Megabytes Versus Megapixels

Megapixel Chart

One of the more common questions I get (I think it’s due to my technical background) is one where people are asking how many images they can store on their media cards.  The answer, as always, starts off with an “It depends…”.  For the quick feed readers curiosity, here’s the laundry list:

1.  File Format
2. Quantity of Light
3.  Varying Degrees of Color
4.  Bit Depth
5.  Megapixel Count (Resolution)
 

And for the more detail-oriented, here’s the extended version…

1.  File Format

Just one of the many considerations here is how you are saving your images.  There are also many facets in the “how you save your images” too.  For instance, RAW as a file format will always have more data in it than its more lightweight sRaw counterpart. Even further, jpg does a certain degree of compression in camera to help save on file storage space, so it will also decrease your file size usage.

2.  Quantity of Light

Another consideration to factor in is whether you are shooting in low light or bright light.  With digital photography, the more light you have in a scene, the more data there is to the image.  Conversely, darker images will have less data and take up less storage space on both your media card and your computer. Take, for example, this series of images I took of “Dino” outside on Sunday.  The exposure is set to under-exposed by 2 stops, neutral exposure, and over-exposure by two stops, as defined by my shutter speed.  The amount of storage space that was consumed on both the media card and my computer is indicated beneath each image.

Exposing Dino

3.  Varying degrees of color

The amount and types of color can also factor into how much storage space an image takes up.  I’ve actually addressed the issue of color in these exact terms before, so for a more thorough explanation of that, check out this article here.  Here, the summary is really the only relevant part, where green encompasses the largest amount of data, blue comes in second, and red encompasses the least.  Other color hues will fall somewhere between these three primary colors, so storage space will be a function of colors in your images as well.

4.  Bit Depth

Another factor that will enter into play (mostly in post production though) is that of bit depth, which is basically how you are saving your file out from processing.  Most cameras will capture in 16 bit depth, and will be imported in Photoshop or Lightroom at their native bit depth, unless you manually change it from 18 bit to 8 bit (which a lot of people do when using the full version of Photoshop, because that enables tools that are not available in 16 bit mode).  You can also output to24 bit or 32 bit mode, but these are mostly used for offset printing (think CMYK) and HDR imaging, which while popular does not speak to standard storage space for images captured natively in camera.  So, the bit depth is really beyond the scope here, but if you’d like to learn more about various bit depths and their usage, feel free to get started at the Wikipedia article here.

5.  Megapixel Count (Resolution)

The final element to consider in how much storage space an image will take up on either a card or a computer is the megapixel count.  Now, unfortunately there is no direct correlation from megapixel to megabyte as one is defined by the resolution of the image and the other is defined by a byte in computer terms.  While the former is mealleable depending on sensor type (CMOS vs CCD vs Foveon, etc.), the latter is pretty well delimited as a byte is a single unit of data.  So

Having said all of the above – raw versus jpg, light versus dark, one color versus another, bit depth, and megapixel count…there are some general rules of thumb we can draw based on significant research that has been done in this area (and by significant, I mean me hitting Google, Wikipedia, and various communities, asking if anyone knows of any authoritative resources I could check out).  The upshot is that images will largely be a function of their megapixel count.  Now, because of the variances in the other factors here, there is no hard and fast rule that is set in stone for image file size relative to MP count, but on a very rough scale, each megapixel of data will typically contain about a megabyte of data.  So the conversion is almost a 1:1 ratio.  Keep in mind of course that this is very rough, because I have seen with my 10MP camera that I have raw file sizes in excess of 17 MB!  It’s always better to work within an expected range, rather than using hard and fast rules anyway, so for that reason, here’s a chart:

Megapixels Resolution File Size
1.6 Megapixels 1536×1024 px 1.6-2.4 MB
2.8 Megapixels 2048×1365 px 2.8-4.2 MB
6.3 Megapixels 3072×2048 px 6.3-9.4 MB
10.1 Megapixels 3888×2592 px 10.1-15.1 MB
11.2 Megapixels 4096×2731 px 11.2-16.8 MB
17.5 Megapixels 5120×3413 px 17.5-26.2 MB
25.2 Megapixels 6144×4096 px 25.2-37.7 MB

Keep in mind that these formulas are very generic in nature as the methodology is not completely scientific, but can help you determine the expected capacity of your media cards for photos!

