camera USB or compact Flash card reader

Oh ... and to answer the OP ... I have been using various card readers instead of the USB to camera (which takes 8-10 times longer).

JChin wrote in post #9659560
I strongly recommend AGAINST your suggestion to "move" the photos. When you "move" the photos, you are effectively doing a "copy" and then a "delete", which adds an extra write cycle to your NAND flash chips.

For years now, I've been copying the photos to my computer and then doing a "quick format" of the flash card. A "quick format" only rewrites the FAT table on the card and not touch the other NAND cells, this limits the "wear" on the flash cells.

Wow. A little 'knowledge'....

1. When you delete a file the actual data of that file isn't changed at all. The only thing that happens is that one byte of the FAT table is changed to signify that the space reserved for that file is now free for use. So moving the files will not write to the card any more than doing a copy and format. In fact the formatting re-writes the whole FAT, so it's actually worse.

2. Even if it did totally overwrite the file - the SLC NAND normally used in fast CF cards is specified at 100,000 read/write cycles per address block. The cards also use wear-levelling to spread the load over all address blocks. So, with an 8GB card, you'd have to write 800 TB of data to reach that limit. That's 32 million 7D raw files. Let's be generous and assume a full overwrite on deletion. That's still 16 million images. Let's asume you use that 8GB for 10 years - that means you'd need to shoot 8767 images a day, every day.

So, to use up your write cycles you'd need to shoot one image every 10 seconds, all day (no sleep), every day, for 10 years non-stop.

I'm not going to lose a lot of sleep over it.

As a professional, I follow a rule that isn't a bad idea for an amateur enthusiast who cares about his images: I never mix two jobs on a card. When I finish a particular job, I remove that card from the camera as soon as practical and it goes into a "secure handling" regimen to get it my card reader and downloaded. Even then, I don't remove the images from it--it serves as an original source file--until I actually need to use it again. In the meantime, I'm using other cards for other jobs with that camera. This was the same practice I followed with film--I never mixed two jobs on a roll of film.

RDKirk wrote in post #9664312
As a professional, I follow a rule that isn't a bad idea for an amateur enthusiast who cares about his images: I never mix two jobs on a card. When I finish a particular job, I remove that card from the camera as soon as practical and it goes into a "secure handling" regimen to get it my card reader and downloaded. Even then, I don't remove the images from it--it serves as an original source file--until I actually need to use it again. In the meantime, I'm using other cards for other jobs with that camera. This was the same practice I followed with film--I never mixed two jobs on a roll of film.

Good point mate.. When I worked for a newspaper I would create a new folder for each job.. I just didn't have enough cards for each gig..There were days I was shooting eight jobs before I could get back to edit and file..

I have never heard of a bent or broken pin within a department of 12 photographers over how many years??

yogestee wrote in post #9664681
Good point mate.. When I worked for a newspaper I would create a new folder for each job.. I just didn't have enough cards for each gig..There were days I was shooting eight jobs before I could get back to edit and file..

When I worked on a paper in the early 70s, I always unloaded the camera and marked and stashed the roll after an assignment, then immediately loaded a new roll. Back at the office, the editor might have already decided which job had priority, and that roll would go into the soup first. We developed film in small booth-like darkrooms that had 4x5 tanks of chemistry: D-76, HC-110, stop, rapid fixer. We loaded the reel(s), slid them onto lift rods, set the ubiquitous Gralab 300, and just dunked them into the tanks in the dark.

Card reader for me.

In the process of transferring data between the card and the hard-drive, I want as little software as is possible in order to minimize the risk of corruption.

USB for me, card reader for my gf.

birdfromboat wrote in post #9629248
I was told to use the card reader, for this one reason alone: the camera is a computer, the USB is a link to another computer, that 2nd computer is on the internet alot and full of downloads. If I don't have to link my camera to a computer that might hold unknown who knows what, why would I? I pull the card, load it in a reader that is integral to my computer, and move the files, not copy, into a file on the computer and onto a CD. That leaves the card empty and i reformat it when I put it back in the camera. Also, I have ZERO corrupt files.

hollis_f wrote in post #9664213
Wow. A little 'knowledge'....

1. When you delete a file the actual data of that file isn't changed at all. The only thing that happens is that one byte of the FAT table is changed to signify that the space reserved for that file is now free for use. So moving the files will not write to the card any more than doing a copy and format. In fact the formatting re-writes the whole FAT, so it's actually worse.

2. Even if it did totally overwrite the file - the SLC NAND normally used in fast CF cards is specified at 100,000 read/write cycles per address block. The cards also use wear-levelling to spread the load over all address blocks. So, with an 8GB card, you'd have to write 800 TB of data to reach that limit. That's 32 million 7D raw files. Let's be generous and assume a full overwrite on deletion. That's still 16 million images. Let's asume you use that 8GB for 10 years - that means you'd need to shoot 8767 images a day, every day.

So, to use up your write cycles you'd need to shoot one image every 10 seconds, all day (no sleep), every day, for 10 years non-stop.

I'm not going to lose a lot of sleep over it.

Not to banter over this ... however, just to clear something up for some novice.

Writes to the flash are not done in NAND "address blocks", single bits or even single bytes. They are written by the OS (camera OS or computer OS) in clusters of blocks. So just deleting one file actually does rewrite entire the FAT cluster (which contains several NAND "address blocks"). Thus deleting multiple files is causing the FAT cluster to be rewritten multiple times.

