative development gamma at .65 gradient, or for every 1 unit of energy that falls on the film, .65 density will occur on the straight line portion of a plotted characteristic curve of densities (such as occur on a calibrated step wedge). A 1 stop push is .75 gamma and a 2 stop push is at .85 gamma. Conversely, a 1 stop pull is .45 gamma.

Since film development and printing, or film scanning and manipulation,is designed to bring the image back to represent the initial scene values, one would think that we would attempt to print a .65 gamma negative with a paper/developer combo that would bring it back to the original 1.0, but it was found that 1.0 was far too low in contrast for pleasing visual contrast. Through experimentation and testing, it was determined that end result gammas that were most pleasing to people were roughly in the 1.4 to 1.8 gamma range.

You can arrive at end result gamma by multiplying your gamma chain of reproduction:

.65 gamma original x 2.20 gamma paper/developer combo = 1.43 gamma end result.
.65 gamma original x 2.40 gamma = 1.56 and so on..

Strangely enough, when CRT tubes were being developed for television, they found that the typical gamma response of the camera target tubes was very close to film gamma and that if they biased the tube to a gamma of about 2.20, pleasing images resulted.

Long story short, when computer monitors were CRT based, the 2.20 legacy (yes there were exceptions) was retained and by the time LDC and other displays came in, it was required to retain this gamma response characteristic to retain backward compatibility with existing electronic images.

OK, so what does that have to do with the 2.20 gamma figure in the scanner software? This is a Look Up Table (LUT) adjustment to bias the gamma response of the digital data as it is rendered into the file. In one way of thinking, it's the grade of the digital paper you are using; lower numbers are lower contrast and higher numbers are higher contrast.

However, you must remember that you are taking a digital image of a negative that is being inverted to a positive OR a positive image that is being translated into what the software developer has determined is a "pleasing" image contrast from a 1.0 gamma original; often a rendition of statistically normal pixel values within what is defined by the histogram.

So, you can basically vary the gamma control to whatever is pleasing to your taste and not worry about it if you get good results, but if you get weird results, it's because the LUT doesn't map your image data pleasingly.

The only way you can offset this is by manipulating the dmax, dmin and pixel value mapping PRIOR to the positive/negative LUT being imposed. The only way you can see this take effect is to change controls prior to the scan, make a scan and then evaluate. I tend to use the curves tool to elevate my dmax and lower my dmin values, as well as expand my midtones with a "balloon" type curve to create a lighter, low contrast image that can be corrected later (also with curves) in Photoshop. You can't recover graduations that are not clearly demarked by code values without heavily degrading the image, but you can make substantial corrections if there are code values there to manipulate.

I hope this makes sense.

Kino said:

OK, lets see if I can navigate the gamma minefield without making too many errors!

Click to expand...

Respectively there is a lot of info there, but also a few significant errors that will lead people up the garden path. Gamma is a complex topic because IT IS used for more than one purpose in the imaging chain. The errors in this post are because you have mixed the different purposes of using a gamma other than 1. Publications that refer to Video/TV etc will explain gamma in one context, analogue photography another, modern digital in another, hybrid photography etc. Hybrid photography has the worst explanations IMHO.

A reasonable summary:

In 2018, all your files that wish to open directly and don't have a colorprofile assigned should be saved with sRGB gamma, or a value of 2.2 which is close enough. (this is the default or assumed value BTW). 16 bit Files that you intend to open with something like Photoshop and first manipulate can be saved with whatever gamma you like as long the program reading the file knows what the correct value is. You can even use the wrong value and correct it in POST

Ted Baker said:

I think I gave the detailed answer earlier... If you want a simple answer, then in your specific case, i.e. scanning a transparency, saving 16bit, there are no additional bits to be gained from epsonscan. All controls in epsonscan are POST scanning controls. Vuescan and silverfast both offer controls that actually affect the scan itself. i.e. exposure time and multi exposure.

Silverfast offers a lot more POST scanning features. I just use vuescan it's cheap, and use the manual exposure feature and run the scans through my own software.

Click to expand...

So I should check No Color Correction and forget about WP and BP for the scan???

Ted Baker said:

Respectively there is a lot of info there, but also a few significant errors that will lead people up the garden path. Gamma is a complex topic because IT IS used for more than one purpose in the imaging chain. The errors in this post are because you have mixed the different purposes of using a gamma other than 1. Publications that refer to Video/TV etc will explain gamma in one context, analogue photography another, modern digital in another, hybrid photography etc. Hybrid photography has the worst explanations IMHO.

Click to expand...

Yes, I mixed Gamma definitions across boundaries, but I warned the reader in advance my methods would be fast and loose.

I am not Charles Poynton, nor will I ever be.

Ted Baker said:

A reasonable summary:

In 2018, all your files that wish to open directly and don't have a colorprofile assigned should be saved with sRGB gamma, or a value of 2.2 which is close enough. (this is the default or assumed value BTW). 16 bit Files that you intend to open with something like Photoshop and first manipulate can be saved with whatever gamma you like as long the program reading the file knows what the correct value is. You can even use the wrong value and correct it in POST

Click to expand...

Yes, you can do all this, but if you scan a problematic negative and your values are insufficient in the areas with the most information, it will require a vast amount of tinkering and digital patchwork to get truly good results.

To each their own.

Follow up on my comment earlier.
Here is a set from my V600 (not a great fan but all I have).
One image is without crop when analysing is done by the scanner and the other is with so the data is analysed within.
I think one can do as well manually. Alan I think you may be able to save yourself some time doing the correction on the scanner

ced said:

Follow up on my comment earlier.
Here is a set from my V600 (not a great fan but all I have).
One image is without crop when analysing is done by the scanner and the other is with so the data is analysed within.
I think one can do as well manually. Alan I think you may be able to save yourself some time doing the correction on the scanner

View attachment 202147

Click to expand...

I found I was wasting more time adjusting before the scan. What happened was if I did not like the results, I had to rescan which takes a lot of time. If I scan flat with no adjustments, or only adjust WP and BP's, I never have to re-scan again. I save the original scan and never edit the original. I can play with copies all I want with Lightroom or Elements or other post-scan editing programs.

On the other hand, in the beginning when I scanned a bunch of old slides I had, I did use auto exposure during the scan. I even used ICE. The results I got were acceptable for creating slide shows to be played on an HDTV. However, when I shoot landscapes with my MF camera, then I save the adjustments for after the scan.

My suggestion is to try it using both methods on the same picture and see what works for you. They're different ways to skin a cat.

Hi there,

I've used an epson V550 for all my medium format scans, from 6x4.5 to 6x8, with both stock and third party holder. All of the film shots in my gallery (check the blog) here: Dead Link Removed
have been scanned at various settings, which have molded over time due to learning small things that seem to make a difference (dpi adjustment, height etc). All in all, if you can find a refurbished version of the V550 (occurs sometimes) you will get a really capable medium format scanner, which will keep you in good stead for most practical applications. You really have to pay a premium to move up in the scanner ranking, to say the V800/850, or dedicated film scanner, and I think you will be pleasantly surprised at just how capable epson's baby V600 works - I've no complaints after at least 5 years of scanning.

Feel free to ask any questions, and enjoy whichever model you choose - either will make great scans.

Cheers. As mentioned a few posts back, I got the v550, and am happy with it. Easy to operate, and decent-quality scans.

sixby45 said:

Hi there,
...

Feel free to ask any questions, and enjoy whichever model you choose - either will make great scans.

Click to expand...

Can I use one of your models?