Is my Bronica SQ-A an 81-MegaPixel camera??

[keithwms]

Keith,

...do you still feel that 16-bit depth is necessary for continuous tones using b/w film?

Daniel, I think the "safe" route is to go with 16 bit, but as you quickly learn, that means enormous files which may contain way more information than you actually need for direct print.

Truth be told, I frequently do simple edits in 8 bit and it works just fine. When working with a 100+ mp file, just doing basic crops or rotations can be pretty unwieldy when the bit depth is maxed out! So I ask myself if my edit is a tonal edit or an atonal, pixel edit and take it from there. If you are doing curves, levels, or interpolation then it's safer to go with more bit depth and to work in the most lossless file format you can. If you're doing a simple crop before sending the file to an inkjet then doing that in 16 bit is silly. But if you have unlimited computational resources (or time) then go for it!

Just keep an eye on the histogram during edits and watch out for posterization.

As with many other things, I think you have to find your own answer because it depends a lot on what you will actually be doing with your files. Probably you will get many different opinions...

 

[donbga]

Keith,

but do you still feel that 16-bit depth is necessary for continuous tones using b/w film?

- Daniel

In a word yes. B&W image files are not immune to posterization either. Yes files scan in 16 bit can get quite large especially for scans of sheet film. But using a Guide File work flow can reduce some of the problems of using a large file.

Look here for details about Guide File Workflow.

http://www.westcoastimaging.com/wci/page/info/photoshoptip/podcast_guidefile.html

I prefer to make my initial scans as large as possible and then down sample them as needed for a specific enlargement size. Since most modern scanners and scanning software allow for larger than 8 bit scanning (if I remember correctly the Nikon 9000 is 12 or 14 bit) doing a full size native resolution scan means you only scan the film once for whatever use one may have for it.

Don Bryant

 

[Daniel Balfour]

Safest bet seems to experiment -

I'll scan the same file in both 8 & 16 bit depth and apply tonal/contrast corrections to it while watching the histogram. I'll report what happens.

 

[sanking]

Safest bet seems to experiment -

I'll scan the same file in both 8 & 16 bit depth and apply tonal/contrast corrections to it while watching the histogram. I'll report what happens.

Actually, there is really no need to experiment if you want to get the best quality print. For that you scan in 16 bit and do all of your editing in 16 bit. With some types of images you might be able to get away with 8 bit scan, but the risk is always that you will dramatically reduce the total number of levels possible in Photoshop and cause posterization. In some cases even one curve adjustment can reduce total levels form 255 to 150 less.

For purposes other than high quality printing (web gallery, etc.) you may be able to get away with 8 bit files. But for really high level work you should scan in 16 bit and do all of your editing in 16 bit. I believe that is the advice you will get from almost everyone who understands the subject.

My procedure is similar to that of Don, and involves scanning to archive. That is, I always scan at the highest reasonable resolution and save the file on the hard drive or DVD. Then I will edit from this original, saving the changes at the size I need to print.

What is the highest reasonable file size? This depends on the RAM, hard drive space, and processing power of your computer. I work with an Intel duocore iMac and can easily save and process files in the 500 mb to 1.5 gig size. I typically scan 5X7 negatives in 16 bit grayscale at 2540 spi, and this works out to a file size of around 430 mb. I archive these files 8-10 at a time on a 4.7 gig DVD. This would allow me to make a print of 20X28" size at 635 dpi, or one of 40 X 56 at 300 dpi.

I could scan the 5X7 negatives at even higher resolution than 2540 spi, but that is already more than I needed since I am not interested in making prints larger than the sizes mentioned above.

Sandy King

 

[keithwms]

Editing in 8 bit is almost always a bad idea, if the editing in any way affects tonality. But I have found that most of the time I can scan to 8 bit, do crops and rotations in 8 bit, and then go to 16 bit for any tone edits. And all is well; the Earth keeps rotating on axis, and there is no posterization. As long as you are very clear on what constitutes a tone edit, you can skimp by on 8 bit for everything else, if you want small file sizes and you want to work quickly.

For archival purposes, sure, scan to the max.

As for me, and I realize that these words will sound ironic coming from a device physicist, but... to me the ultimate archival medium is..... a piece of properly stored film So I scan to sizes that are roughly double what I can ever imagine needing, and then I store my film in a nice, comfortable place.

 

[lenny]

Your film has a continuous tone scale so you want to scan to a high bit depth to try to pull out as much as you can. The downside: you have to store much bigger files!

The only good news is that hard drive costs continue to come way down. I built a 5 Terabyte RAID last fall for less than 2K. Of course, PhotoShop still has some work to do getting it work better with larger files as well... all those memory issues, etc.

Lenny

 

[keithwms]

One thing I suppose one could do in PS is to do computationally intensive edits in 8 bit mode (or even smaller) and record those actions, and then have PS redo them in 16 bit while we sleep.