Depending on the logic in the flash card, that FAT entry cluster maybe in a restricted section of the flash card or it may roam within a restricted area on the flash card (similar to the first track on a hard drive). In most cases (as far as I know) that restricted area does not float over all the address blocks on the card. This is the reason why so many flash cards fail when people try to use them as a cheap SSD for a computer (using those internal CF to IDE/SATA adapters).

Oh, I also believe that most of the CF and SDHC cards are MLC NAND, not the more expensive SLC NAND. The typical MLC NAND flash card lifespan is like 5K-10K erase-write cycles.

CF and SDHC cards are cheap today, so buy extras. However, I would hate to have one fail on me prematurely or in the middle of a shoot.

I've been told by several pro photographers:
1. never delete images on camera (especially not during a shoot)
2. use a card reader to copy (not move) all images to your hard drive and then format the card
FWIW, these rules have worked for me flawlessly for years.

One reason I don't delete images in the camera (especially during a session) is because too many times I've found an image that didn't appear useful at first glance to be useful later during editing. For instance, it might have showed the subject blinking--but have everything else in a more perfect position than any of the other shots of that pose. So the shot I thought was better at the session becomes mere "swap fodder" source for the subject's eyes...and the shot I thought was useless becomes the main image.

Something like that has happened more than once, so I give every image a chance to be seen a second time, outside the heat of the session.

Card Reader, if the camera loses power during photo transfer, you risk corrupting your photo(s), which would be irretrievable

The card reader is powered by the computer's usb port

JChin wrote in post #9666029
Not to banter over this ... however, just to clear something up for some novice.

Writes to the flash are not done in NAND "address blocks", single bits or even single bytes. They are written by the OS (camera OS or computer OS) in clusters of blocks. So just deleting one file actually does rewrite entire the FAT cluster (which contains several NAND "address blocks"). Thus deleting multiple files is causing the FAT cluster to be rewritten multiple times.

Ah, yes. I should have remembered that.

JChin wrote in post #9666029
Oh, I also believe that most of the CF and SDHC cards are MLC NAND, not the more expensive SLC NAND. The typical MLC NAND flash card lifespan is like 5K-10K erase-write cycles.

Yeah, I couldn't find anything definite on this. Just a few links (eg) that say that the fastest cards use SLC. But even 5,000 cycles equates to writing 40 TB of data - still over a million images.

JChin wrote in post #9666029
CF and SDHC cards are cheap today, so buy extras. However, I would hate to have one fail on me prematurely or in the middle of a shoot.

I've been told by several pro photographers:
1. never delete images on camera (especially not during a shoot)
2. use a card reader to copy (not move) all images to your hard drive and then format the card
FWIW, these rules have worked for me flawlessly for years.

I've only had one CF card fail one me in over 15 years. And that was my own stupid fault (I pulled it out of a PDA while it was being writ to). Those two rules are good advice (although I'd expand rule 2 - copy the images, make sure they've copied OK, then format the card.)

hollis_f wrote in post #9671503
...
But even 5,000 cycles equates to writing 40 TB of data - still over a million images.

My point was that the area in flash where the FAT is stored is constantly getting overwritten. Those NAND cells gets the most erase-write cycles and would go first, thus making the card useless (kind of like track 0 on a hard drive going bad).

hollis_f wrote in post #9671503
...
Those two rules are good advice (although I'd expand rule 2 - copy the images, make sure they've copied OK, then format the card.)

Very good point! It would be a good idea to at least verify the copied images are the same size as the originals on the card before formatting the card.

JChin wrote in post #9672183
My point was that the area in flash where the FAT is stored is constantly getting overwritten. Those NAND cells gets the most erase-write cycles and would go first, thus making the card useless (kind of like track 0 on a hard drive going bad)

Is the FAT always in the same place on a solid-state drive? I'd have thought that would have been handled by the wear-levelling firmware. If it isn't then 10,000 cycles is very worrying. I'm sure there are pros that would exhaust a card in a month or two if that was the limit.

hollis_f wrote in post #9672285
Is the FAT always in the same place on a solid-state drive? I'd have thought that would have been handled by the wear-levelling firmware. If it isn't then 10,000 cycles is very worrying. I'm sure there are pros that would exhaust a card in a month or two if that was the limit.

As far as I know, the FAT is always within a certain area of the flash card. The wear-leveling algorithm does move it around, however, it still resides within a "restricted" section. It was explained to me that is similar to using a different area of "track 0".

This is the main reason why people who have used CF cards (with CF-to-IDE adapters) as a cheap SSD drive for an operating system (such as Windows) has experienced "drive failure" within a few months of daily use.

True SSD hard drives, according to my in-person discussion with an Intel technical rep., uses different algorithms than traditional flash cards, as well as "different NAND flash" (assumed to be SLC) chips. He said, at least that is what Intel does. He also noted that there are some SSD manufacturers that use the cheaper NAND, but allocate a larger portion for "track 0" type usage within the wear-leveling algorithm so that they don't fail as quickly. Guess you get what you pay for. Maybe that is why Intel SSD drives cost so much more than others.

JChin wrote in post #9672715
True SSD hard drives, according to my in-person discussion with an Intel technical rep., uses different algorithms than traditional flash cards, as well as "different NAND flash" (assumed to be SLC) chips. He said, at least that is what Intel does. He also noted that there are some SSD manufacturers that use the cheaper NAND, but allocate a larger portion for "track 0" type usage within the wear-leveling algorithm so that they don't fail as quickly. Guess you get what you pay for. Maybe that is why Intel SSD drives cost so much more than others.

Intel do two types of SSD, the M-series uses MLC and the E-Series uses SLC. The M-series drives are expensive, the E-series drives are stupidly expensive.