Can I make 50Mb Tiffs with my K10D?

[thomas5]

I've been shooting film and producing negs and slides I can scan for the stock agency I work for. I figured

it'd be nice to suppliment that work with a digital camera and so far I've been fine.

 

What I want to know is can my new digital camera do what I ask if I crop my images to a square and what

software do I need to make TIFFs that match the ones I'd been scanning from film?

 

Thanks.

[strpichu]

Shoot RAW, and resize. What film gives you 50Mb TIFFs?

[thomas5]

Pretty much whatever I like. I shoot a Hasselblad.

[strpichu]

Ahh, medium format. That explains. Well. You cannot match good scan from medium format. Stay with your Hassy.

[thomas5]

I don't plan to ditch the film. Believe me.

 

 

What I want is a digital alternative though.

 

 

I downloaded the free version of Lightroom to see how it worked. It's still plenty

confusing, but I managed to output some TIFFs that were only 20Mb. What amazingness

am I missing to make my images bigger? I mean, I'm shooting in RAW (which I

understand to be proprietary anyway), what more is there?

[mountainvisions]

I get 60MB tiffs from my film scans...just plain old 35mm.

 

Better scanner could yield bigger files.

 

Every scan yields the same size file at X DPI and bit. I scan at 4000dpi input and 16bit.

 

It's not magic. Coincidentally, the resultant file is a 10MP file. The K10D will actually yield the same size tiff.

[mountainvisions]

thomas, specifically replying to your lightroom issue, i don't know what size you saved at but the files in TIFF should be quite a bit bigger then 20MB as the original RAW are 14MB and the uncompressed finalized tiffs are significantly bigger.

 

I don't really pay attention to this stuff but I'll take a look at the exact size of both my 6MP RAW-TIFF's and my 10MP. If I remember correctly the 6MP RAW yielded a 30MB TIFF and the 10MP over 50MB.

 

I'll look later when I turn my desktop on.

[kelly_flanigan1]

You can make 50Mb tiffs from a 1/4 VGA Barbie cam or Polaroid Webster sticker cam. You can make 50Mb Tiffs off a drum scan of a 4x5 trany. One can also downsize the 150meg file off a 50megapixel Phase One scan back to 50 megabytes too. One can also take one dead mouse and add hamburger helper and serve the meal to a football team; or take them to a steakhouse and get each one a 1 lb steak. Stock agencys take a dim view to bloaded files and upsized to the moon images. The reviewers didnt just fall off the turnip truck and can smell BS from a mile away. Some actually ban inputs from folks who lie and upsize images and claim they are not upsized. If you fib to an agency about not upsizing it just means they will wonder about other things you are fibbing about; image ownership, doctored images, stolen images. A cropped square image from a 10.2 megapixel K10D is going to be roughly in the 7 megapixel range; about a 21 megabyte RGB TIFF file. The header info will contain the camera info ; and unless stripped out will show that you upsized the image when you fabricate a larger image size. You should check if your agency really wants or accepts an upsized/bloaded image before submitting them so one is not crowned with a negative marker.

[thomas5]

Sorry, Kelly. I didn't mean to offend you.

 

 

They asked for the images and I told them what I'd shot them with. There really wasn't

any subterfuge. I'm nowhere near that clever.

 

You have, however, answered my question: I can't. The biggest I can get the file to be is

the 20Mb I've managed to achieve. Thank you for that.

 

And, since I've been an English teacher for the last 10 years, I think I can say with a level of

certainty that many football players wouldn't be able to tell the difference between a

mouse (but I understand there's more dark meat on a hamster) mixed with hamburger

helper and they'd be none the wiser if you chose to feed them thus.

 

Thanks a million.

[amund_aaeng]

An uncropped K10D file is 57Mb, cropped to a square it will be 38,5Mb as a TIFF...

[thomas5]

Appreciated, Amund.

I can only seem to make my RAW files 27 or 28 Mb. Where (or better yet, how) are you

getting 57Mb from?

 

I know this all seems simple to many of you, and frankly I'll probably use the camera mostly

to shoot personal family-type stuff, but on the occasion that I can get it to work like this, I'd

like to already be in the know.

