How much does an SD card affect quality of photos?

[Stacie]

Does certain SD card have much affect on the quality of photos? What ones do you recommend? Expensive and inexpensive. Thanks!

 

[fjrabon]

no, a file is a file. What differences in quality of card effects is failure/corruption rate, speed and capacity.

I don't care about brand, to me that top 10 or so brands are more or less the same. SD cards at this point are a commodity.

Figure out what speed you need/your camera supports. Figure out how much space you need/want on it. Buy the cheapest one you can find that hits those specs.

I typically buy 95 mbps 32 GB cards. They're the sweetspot to me. I don't own any cameras that will write faster than that anyway, so anything faster will be wasted, but I definitely prefer the fastest cards my camera can use. 32 GB is big enough that I almost never need to worry about changing cards, but small enough that I won't let it build up until it's a completely unmanageable amount on there. Also, it doesn't break the bank if a 32 GB card corrupts (i've only had it happen a couple of times, thank god for the D7000D7100 having dual slots).

 

[Stacie]

Thank you. I have a Canon T2i. With this lens. Canon EF70- 300mm f/4-5.6 USM I want something all around, but prefer something better for action.

 

[bratkinson]

When they work, the quality of one SD (or CF) card over another brand is of no concern.

But, the difference is how often they fail...damage or completely lose photos, lose directory information, etc. The higher quality card(s) have a much lower failure rate. In short, you get what you pay for.

As far as buying faster cards, perhaps as high as 95 mbps, unless one shoots movies and or continuous mode, the higher speed is generally not an issue for the average photographer. BUT...my experience with RAM for computers tells me that the higher speed-rated cards are better construction quality overall. It was (and perhaps, still is) true that RAM that is rated for lower speeds is often higher-speed RAM that didn't successfully handle the higher speeds, but at slower speeds is OK. For this reason, I buy PC-2100 RAM when all that is required is PC-1600 on my computer (I overclock somewhat, so it's running at PC-1866 speeds), and buy 95mbps memory cards for the same reason.

 

[Derrel]

I buy only SanDisk or Lexar. They are rock solid.

 

[chuasam]

I've had a Sandisk and a Lexar fail on me... As well as a SiliconPower. The biggest difference is how quickly Sandisk and Lexar took care of me and send a replacement within days.

 

[table1349]

Sandisk only. Lexar's had an issue with Canon cameras many moons ago when I went digital. Problem is fixed now, but I got used to and have a number of Sandisks and never had an issue so I just keep on buying them.

 

[table1349]

When they work, the quality of one SD (or CF) card over another brand is of no concern.

Au Contrair there Pierre! Rob Galbraiths extensive testing disputes that.

Rob Galbraith DPI CF SD XQD Performance Database

Many others have done similar tests that with newer camera bodies.

 

[Alexr25]

When they work, the quality of one SD (or CF) card over another brand is of no concern.

Au Contrair there Pierre! Rob Galbraiths extensive testing disputes that.

Rob Galbraith DPI CF SD XQD Performance Database

Many others have done similar tests that with newer camera bodies.

Where in that article does he dispute that??
Cheap SD/CF cards may have shorter life spans or slower read and write times than high end cards but as long as they work memory cards can in no way affect the image quality, only the camera can do that. That is the advantage using digital capture.

 

[unpopular]

It seems that large cards fail more often. Maybe there is just more there TO fail, but it seems that when you read about card failure, it's never on an 8gb card... is there any truth to this?

 

[Alexr25]

Large cards would be using multi-level storage cells in their flash memory where each storage site can hold more than a single bit of data, older lower capacity cards store 1 bit per each memory cell. This multi-level cell technology makes it possible to make high capacity memory cards cheaply but error rate increases and card reliability does suffer.

 

[Solarflare]

Does certain SD card have much affect on the quality of photos?

Thats exactly the point of digital: all data is either a 0 or a 1.

So if the card has the (analog) values 0.8, 0.3, 0.9, 0.7, 0.9, 0.1, 0.2 stored the data will still be retrieved as 1, 0, 1, 1, 1, 0, 0, making any data degredation much less significant than with analog recording.

Thus, either you still have the data - or its completely gone.

Thus all the SD card can do is
(a) provide data security - dont lose the data, possibly even under hardened conditions (long storage time, radiation, cold, heat, water, corrosion, ...)
(b) provide writing speed - well also reading speed, but writing speed influences how fast photos can be stored on the card during actually shooting, while reading speed only helps with reviewing photos afterwards.

 

[chuasam]

It seems that large cards fail more often. Maybe there is just more there TO fail, but it seems that when you read about card failure, it's never on an 8gb card... is there any truth to this?

Both of the cards that failed on me were both 8gb

 

[unpopular]

So if the card has the (analog) values 0.8, 0.3, 0.9, 0.7, 0.9, 0.1, 0.2 stored the data will still be retrieved as 1, 0, 1, 1, 1, 0, 0, making any data degredation much less significant than with analog recording.

