SLR: Digital vs. Film

[JonEdangerousli]

http://www.snopes.com/business/genius/spacepen.asp

FTFA:
"The lesson of this anecdote is a valid one, that we sometimes expend a great deal of time, effort, and money to create a "high-tech" solution to a problem, when a perfectly good, cheap, and simple solution is right before us."

&

"Fisher spent over one million dollars in trying to perfect the ball point pen before he made his first successful pressurized pens in 1965."

Fisher being an American it is correct to say that the US spent a million dollars developing the pen. Whether the US government paid that much for them is a different statement, although it is semantics. The Russians did use pencils. Again, whether they used the Fisher pen when it was developed or not the statement still holds.

 

[farace]

But where it doesn't work is 1) Fisher was never asked to develop the pen; he saw a business opportunity and used his own money and approached NASA only after he had developed the pen. He never asked for reimbursement of his R&D costs. 2) Pencils are dangerous in an oxygen environment; the original story implies that pencils were the low-tech perfectly good solution, when in fact they were not a solution at all, they were the problem. Besides the fire danger, there was the danger from floating pencil debris. Yes, the Russians used pencils, but so did we, and they needed to be replaced with something safer.

I agree that we often go high-tech when it's unnecessary, but Snopes also stated that this was not an example of that. I'm sure we can find another example of an overengineered "solution" to a very simple problem. The Power Fusion comes to mind . . . (And I'm still trying to figure out diet water--what the hell is that? But that's a solution where there was no problem to begin with.)

 

[JonEdangerousli]

Isn't that the ultimate description of marketing? Creating (and selling) a solution for a problem that doesn't exist?

 

[farace]

Isn't that the ultimate description of marketing? Creating (and selling) a solution for a problem that doesn't exist?

Or sometimes going so far as to create the problem in the first place?

 

[Duggo]

After going digital, I could never shoot only a roll of film every two weeks. I used to shoot about that much with film, because I just didn't want to spend more on it. Now often shoot 50 shots in a day, sometimes many more. I shot over 5000 shots last year. At $15 a roll for a 36 shot roll, that would be over $2000, twice what I paid for my D70.

With me, I took a class where I had to finish a roll of 36 exposures in two weeks. I couldn't do it because, I was too worried about getting the perfect shot, for the money I was paying. (15.00 - I think for the roll and 15 to develop) so it had better be good!

When I bought my Olympus C-2100UZ in 2001, and I went into class the next week, the teacher said, "I expect great things from you Duggo." Suffice to say, she was pleased with what I handed in. I definitely felt a certain freedom and I wasn't always thinking about the cost involved. It was definitely the right move for me. Unfortunately, 6 years later, 1, they still haven't come out with the perfect camera, and 2, I'm in need of an upgrade.

 

[paydepst]

Isn't that the ultimate description of marketing? Creating (and selling) a solution for a problem that doesn't exist?

Isn't that sort of what King Gillette did with his safety razor?

 

[JonEdangerousli]

Exactly the example I was thinking of.

 

[mnealtx]

I disagree with the posting about startup times and costs. Modern digital SLRs have startup times less than 2 seconds - by the time you have the camera up to your eye, it's ready to take the shot.

As to cost - yes, film is cheaper than ever but there is still the processing cost for the film. The price difference between a decent film camera and a decent digital camera will be absorbed by the processing costs of 100-150 rolls of 36 exp film. Prints, of course, would be the same cost. Digital has the advantage of being able to change "speed" in midstream - going from ASA 100 to ASA 400 in a film camera requires wasting at least part of a roll of film.

A final thought, for those considering digital - while both Canon and Nikon are good cameras, Canon has less noise at the higher ISO speeds.

 

[mnealtx]

Amongst other things Ken Rockwell also discusses film SLRs and large and medium format cameras on his website.

I am not affiliated with Ken Rockwell in any way but I did stay at a Holiday Inn Express the other night.

There's plenty of better sites to get info than KR... he's been caught in quite a number of (shall we say) exaggerations. Sites like POTN, FM and others would serve you better.

 

[Roman414]

How long do you want the photos to last? Do you want your grandchildren to have a photographic record of what the world looked like in Grandpa's day? There are prints and negatives a hundred years old. What is going to happen to your digital images? If you want them to have any permanence, every few years you are going to have to copy them to new media as old media becomes obsolete. Could be a pain in the *** if you have thousands of shots. And every time you copy, you lose a little image quality.