[graybrick]

You can get a 39MP back from Hasselblad... otherwise you can't compare DSLR output to MF film... or 35mm film for that matter. Pony up 25G and you're in the ballpark. I want an H3D39 myself, but as a lowly student I'm stuck with my punt little Canon toy...

[mountainvisions]

Thomas, you have no idea what you are talking about.

 

The uncropped full RAW to tiff 16bit file from a K10D is 57MB.

 

Not sure where the 20MB is coming from but I'd check your settings.

[thomas5]

Score!

 

You're right, Justin. This is completely new territory. I want to understand though, so I'm

paying attention.

 

I have been making 8 bit files, not 16. That one piece of information is where I consistently

was going wrong.

 

Thank you for pointing that out! Now I just have to figure out what the difference is... :)

[nealcurrie]

I'm not sure why some of the responders you'e had so far have been rude. You haven't said anything to deserve it.

 

The difference between 8bit and 16bit files are the number of available colours. A bit is a 0 or 1.

 

A TIFF file stores each pixel using 3 values: one for each of Red Green Blue. Using 8bit, each colour can have 2^8 = 256 (2 to the power of 8) different intensities. This means the total number of colours available to each pixel is 256 * 256 * 256 = 2^24 which is about 16.7 million.

 

Eight bits happens to correspond to a single character (a "byte"). Each pixel in an 8bit TIFF requires 3 characters (3 bytes) of information. For a file size of 3872 x 2592, this is about 10 million pixels. Each one needs 3 bytes, so about 30 million bytes (or about 29MB).

 

 

Using 16bit, each colour can have 2^16 = 65536 different intensities. Each pixel in a 16bit TIFF requires 48bits (3 times 16), or 6 bytes. The total number of colours available to each pixel is 65536 * 65536 * 65536 = 2^48 which is about 280 trillion (28 followed by 13 zeros).

 

For a file size of 3872 x 2592, this is about 60 million bytes (6 bytes for each of the 10 million pixels). 60 million bytes is about 57MB.

 

Any questions about the above, just ask me.

[thomas5]

Aye, Neal. What you've said makes sense.

I think I ken how the numbers might make one compression different from the other. What I

don't quite understand is perhaps why the people who've requested my images would ask for

8 bit (and a quick look determined that) when the 16 would (apparently) be more

advantageous.

 

Thank you, kindly.

 

And cheers on the teaching of maths. I'll stick to English. Maths is all Greek to me. (grin)

[nealcurrie]

Welcome. Did I mention I'm a math teacher when I'm not busy in my darkroom or out taking pictures.

[kelly_flanigan1]

Most all stock agencies want their "megapixel" requirements based on 8 bit RGB files, not 16 bit ones. <BR><BR>Thus if your camera outputs a 10 meg 8 bit file and you go into photoshop and convert it to 16 bits; the file size is twice as big; ie 20 megs. You still have the same number of pixels. <BR><BR>Its like cashing your paycheck in single dollars instead of larger bills.<BR><BR> It doesnt impress a stock agency to play games with numbers; it insults them. Its not going to impress you wife or kids by giving them jars of pennies and act like its more money. You can make any size file you want to with upsizing; some systems limit the max ones to about 4 gigabytes. <BR><BR>Its quite rude to "pass off" a lessor file to as stock agency like it was shot with a bigger rig. This type of event is real common with the lay public; it ticks off publishers to no end. It wastes folks time; the images take longer to rip for printing; the images clog up the rip and are just useless info. <BR><BR>This type of rudeness by photogrphers gets them blackballed, shunned, and earmarked as creators of bload; hated as much as spammers. If you fib about the matter and they know it then trust is lost; your professionalism is tarnished,. <BR><BR>Strive to find folks actual requirements and avoid the rudeness of creating fake upsized images. Publisher clients of ours really swear at this amateur jackassery of creating useless bloaded files. Some employ a chap/filter to cull out these images. If you dont get any replies from a publisher or stock agency with bloaded/upsized inputs look yourself in the mirror and repeat I will be less rude and submit proper images; ones that dont insult folks.<BR><BR>This upsizing to fit minimum requirements is a nightmare for some receivers. <BR><BR>If you do upsize images; LET THEM KNOW what the 8 bit RGB un-upsized image size is too. This defuses the hatred and helps build some needed dialog. Its not that it is bad to upsize; its bad to pass it off as not upsized; or to create too much bload factor.

