Scanner Optical Resolution vs. Effective DPI?
 

I was looking at the Canon LiDE 110, and the listing says that it has an optical resolution of 2400x4800 DPI.

Perhaps I am in the dark regarding recent tech updates, but I thought most scanners only scanned at 300-600 DPI? Maybe these numbers are different from one another, but I keep trying to figure out the math and it doesn't quite add up to me.

Amazon Listing:
http://www.amazon.com/Canon-CanoScan...words=lide+110

No, you've always been able to optically scan at better than 300-600dpi. You're thinking of cheap consumer scanners (mostly flatbeds), not professional scanners.

Most pro scanners can easily start at a true 1200dpi, and some go up to 4000dpi or more. Pro scanners also offer 12-bit, 14-bit and 16-bit color, depending on the model. I was using a Kodak RFS scanner in 1999, which offered 16-bit color and 1800dpi. The thing was many thousands of dollars, owned by the newspaper where I was staff.

After that, you mostly interpolate the results. Cheap home scanners that claim greater than 600dpi are always interpolating to get those crazy numbers (1200-9600dpi).

Most of our high-end Nikon and Minolta scanners give true 3200dpi+ optical resolution, 14-bit+ color, and have ICE (and ROC and GEM). Our flatbed also give 1200dpi and have 16-bit color.

Yes, that cheap Canon scanner is probably 600dpi at most, though more likely just 300dpi.

Good flatbed scanners usually cost $500+, while negative/slide scanners are $1k+. You pay for the glass especially.

So at 2400x4800 optical (a different number is listed as interpolated), listing 48-bit color, how do you convert that to what you actually get out of it?

48-bit = BS
2400x2400 = BS

It's become marketing. It was the megapixel myth before there were megapixels.

48 bits is just 16-bit. You use to say "8-bit!" then 16-bit came out. So you took "8 bits per channel" to create "24-bit!" -- and suddenly 24 was bigger than 16, even though it was bollocks. The 16 was greater than the "24". We did that with every increment to date -- 10, 12, 14, and 16. So "48-bit!" is just 16. And when cheap hardware is used, you can pretty much guarantee that number is fudged, with the 16-bit being rounded or dithered (or whatever) to get that magic number. In true lab tests, it'd surely fail.

Same for the resolution. If the optics are smeared with vaseline, the actual image will be nothing like the supposed pixel acquisition. We see this in both photo and video. An iPhone claiming X megapixels will easily be crushed by my Nikon D3s, with it's $2k+ optics and giant sensor. The iPhone may acquire X pixels, but the glass to actually resolve them does not exist. Again, in lab tests, a cheap scanner promising big resolution would certainly fail.

If it's too good to be true, when photo and video is concerned, it always is. :2cents:

Just to make sure we're on the same page, because I am not the photo expert:

Optical - actual scanned resolution
Mechanical - artificially slowed down to scan more (similar to intelligent zoom in video)
Interpolated - artificial software resized with no detail gained, and in fact, some lost.

My camcorder technically takes real 16 MP stills, but they are B-grade, and often times I end up taking a video instead of a photo because I can choose between 60 stills per second.

I am getting the impression that this situation is similar to what I am finding here. There must be some truth to this resolution, as the next scanners up can scan slides, which I imagine to be impossible without a much higher resolution than 300 ppi considering the size of film. In practice, though, the sensors may not be all that great.

I suppose an equivilant comparison would be that just because some people have drivers licenses doesn't necessarily mean that they can drive :)

Optical, yes.
Interpolated, yes.

But there's no term for "mechanical" -- and that's where the funny business happens. It's the gray area.
Welcome to the Nonsense Zone!

Camcorder sensors are generally very small, the glass not optimized for greater than 720x480 SD (or 1920x1080 HD) images, so the "16MP" is complete BS. Calling it "B grade" is probably adequate, if not outright generous.

It doesn't take much to scan slides. 300dpi = dots per inch. How many inches is the slide?

Slide scanning is inherently interpolated in a standard flatbed scanner. You see something like "300dpi, also scans slides!" -- note the order of the text.

You want an analogy? I can cook, but I'm not a chef. No more than those items do what they promise.

Sorry, I just double-checked and the photo resolution is 20.4. I think they make that claim because it has 3 sensors, though I wouldn't compare the quality to a 5 MP Rebel by any means.

Are the Epson Scanners that you've mentioned in other posts any different from what other manufacturers claims are, regardless of ICE?

I knew you were going to ask that. :smack:

Right now, the scanner is disconnected and put up. We're sorting photos, and need as much room as possible.

The V600 is at least 1200dpi or more. You can tell by the sound it makes. When it "goes back" to re-scan, it's fake resolving. I forget where that barrier is off-hand.

I believe that the slide scanning makes it go into another mode internally. And then dpi doubles.

There have been advanced talks on this scanner, and others, over the years. In lab tests, the scanned slide was around 2500-3000 actual dpi, and mostly good for 8x10 prints. If you want larger, take the slide to a photo lab.

Quality of the film also plays a huge part in this!