Scanning with the emulsion facing the wrong way

[pkr1979]

Hi all,

So, I am scanning with my ScanMate 11000. And as its been some time since the last time Ive ended up attaching all the negatives with the emulsion facing out instead of in. Does anyone know how much this will affect the scans? I am considering if I need to re-scan it all.

Cheers
Peter

 

[Don_ih]

I've never noticed a difference. I will put the negatives in a negative holder (in an Epson V600) in whatever way keeps the surface of the film off the glass. If there's no choice (as in, some 120 will fall on the glass in the middle no matter what), I'll put it emulsion side down to avoid newton rings.

 

[radialMelt]

Same experience as Don_ih here. I've scanned both ways (similarly, to avoid newtons rings in glass holders) and not noticed any differences.

 

[pkr1979]

Might it matter that the ScanMate is a drum scanner?

 

[koraks]

Might it matter that the ScanMate is a drum scanner?

Yes, but even so, I'd expect the difference to be so marginal as to make the option of re-scanning not very relevant. I'd not bother unless you are seeing severe problems in your scans, which I bet you're not experiencing.

 

[Don_ih]

Do a check. Scan the same negative twice, both ways (emulsion in, emulsion out) and compare. That will give you the answer.

 

[MattKing]

This may depend slightly on how much base fog you are dealing with.
I would aim for having the emulsion closest to the scanning sensor.

 

[brbo]

Do a check. Scan the same negative twice, both ways (emulsion in, emulsion out) and compare. That will give you the answer.

Easier said than done when negatives are already mounted on the drum. Mounting fluid, mylar and all...

@pkr1979, scann a small section of one negative and see if auto-focus picks the proper (emulsion) side. I don't know the SM 1100, but if it has a way to manual focus the scanner that would probably be the best in this situation. A lot will depend on the chosen scanning aperture, you might get away with a larger aperture, but know that mylar can't offer the uniformity of the drum on the contact with the emulsion.

 

[Don_ih]

Easier said than done when negatives are already mounted on the drum. Mounting fluid, mylar and all...

No, I mean a separate test. Scan a negative, scan it again facing the other way, compare. That should indicate if it mattered all along and show whether or not some of the previously (backward-mounted) scanned negatives need to be scanned again.

 

[brbo]

Of course, but that is a lot of work if, like OP, you have the entire drum mounted with various strips or even tens of individual frames and you don't have another drum to make the test for comparison prior to scanning the already mounted drum. Entire drum at high resolution means a full day of scanning on a "desktop" drum scanner.

Anyway, it just dawned on me that I used to deliberately mount the film with emulsion away from the drum when I only had a drum with slight crazing. That way the crazing was less noticeable in scans and quality was still good.

But, if OP can't do a comparison, somebody else surely can...

So, here we go....

Mounted properly:



Mounted with emulsion facing away from drum:



(click for full resolution)


If your auto-focus can pick the emulsion side or you can focus manually, I wouldn't worry and scan the film...

 

[Alan Edward Klein]

Wouldn't you get moire patterns if the film is against the glass? Isn't it better when the emulsion side is one way or the other?

 

[koraks]

Wouldn't you get moire patterns if the film is against the glass?

The film is fluid-mounted on the drum; this prevents Newton rings from happening.

But, if OP can't do a comparison, somebody else surely can...

That's pretty compelling. In the sense that I really don't see much of a difference at all.

 

[brbo]

Yes, virtually no difference. Maybe scanning a fine-grain slide film with the smallest aperture (this was with the second smallest aperture) would show an observable difference, but I doubt anything as big as to require going through the pain of remounting the film on the drum.

 

[pkr1979]

Thanks guys! So, to avoid brain-itch I re-scanned the negative emulsion the right way. I cant see any difference at all. @brbo - what scanner are you using?

 

[brbo]

@brbo - what scanner are you using?

The poor man's drum scanner - Howtek 4500.

 

[250swb]

Isn't it ultimately just a software issue for most scanners? Whichever software you are using will say to scan emulsion down or emulsion up so you don't have to reverse the scan afterwards. The Epson Scan (Epson V700 etc.) for example says to scan emulsion up, but Vuescan says down. It doesn't matter to the scan itself as it's just a post-scan management issue.

 

[koraks]

Isn't it ultimately just a software issue for most scanners?

In terms of orientation of the digital image, yes. I think the primary concern here was about actual image quality; please see the exchange on this, including examples, in the posts above. Also note we're not talking about a flatbed scanner.

 

[Harry Callahan]

Out of curiosity i also took a look at the two scans brbo showed - and to check for sharpness and grain i zoomed in on both scans to the same spot, to the cylinder lock on the trunk of the blue car. I had both pictures in to different tabs and then switched rather fast between both pictures, to check for small differences in sharpness or grain.
And i found that sharpness and grain basically is the same...
... but not subject detail!