[mountainvisions]

Kelly,

 

Megabytes and pixels are two separate things. 16 an 8 bit have no reflection on the actual mega PIXELS. They do, however, affect the mega BYTE size.

 

Thomas, for print purposes most printers (although this is changing)prefer the 8bit color space. The gammut of 16bit is so big that printers sometimes have a problem. 16bit AdobeRGB is a huge color gammut compared to 8bit sRGB.

 

Likewise, they often prefer sRGB even though this is a very small gammut compared to AdobeRGB and also the newer and larger (??name is escaping me) RGB.

 

The newer desktop printers are now able to recognize larger gammuts and print them.

 

Hopefully this helps some. BTW, you should be scanning your film in 16bit and AdobeRGB to get the most data from the original scan. Then downsample as needed for your ouput source.

[thomas5]

Thank you all. This has been very helpful for me. The short subtext of it seems that since

I know my film camera already, there's no reason to deviate far from it. The digital is

considerably smaller and lighter, however. (smile)

 

Kelly, I see that you're really concerned that I might fudge my numbers, but I have to

emphasize to you that I don't speak fluent enough digital to want to try to pull the wool

over anyone's eyes on this. I'm asking for my own edification and I plan on asking the

agency with whom I'm working as well to make sure I'm all clear.

 

I'll certainly check in if I have other queries.

 

Cheers.

[andrewg_ny]

I am also wondering why this thread is attracting rude responses. As for 8 vs. 16-bit, I don't claim to be an expert but my understanding is that in addition to lack of printer support a lot of software has limited support for these images as well.

 

-Andrew

[mountainvisions]

Not so much anymore. Beyond PS 6 16bit was supported in almost 100% of the features. Now elements cripples 16bit to some degree but even it will eventually fully support 16bit (as it seems to allow more and more 16bit editing with each generation).

 

Likewise, the other programs and filters I've used support 16bit pretty much universally.

 

it simply gives a larger gammut and more gentle gradiations, less of things like banding and artifacting as you work on an image and make changes.

 

I should note that many monitors are calibrated to sRGB so for web display always downsample to 8bit sRGB.

 

I don't think this thread has been nasty. bit depth and color space have been hotly contested for ages. Simply, IMO, it is best to record in the highest bit and largest color space for your original. This retains the most information, that is closest, to what the eye sees. It also gives more headroom for editing a file or restoring a poor original (ie overexposed, underexposed, wrong WB in JPEG, etc).

[karel_peijnenborg]

Thomas: "What I don't quite understand is perhaps why the people who've requested my images would ask for 8 bit (and a quick look determined that) when the 16 would (apparently) be more advantageous."

 

The 16 bit images are most useful for image editing: you will be working with as much info as possible. When you tweak contrast, the colors will not be squeezed into 256 steps but smeared out over 65K steps, and when you then tweak a tone, the same applies. You *will* be compounding errors, but much smaller ones. For output however, 8 bit files are probably more than enough for most people, unless we get a serious eye upgrade.

 

Regards,

 

Karel

[thomas5]

Well since you're all still answering queries, how about this: You can see that I'm pretty well

clueless as far as the digital side of my work goes. I've been paying for scans of my negs but

I've recently gotten access to the local university's photo department and have scanned a few

of my own negs. I also recently downloaded Lightroom, which is beginning to make sense to

me. Given my novice level of digital understanding and the K10D I have, is Elements 5 a

better choice for output or can I stick with Lightroom?

[mountainvisions]

Elements for editing of your film scans. Light room for the K10D RAW files.

 

Lightroom is mostly designed for RAW workflow from D-camera. Elements and PS are more for photo editing from scans, tiffs, and jpegs. Unless of course something changed with the later versions of lightroom.