So, the natural extension of this takes us back to the original question:  How many images you can store on a media card given a certain pixel count?  Extrapolating things out is just a simple matter of math and Excel! 🙂

Megapixel Chart

With smart phones exploding their own megapixel counts, many are also now looking to use the MP count as the benchmark for identifying image quality, thinking that more is always better, right?  Again, as a general rule of thumb, this is true, but there are laws of diminishing returns.

Think of it this way – a sensor in a camera is a finite size – it’s not going to change substantially within a given form factor.  So, an SLR will have a certain size of sensor, a point and shoot will have another sensor size, and a cell phone (smart or otherwise) will have yet another size of sensor.  If you had to choose between an SLR that has 10  megapixels in any given photo (of roughly 10-15 MB of data), is it safe to assume that this will produce the same quality image as a 10 MP camera on a cell phone?

The answer, of course, is no.  Sensor size is really what matters here because you can capture much more data in a megapixel (or a megabyte for that matter) if it’s a bigger size.  So again, bigger means better! 🙂  What this hopefully tells you is that you can get some pretty big photos coming off a pretty tiny camera, and still get pretty lousy results.  In general, there are limits to what really matters on any given sensor size, because even though you can pack more megapixels on a sensor, the image quality really doesn’t return that much better a result after a certain threshold is reached.  What is that threshold?  Excellent question!

The answer:  It’s subjective, and open to interpretation, but here’s my take:

  • Smart Phone Cameras – The sensor is teensy tiny, so anything above 8-9 MP is just for fluff
  • Point & Shoot Cameras – A somewhat bigger sensor, and with the RAW capability, I’d say these can see benefits up to the 16-20 MP range…
  • SLR Cameras – With the biggest sensor in the category of portable cameras (I would not define a medium format or large format camera as “portable” in most scenarios_, these are seeing pixel counts in the area of 25-30 these days…a bit overzealous, and perhaps there is a difference, but certainly not for the purposes I use images for.  Even for stock images, I wouldn’t be using MP counts that high.  The logic is that if you start with a larger file, you have more capability to crop.  My response is, if you need to crop, you didn’t frame it right to begin with!

Until next time, keep on shooting!

How Fast Can You Shoot?

Whole Lotta Shaking

Whole Lotta Shaking

On initial glance from the title, one might think this blog is segueing into a dialog from the Top Shot show over on the History Channel (very cool show btw), but I’m actually referring to the notion of speed often associated with shooting digital.  Sports shooters, action shooters and the like often will get the fastest cameras, the fastest cards they can get, with the fastest processors, and go to all sorts of degrees to eliminate bottlenecks in their capacity to shoot fast and on the go.  To an extent, their efforts are justified, but how fast does your card need to be?

As you may recall on Monday, I talked about how the real meaning of memory in media cards.  We talked about Megapixels and Megabytes, and I gave some real world number ranges for what you could expect a media card to handle.  So, today, we’re going to follow up on that topic of discussion and take a look at speed ratings for cards.  As always, there’s a lot more to it than meets the eye, primarily because there are lots of numbers bandied about when speaking about card transfer rates.  Before we wade hip dip into this, one brief note here is to give serious props to Rob Galbraith for compiling a pretty extensive database on his website of data transfer rates for a whole slew of cards with a whole slew of cameras.  If you really want to dig into the minutia of how fast cards really are, be sure to check it out!  Just one note though, that this database has not been updated since 2008 so cameras made since then will not be found with any reliable information.

With that in mind, today it’s probably more helpful to walk through the various points that can contribute to data transfer rates for cards.  The natural first stopping point is in the camera itself, in what is commonly known as fps, or “Frames Per Second”.

Frames Per Second

Quite simply this refers to the number of actuations a given shutter in a camera can cycle through in exactly one second.  As indicated above, the abbreviation for this is typically fps, and is easily culled from the spec sheet for pretty much any camera on the market today.  For those of you that want to see a compilation though, here you are (I should note that I grabbed all this data from the vendor websites on claimed fps for the highest resolution image settings.  If you set your file size to sRaw, or jpg and add compression, your max frame rate may increase…:

Vendor Camera FPS Vendor Camera FPS
Canon 1D Mark IV 10 Nikon D3S 9
Canon 1D Mark III 10 Nikon D300S 8
Canon 7D 8 Nikon D7000 6
Canon 50D 6.3 Nikon D3X 5
Canon 60D 5.3 Nikon D700 5
Canon 1Ds Mark III 5.0 Nikon D90 4.5
Canon 5D Mark II 3.9 Nikon D5100 4
Canon T3i 3.7 Nikon D5000 4
Canon T2i 3.7 Nikon D3100 3
Canon XSi 3.5 Nikon D3000 3.0
Canon T1i 3.4
Canon T3 3
Canon XS 3

 

The bottom line here though is that with most DSLR’s on the market, if you expect more than 10 fps, that’s probably not going to happen (assuming also you are shooting continuous and on the highest resolution setting.)