 

[paydepst]

There's plenty of better sites to get info than KR... he's been caught in quite a number of (shall we say) exaggerations. Sites like POTN, FM and others would serve you better.

Really? I always found him to be spot on. Do you have an example to cite? I don't know why he would be motivated to embellish since he doesn't accept advertising.

 

[Haiku]

For my money, the best site on the net is rogerandfrances.com It's run by Roger Hicks and Frances Schultz. What those two don't know about photography, including digital, isn't worth knowing.

They have a photoschool with both free and paid lessons. Read the free lessons before spending the money to subscribe. It cost $20 per year to subscribe and it's the best $20 I've ever spent on photography.

 

[rusirius]

A few comments about some posts (since this is the first time I've even paid attention to this one). Anything not stated as fact is of course my own opinion...

Wil: Shoot time on a digital SLR often times can exceed that of a film camera. Take for instance the Canon 30D... It has a .15 second start up time. That may not be the 0 second start up time of a film, but you also have to consider focus time, etc. MOST digital SLR's (if you're using newer glass intended for digital SLRs) actually have much faster drives for the autofocus... Therefore it can actually get a shot off faster than with a film camera. Unless of course you're talking manual focus, but in that case a .15 second difference isn't gonna make any difference at all.

Ontario: You said "To process and print film costs the same as printing digital images." This is not true at all. To begin with, when processing film images, you're paying for the development as well as the printing. When processing digital images, you're paying only for the printing, thus it's by far a cheaper option. Further, when processing digital images you can process ONLY the images you want. You have no choice however than to develop an entire roll of film. Even if you choose not to print the entire roll, your costs are still higher. (Not to mention the original price of the film).

You also said "pick up a used film camera and a roll of 400 film and you'll be getting better pictures than from a $1000 digital". This also just isn't true. Leaving the debate that it's the photographer not the camera that determines the quality of the pictures behind, keeping all other things the same, almost ANY digital SLR will take "better" pictures than a 35mm film SLR. While it's true that the CCD in a digital camera is either the same size or in many cases smaller than a 35mm frame, the actual sensors on the CCD are actually smaller than a grain of film. I have printed 24x36 prints off my digital SLR even without post-processing to remove noise which had almost no grain at all in the image. On the other hand, I also have 24x36 prints from 100iso 35mm film which you have to stay back quite a ways to keep the grain from becoming rather distracting. Obviously this doesn't apply for the medium format and landscape cameras, but you'll not be picking up and operating any of those for pennies...

You also mention scanning the negatives if you want to manipulate the images on a computer. Have you ever done this yourself? If you have, have you ever compared it to an equivelent picture pulled off a digital camera? I have a film scanner I paid $1200 for USED, and it won't scan a negative with nearly the quality of an image from my digital SLR. That's WITH all it's extra processing it does to improve the image. Again the issue is grain. Try using one of the "negative" attachments that comes with your typical consumer grade scanner... Those things provide scanned images from negatives that are absolutely worthless.

LetterK: You mention the megapixels and the size of the prints. That's very true, but it's also important to mention that megapixels aren't everything... Many consumers get wrapped up in the whole "megapixel" frenzy... I'll never forget the time I was shooting with my 20D (8.4MP) and a guy started asking me questions about it, including megapixels and cost... Then immediately pulls out his point and shoot and says, "Yeah, guess you're kinda mad about all these new cameras coming out. I bought this one two weeks ago and it's 10 megapixels, but only cost me $150! Bet you wouldn't have thought when spending that much money that a few months later you could have bought a better camera for a fraction of the cost?"
Needless to say I busted out laughing so hard I couldn't even explain. He walked off kinda bewildered.
Megapixels mean NOTHING... Well okay, it means something... It means how many pixels they've crammed onto that CCD... It has NOTHING to do with the quality of the image it can generate. Only, as John pointed out, how large of a print can be made without seeing those pixels (i.e. like film grain). However, what is MUCH more important is the amount of noise the CCD generates. When all those sensors and capacitors are fired up, they can actually capture "noise" from their neighbors. This "noise" actually gets worse the tighter you pack these things in, which is why in a lot of cases, given an 8x10 print for example, a 4 megapixel camera can often make a print FAR superior than a 10 megapixel camera.
This is also why an 8 megapixel camera of one type or brand may only cost $100, while another costs hundreds more... It's not because of extra "features" on the camera, etc... It's because the CCD is much lower noise.