Especially the license plate holder on the blue trunk does differ in shape - there is a lug on the emulsion out scan, which is not present on the emulsion in scan:

Red arrow, no lug -

- but now there is some sort of lug, which wouldn`t make sense as it would lift the license plate. This "lug" must have been inserted by some image-improving-software.
Red circles show areas which also differ, on the upper red circle chromium seems to bleed into the blue paint, in the lower red circle the chromium sheet itself differs in shape.
To see the differences in both red circles best, one should open each picture in an individual tab and then swap between both tabs rather fast, like within one second.

Another part of the subject also is altered:

Here is some brown dirt on the black tire of the car standing behind the blue car -

- and here some of the brown dirt is gone, as if someone cleaned the black tire a little. Also the tire tread differs a little in the red circle - but this only is visible if you swap between both pictures rather fast.

Then there are orange circles in several perforation holes, these maybe are air bubbles from the fluid used to prevent Newton rings - but quite some of these orange circles are symmetrical to 100%. Some are oval or non-circular, but a lot are incredible symmetric. The symmetric ones are indicated by a red arrow - you need to zoom in to see how symmetrical they are:

Making me wonder whether you can have so many entirely circular air bubbles or whether an image-software decided to shape several air bubbles to perfect symmetrical circles.
Also i wonder why these circles are orange at all. If these indeed are air bubbles, why do they have basically the same orange as the edge markings on the film?
Seems like an image-software decided to color possible air bubbles to the next nearby color which would be the orange of the edge markings.

 

[koraks]

@Harry Callahan the so called missing details look like fixed spots of dust to me. I.e. clone tool artifacts. I don't see how a drum scanner could make up or forget details. The supposed 'lug' for instance doesn't look like one to me. Pieces of mud, bodywork seams etc. can easily be affected by clone/healing tool actions.

Note how very clean @brbo's scans are. Apart from tidy handling practices, he may have done some cleanup on the files. See e.g. this bit where he 'missed a spot' (which sounds horrible to say in relation to such otherwise impeccable scans):

If you compare both images by overlaying them (they align virtually perfectly) you can see plenty of tiny differences in blemishes that are present in one file and missing in the other, and vice versa. I reckon he did some cleanup on the more objectionable ones.

The round circles do indeed look like air bubbles in the mounting liquid to me; that's why they're in the sprocket holes.

Also i wonder why these circles are orange at all.

For the same reason the film edge itself renders orange. Apparently a diffraction artifact.

Interesting detective work, but I don't see much relevance to the question of film orientation during scanning.

 

[brbo]

LOL!!!

Yes, I did quickly run through the scans with clone tool and yes bubbles tend to be perfectly round when they are not squeezed against something.

 

[Harry Callahan]

@Harry Callahan the so called missing details look like fixed spots of dust to me. I.e. clone tool artifacts. I don't see how a drum scanner could make up or forget details. The supposed 'lug' for instance doesn't look like one to me. Pieces of mud, bodywork seams etc. can easily be affected by clone/healing tool actions.

Note how very clean @brbo's scans are. Apart from tidy handling practices, he may have done some cleanup on the files. See e.g. this bit where he 'missed a spot' (which sounds horrible to say in relation to such otherwise impeccable scans):

....

Oh, havn`t thought about that - yes that does explain the differences i found. I thought that the scanning software did some "improvement" on the raw data. Well, there still are several dust particles on both scans... maybe because of this it didn`t occur to me that some "de-dusting" has been made.

...

The round circles do indeed look like air bubbles in the mounting liquid to me; that's why they're in the sprocket holes.


For the same reason the film edge itself renders orange. Apparently a diffraction artifact.

...

But if the bubbles are colored orange by diffraction, doesn`t this mean that the bubble must be exactly the same thickness as the film or exactly the same surface-properties? Because when diffraction does create a color, thickness/surface-properties usually are critical within the nanometer range...

...

Interesting detective work, but I don't see much relevance to the question of film orientation during scanning.

No relevance, just occurred to me. To get back on topic, when i did my detective work i found that the edge of the picture might be a little less sharp on emulsion out:

and a tad sharper on emulsion in:

But emulsion in also is a little darker, maybe this does make it appear a bit sharper...

LOL!!!

Yes, I did quickly run through the scans with clone tool and yes bubbles tend to be perfectly round when they are not squeezed against something.

...you...You...YOU... how can you do this to me? Are you aware of what i look like now? Like an idiot - on the internet!



When i had a look at the bubbles i already was convinced that some image-software was altering image content... maybe i also expected some flow of the fluid which would de-form the bubbles a little... i mean there also are capillary (?) forces which should deform a bubble when it has contact to a solid body. Several of the perfect circles do have contact to the perforation hole, maybe that`s why i was that sure...

...but ok, let`s settle that it`s all your fault - and not mine .

 

[NMSS_2]

you’ll probably notice a bit less sharpness and maybe more grain, but for most purposes it should be fine. Only re-scan if you need it really crisp.

 

[koraks]

and maybe more grain

How do you figure that? Can you post an example?