So, even the camera you have can cause bottlenecks in write speeds…if your camera can’t write data that fast, it really doesn’t matter how fast your card can write data if it doesn’t have the data to write.  Now granted, each camera will have different MP counts so the Canon 5D (for instance) will of course have a slower fps rate because it’s a larger file, and it’s a full frame camera, so will naturally be slower than the 7D by comparison simply because it has to be.  So what we really need to be measuring is the amount of data that is being transferred per second (see now why I did that article first? 🙂 )  So, we’ve reached the second point of bottleneck now in talking about transfer speeds and that’s the buffer.

Buffer

The best analogy I have is that the buffer in a camera is very much like the RAM in your computer.  It’s a sort of staging area, where data is stored before it gets actually sent to the processor and then saved to the card. Without the buffer in place, fps rates would drop dramatically because there would be no place for the camera to temporarily hold data before writing.  The buffer is what allows you to keep shooting.  So, this brings up two important questions:

1.  How can you increase the buffer?  (You can’t really…)

2. How can I tell how much of a buffer my camera has?  (As of this writing, I am not sure if this metric is reported consistently by vendors, except when reviewers say “an increased buffer size improves performance”…which is still relatively meaningless.)

Since it’s not something we can really measure, nor change without buying into a whole new camera, here it’s just sufficient to say that this is the second bottleneck point, and is usually where you will run into moments of pause.  Why? Because cards are usually transferring data that fills up the buffer, and at that point the camera can’t take in any more data. When this happens your camera won’t let you take any pictures.

The other factor that comes in to play though ties more to the card itself and not the camera, which is the speed factor.

Speed Class Rating

Media card vendors like Lexar, Sandisk and others like to use terms to define their speed.  Catch words like Extreme, Pro, Extreme Pro, 300x, 600X and all sorts of fancy jargon is used for marketing purposes.  Notice the various cards on the market – the more buzz words, usually the faster the transfer rate claim, and of course, the more expensive the card!  🙂

So, rather than pour over all the various vendor semantics, I figured it’d probably be better to stay on what is a more neutral metric – speed class ratings.  While we could also introduce variances between the CF and the SD format in terms of write speeds, since most devices are moving to the SD format and it’s smaller cousins (mini SD and Micro SD), these are likely the ones we’ll see more often in the future, so the speed class ratings here are most relevant.

To that end, there are 5 different ratings or grades given to SD cards.  These are 2, 4, 6, 10, and 1 respectively.  The last rating (1) is reserved for the SDHC and SDXC card types, and has a special designation as UHS, for Ultra High Speed…while the lower ratings all have transfer rates classified as normal and high speed.  A full chart is viewable on it here.  In a nutshell, the higher the number, the faster the card can transfer data, so keep that in mind as you shop for media.  The ultimate barometer really though is not how fast you need to capture, but more what you need to capture, and as the chart indicates, video needs faster transfer rates than stills, so naturally the higher ratings are intended primarily for videographers.

If you really wanna geek out on data transfer rates, a better place to go for that is the Rob Galbraith database I mentioned upthread.  Again, it’s not been updated since 2008, but the general trends are probably consistent with the current market of vendors we have to choose from.


One final note on data transfer rates…card technologies have changed substantially in recent years, enough that there are now cards referred to as UDMA cards.  While the current generations of cameras from both Canon and Nikon support this mode of reading and writing data to media cards, older cards may not and will read the card at the slower rate. So, if you have an older camera, you may want to check for UDMA compatibility before getting a newer UDMA style card.

 Conclusion

So, what’s the takeaway from all this?  Hopefully three things:

1.  Transfer Speeds are not just a function of your media card, fps rates and buffer rates in cameras are factors too.

2.  Paying more for a faster media card might not be needed if you are not shooting video.

3.  If you are shooting video, or need the extra oompfh of speedy cards, make sure you are using the right speed class, and with the best camera you can.  As the old adage says: it’s a poor craftsman that blames his tools…

Happy shooting and we’ll see you tomorrow to wrap our discussion of Media Cards with a look at the various vendors!