Farace: A photographic negative will always be able to be used? Negatives can have all sorts of things happen to them, the most obvious of which is scratches... Not to mention they're just plain easy to lose. You're looking at digital in a completely wrong way... First, it's completely medium independant... That's what makes it so wonderful... It's nothing more than a stream of bits (ones and zeros)... It can be transfered from one storage medium to another at whim... I have images on my hard drive, backup drive, memory cards, CD's, DVD's, memory sticks, you name it... I can go out to my PC and make a full backup of every photograph I've ever taken (including film that I've scanned) with a couple of quick clicks... How to you suggest making backups of those film negatives? Not with a few clicks I can tell you that... As technology changes, a few more clicks will be all that's needed to "migrate" to it... It couldn't get much easier than that. Also, CD-Rs and DVD-/+R's do have a finite life span, since the ink will break down over time.. Right now that's estimated to be around 150 to 200 years.

Roman414: You say, "Every time you copy you lose a little image quality". Again this couldn't be further from the truth... You're aligning your thinking back to analog technolgy (and in truth the same technology you're defending). It's ANALOG that suffers generation loss from copying... Not digital... Audio cassetes for instance loss parts of the data they represent each time you copy one... However, I can copy a CD hundreds of thousands of times over, and it'll still be EXACTLY the same as the one before it... Why? Because digital is just that, binary... Ones and zeros... If I have a byte (01100101) and copy it to a new byte it's STILL going to be (01100101).. It's not going to suddenly become (0[maybe one]1[almost zero]01[not quite zero][maybe 1])... I can copy it a billion times over, and it STILL stays exactly the same... (not to mention I can even introduce things like parity bits to actually PROTECT against any sort of loss suffered by a storage medium failure.)

 

[ouch]

You mentioned parity bits, which is one of the great advantages of digital transmission- error detection.

Any error detection scheme does basically this- a calculation is performed on the packet of information being sent at the transmitting end and the result of that calculation is sent along with the packet. At the receiving end, the same calculation is performed, and it is compared to the answer that was received. It may be as simple a calculation as "add up the total number of 1's sent, and determine if the sum is odd or even", or involve the heady mathematics found in cyclical redundancy checks (CRC errors). The fromer case is called a parity check, and will tell if you're off by one single bit (or an odd number of bits, for that matter). Errors in transmission involve a single bit over 96% of the time, so even this simple technique is effective.

What's more poweful than error detection is error correction. Calculations may be performed to isolate which bit (in a stream of predetermined length) has suffered the error. Then it's a simple matter of flopping the bit (1 for 0, or vice versa) in that slot to acheive error correction.

 

[jazzman]

almost ANY digital SLR will take "better" pictures than a 35mm film SLR. While it's true that the CCD in a digital camera is either the same size or in many cases smaller than a 35mm frame, the actual sensors on the CCD are actually smaller than a grain of film. I have printed 24x36 prints off my digital SLR even without post-processing to remove noise which had almost no grain at all in the image. On the other hand, I also have 24x36 prints from 100iso 35mm film which you have to stay back quite a ways to keep the grain from becoming rather distracting. Obviously this doesn't apply for the medium format and landscape cameras

That's true as far as it goes, but the one area in which 35 mm film still has an advantage is contrast. Film will record a greater range of dark to light "data" than digital. If the digital medium records pure white or black because it can't get that last portion of information, no amount of work on the computer will put it back, but if film catches it, a print can be made that shows it. In practice, this is almost meaningless for most photographers and most images, but I have a few framed prints that would have ended up as deleted files if I had used digital.

 

[mnealtx]

Really? I always found him to be spot on. Do you have an example to cite? I don't know why he would be motivated to embellish since he doesn't accept advertising.

A few examples:

"The low magnification of an ultra-ultra wide 14mm lens allows longer speeds than a normal lens." - um...no, it doesn't. With that said, you can get away with handheld at a slower shutter speed with a short lens, but it's it's not due to magnification.