Just to clarify. There are two distinct parts of a digital camera - the analog side and the digital side. Analog data is continuous signals, 0-5v and everything in between. The sensor measures and amplifies information as an analog signal. From there, the analog signal is converted to a digital one that is recorded by a memory card.

The conversion from analog data happens inside the camera, before it reaches the card. The card has nothing to do with the analog information that is measured by the sensor. Either the information reflects the digital information it was given by the camera, or it doesn't. When it doesn't, that's what we call "file corruption" and in general, this makes the file unreadable.

More often though the corruption happens with the files that say "This is an SD Card" than it does your images. If the computer has no way of knowing that what you stuck into it is an SD card, there is no way to get the images off. In this case, you'd need to format the card, which involves erasing it.

I'd imagine also that physical damage could happen, rendering the card completely unrecoverable.

 

[soufiej]

Does certain SD card have much affect on the quality of photos?

Thats exactly the point of digital: all data is either a 0 or a 1.

So if the card has the (analog) values 0.8, 0.3, 0.9, 0.7, 0.9, 0.1, 0.2 stored the data will still be retrieved as 1, 0, 1, 1, 1, 0, 0, making any data degredation much less significant than with analog recording.

Thus, either you still have the data - or its completely gone.

Thus all the SD card can do is
(a) provide data security - dont lose the data, possibly even under hardened conditions (long storage time, radiation, cold, heat, water, corrosion, ...)
(b) provide writing speed - well also reading speed, but writing speed influences how fast photos can be stored on the card during actually shooting, while reading speed only helps with reviewing photos afterwards.

That argument is very old and very misleading IMO. Yes, digital recording/playback is an on/off affair within the digital circuits. On/off is represented by a 1 and a 0 code. Signal/no signal. Works great on paper as theorized back in the 1930's when circuits couldn't process data fast enough to test the on paper theory. Real world conditions though suggest digital is not perfect and, in fact, can never be "perfect". Nor is analog perfect. What's very much not perfect is the conversion process which takes an analog waveform and converts it to a digital set of data points and then the digital process which returns the data to a usable analog signal.

Sample the same analog "write" signal five times and you'll find five different results of the A to D conversion. Sample five D to A conversions and you'll find five different "read" versions of the same data. Read/write errors, error correction, buffers, filters, aliasing, phase shifts, etc all place their stamp on "digital".

One aspect of digital which has always bothered me is the focus on one single piece of the digital pie. You cannot divorce the storage media from the write/read processing system. In general, I'd agree with the idea you typically get what you pay for, both in equipment and supporting devices. Paying more typically gains you tighter control over the processes which create the device. Fewer errors and more consistent performance over time are to gain.

However, any one device (a SD card) is simply a part of a system and it is the performance of the whole which is of concern to the end user. Since on paper specs are often misleading and incomplete, relying on the system's designer to suggest which individual components are best for their system is about the only way to begin. Canon suggests you use cards they have tested and approved for their cameras. Buy those cards and you'll likely have the best system performance from your camera. Buy something other than those cards and you have placed your money down and you've said you're willing to take your chances.

The idea "a file is a file" is not quite the truth either. As I have said in other such threads, a digital file is similar to a CD sitting on a shelf. What do you get from a CD sitting on a shelf? Nothing. It isn't until you place the CD into a processing system that you have anything tangible to speak of. For example, until Fuji began to produce cameras which didn't rely on a "conventional" sensor construction, the typical digital image file was constructed and deconstructed using an aliasing filter - a low pass filter - through which the data had to pass. Filters are known to cause phase shifts which in turn cause errors in the implementation of the data. Yet filters are also used to correct errors within the digital processing system itself. It's a classic Catch 22 when it comes to digital systems. Either you use a filter to correct the errors inherent in the system or you forgo the filter and allow those errors to pass unchecked to the successive circuits where those errors might even be compounded.

The debate over whether aliasing artifacts are more bothersome than the effects of the anti-aliasing filter will continue since many digital adherents believe digital was created as the perfect storage system and, therefore, what digital is on paper must be what digital is to the human perception systems. Yet the results of digital 1 and 0 data processed with the filter and without the filter yield quite different results; Non-low-pass filters explained Goodbye to anti-aliasing TechRadar

I would think the data as interpreted by such diverse systems would lead us to see that a file is not just "a file". It is the processing of the file by a system - both in read and write functions - which makes a significant difference in the end result. For the purposes of this thread, staying with the designer/manufacturer's suggested parts of that system is your best advice unless you have the actual test gear and knowledge to say otherwise IMO.

If you use a Canon system, buy the SD cards Canon has tested and approved. Buying a "faster" card simply means the system can buffer the data at a faster rate of speed. How fast do you require your camera to process data? If you are the rather typical photographer who shoots, reviews and then shoots again, you do not need a fast buffering action. If, however, you shoot in continuous frames and that buffering speed is important to your results, then spend the money for the faster card. For most of us, "fast" just isn't a requirement.