High res 8x10 scan sample, anybody?

[Sizam]

The crops are almost comparable size in a pixel sense - what you're missing is that the 8x10 scan crop is from an image which is 19500x15500pixels, whereas as the Hassy is only 7200x5400 pixels - so is you printed both at 360DPI, the scan would produce a print of about 53x43 inches with about the same level of detail as a 20x15 inch print from the Hassy... The print from the scan would be 7 and a half times bigger than the print from the Hassy. Apples and oranges. Even then, I think that the scan crop shows more detail than the Hassy crop - there's no texture in the bushs on the Hassy crop.

Yes, no doubt the scan has more detail then the hassy it just appears to be close, ie marginal, where I'm expecting 4x at least (7x as you say would be great). Could you please explain the logic in the above paragraph a little more. It seems like what you're saying is you'd have a 53x43 print vs a 20x15 print but the 53x43 print would just be larger, ie the people would be 7x larger which would net nothing since that would just be 'blowing up' the detail we're seeing in the sample images. If however you mean that we have a 53x43 print vs a 20x15 print and the size of the people remains constant, well then now we're talkin.

[Sizam]

Ok,
We noodled it out and got that you ment the latter (more awesomer) option which really makes my day (weekend even).

Thanks!

[Don Hutton]

I just did this crop - it's exactly 467x498 pixels - pretty close in actual pixel size to the Hassy "crop". However, the scan is 19500 pixels wide by 1500 pixels high vs 7210 pixels wide x 5410 pixels high for the Hassy full image. So if you print both images at 360DPI, the scan will produce a 53x43 inch print whereas the Hassy print will only be 20x15 inches, BUT the level of detail in both will be comparable, as shown in the actual pixel size crops.... i.e. for the people in the crops to be of similar size, the print from the scan will be 7 and a half times larger than the print from the Hassy file.

[Sizam]

Yea,
I get yea now. Freakin awesome.

[Don Hutton]

Don, what size did you come up with?

Sorry I missed this earlier Will

My conclusion was that around 24x30, the print from an 8x10 starts to show some worthwhile advantages. This is presuming very careful and high quality capture on both formats, drum scanned digital workflow and inkjet printing. I threw an extra stop on the 8x10 because working apertures in my experience are typically about a stop down from 4x5. With a very fine grained high resolving film like Tmax 100 (which is what I shot for the B&W test) the differences probably start to become apparent at slightly larger print sizes than they would for a faster film. This was really about justifying keeping my 8x10, and despite the fact that I seldom make very large prints, I decided to keep it!

[Sizam]

This was really about justifying keeping my 8x10, and despite the fact that I seldom make very large prints, I decided to keep it!

No kidding! Glad I got into 8x10.

[james zhou]

I had used a D300 as a light meter and test shot to analyze the scene, while setting up my 8x10. I later looked at scans from 8x10 provia and some D300 shots of part of the scene. I think the 8x10 film is about 18-20 time bigger than the D300 for the same detail. That is similar to what you two found out. Interestingly, I found Provia gives more highlight and shadow details than T-Max when shooting the same scene---but this could be due to how film is processed. This convinced me to keep all of my film camera including Pentax 67, and 4x5. Nice postings! Thanks! jim

[Steven Barall]

Thanks Don. That is pretty cool when you actually see it displayed like that.

[Ken Lee]

"My conclusion was that around 24x30, the print from an 8x10 starts to show some worthwhile advantages".

That makes good sense. Let's ignore film grain altogether, and presume that a good taking lens for Large Format gives somewhere around 50 or 60 lp/mm at best. This means that under anything less than ideal circumstances, it performs worse, perhaps more like 30 to 50. On a contact print, that's a lot of line pairs. The look is... smashing !

But to get a 24x30 image from 4x5, requires around 6X magnification. Now that 30 to 50 lp/mm is down to 1/6 of the original: somewhere around 5 to 8 lp/mm - inside the range where the average human eye can detect a difference.

Meanwhile, the 8x10 image, requires only 3X enlargement, and still has 10-15 lp/mm. It is still at - or above - the limit of human perception.

A 6x6 image made with the best lenses, may have somewhere around 80 lp/mm. You can make a 8X or 9X enlargement, and get a very sharp image around 18x18 inches.

But we rarely get to shoot our lenses at their best aperture, and it's mostly normal length lenses (like 80mm on 6x6) which do so well. So it may be more realistic to expect a 16x16 image. In any event, that camera had better be on a tripod, or we can kiss all those line pairs... good bye.

[Bruce Watson]

Could somebody(s) post some samples of high-resolution scans from 8x10, color or B&W? I'm curious what you can really get out of a high-res scan so maybe a small 100% crop and a full sample if you will.

Got a couple on my Large Format Pro website. One from color to show detail, one from B&W to show tonality and shadow detail. Both drum scans of 5x4 negatives though. Might be close enough for what you need.