He consistently says that "tripods are for the weak" - not a very astute thought for a purported landscape photographer, where exposures could be several seconds.

Then there's the whole "left handed Nikon" issue...

Try POTN for Canonites, or ByThom if you're a Nikon nut - you'll get better info.

 

[paydepst]

Point taken Mike. Still he was always spot on for the information I consulted him for. The left handed Nikon is a new one however. Good to have some more websites to consult however.

 

[mnealtx]

Don't get me wrong - I'm not saying there's not good info on his site, just that some of his comments are a bit questionable, that's all.

Have fun with the other sites - some great info there!

 

[icemncmth]

Also, CD-Rs and DVD-/+R's do have a finite life span, since the ink will break down over time.. Right now that's estimated to be around 150 to 200 years.


Actually CD-Rs and DVD...r's etc life span can be as
short as a week depending on how you store them...The companies that make them want you to think they will last a lifetime but I wouldn't trust them over a year in perfect storage. the bubbles in the gel begin to break down over time...and UV light, heat have a lot to do with it..

Roman414: You say, "Every time you copy you lose a little image quality". Again this couldn't be further from the truth... You're aligning your thinking back to analog technolgy (and in truth the same technology you're defending). It's ANALOG that suffers generation loss from copying... Not digital... Audio cassetes for instance loss parts of the data they represent each time you copy one... However, I can copy a CD hundreds of thousands of times over, and it'll still be EXACTLY the same as the one before it... Why? Because digital is just that, binary... Ones and zeros... If I have a byte (01100101) and copy it to a new byte it's STILL going to be (01100101).. It's not going to suddenly become (0[maybe one]1[almost zero]01[not quite zero][maybe 1])... I can copy it a billion times over, and it STILL stays exactly the same... (not to mention I can even introduce things like parity bits to actually PROTECT against any sort of loss suffered by a storage medium failure.)

Digital does loose bits as it is copied ...That is why when loading programs onto computers they sometimes don't work right. Most of the time the program used to copy a file is to blame...it also has to do with the buffer overflow. Bits can be lost when the ram fills up and buffers over to the drive...this happens more than you think on large pictures..there is a lot of information on that is being transfered and they do loose some information...

Heat build up is probably the biggest cause of bit loss....Processing large files, even moving, transfering them causes heat build up in a computer. As the information moves through the system the heat can cause some fluctuation in the quality of the information.

Also you have to look at the hardware itself. Let say you are going to backup a 100 gig HD full of your photo's...and the read head is not functioning properly...you could loose everything..

I have several computer customers and I do backup on their systems all the time...but I am redundant and this keeps my loss at a minimum.

Digital has it's faults and so does film...

It still will be a while before digital replaces my 8x10" large format...it makes a great contact sheet!...LOL

 

[rusirius]

Digital does loose bits as it is copied ...That is why when loading programs onto computers they sometimes don't work right. Most of the time the program used to copy a file is to blame...it also has to do with the buffer overflow. Bits can be lost when the ram fills up and buffers over to the drive...this happens more than you think on large pictures..there is a lot of information on that is being transfered and they do loose some information...

With all due respect, speaking as someone who has been involved with computers for the last 25 years, is the Director of Information Technology for a large company, have experience with networking, programming (C++, C#, VB, ASM, Pascal, etc), and yes, even designing both digital and analog circuits, let me be the first to say, You frankly have no idea what you're talking about.
Bits don't just "dissappear". A "bit" in a digital circuit is represented by a potential (a voltage). Let's take standard TTL logic for example. Generally a "0" is represented by a 0 voltage... a "1" is represented by a 5 voltage. When a "bit" is shipped off on a data bus, that bus is either raised to 5 volts, or left off at 0 volts. That's where transitors come in and make computers possible, they are "switches" that turn on or off.
In other words, when a transistor or "bit" is ON, the voltage doesn't just magically dissappear sometimes resulting in it being read as a 0 instead of a 1...

RAM is a different story... current RAM technology is "charged" much like a capacitor. Much like a battery or capacitor, that charge can "fade" over time... This is precisely why RAM is "cycled"... In other words, every so many nanoseconds, every single bit stored in ram is read, and then re-written. This "refreshes" the bit for that next cycle. Further, RAM does use CRC checks and parity to ensure that even if by some chance a bit is "lost", it is corrected...

Programs not working once they are loaded have NOTHING to do with "lost bits"... It can have any number of a thousand reasons... It could be an incompatibility with any of the hundreds of dynamically linked libraries... It could be an incompatibility with certain hardware, hardware drivers, operating system, etc...

You mention "buffer overruns"... First of all, a "buffer overrun" has nothing to do with running out of ram and it "spilling" over to the hard drive. When you are getting low on system ram, most operating systems utilize what's called a "page file". This page file or paging ram is set up to model system ram in how it's accessed, but instead uses the hard drive as a medium. The OS will take the memory allocations (every program running must allocate certain areas of memory by utilizing calls to the OS. If the application wants more ram, it must make another call to alloc more, etc) Anyway, it will take the memory allocations for a specific program (it tries to choose processes that are dormant) and move their memory allocations over to the page file. Still though, this does NOT result in any data loss in these processes. In fact, if you understand the structure of a file (take for instance a .jpg file), especially a compressed file like a .jpg, you'll understand that "losing" bits doesn't result in less quality, but rather a "major destruction" of the image... I urge you to try the folowing... Take a .jpg file, any image you want... Open it in a byte editor. Now just change a few bytes here and there and save the file back out. What you'll find is that you didn't just change a few pixels, or change the brightness, contrast, etc... Instead you probably ended up with HUGE bars of solid colors, or areas which are completely unreadable.
In other words, "loseing bits" is NOT common, and when it does happen, the results are NOT just "small degradations", but rather actual destruction of the source.

Coincidently, stepping back for a moment, a "buffer overrun" has to do with program code. Remember I told you when writing a program you must allocate every area of memory you intend to you? Occasionally a programmer will make a silly mistake... For example, if I allocate a variable for a numeric value to be a double (assuming a 32-bit store, a double would be 64-bit, or 8 bytes). A double has 52 bits for the mantissa, 11 bits for the exponent, and 1 bit for the sign. With that in mind, the absolute largest number I can store in a memory space designated as a doube is 1.7976931348623157 x 10^308. Now let's say I have a variable called "dblMyNumber" that currently has the above number in it, and then I do something silly like this:

dblMyNumber += dblMyNumber (<--- Adds dblMyNumber to itself and then stores the result back into dblMyNumber)

or for that matter even:

dblMyNumber += 1 (<--- Adds 1 to dblMyNumber and stores the result back into itself)

What results is a "buffer overrun"... My number no longer fits within 8 bytes of space, so what happens is, the extra bytes "run over" or "buffer overflow" into another section of memory... If my process is running with high enough rights, this can even be an area of memory that has nothing to do with my process.

In fact, this is why "buffer overrrun" errors are most commonly used to "hack" into a system. If let's say a programmer defines 60 bytes for a string array, but they never "limit" or make sure that an input field doesn't exceed 60 bytes, a very LONG string can be passed to the process... It can contain extraneous information that can overrun even into the system areas, and cause code execution, etc that results in administrative rights being granted, shells firing up under the root user, etc...

Anyway, I digress...

As the information moves through the system the heat can cause some fluctuation in the quality of the information.

Once again, there is no "quality" of the information... it's all boiled down to 1's and 0's... Heat CAN cause a system to fail, but not to lose "quality" in it's files.

Also you have to look at the hardware itself. Let say you are going to backup a 100 gig HD full of your photo's...and the read head is not functioning properly...you could loose everything..

Sure, in the case of a hardware failure something like this could happen... But to begin with, if the files can't be read due to a hardware failure, the OS is going to be kind enough to notify you that it couldn't read the file(s). I don't think anyone on this forum is silly enough to "backup" their files, ignore all the "source file could not be read" messages, and then proceed to delete all the original files...

Yes files can be lost due to hardware malfunctions... That's the point of backups in the first place... But they do NOT lose quality... .jpg, .raw., .tga, .tiff, or .psd images don't just suddenly start looking fuzzy or losing detail because they've been copied too many times... .wav's, .mp3's, .aiff's, etc don't just start sounding worse and worse because of excessive copies... My word documents don't just start looking more fuzzy when they are printed, or having A's start to look like a pre-schooler drew them...

In the worlds of 1's and 0's, there's no in between...