One of the most common questions we receive is “What PPI is your system?” The short answer is “Whatever PPI you need, we have” but as with many things the full answer is a bit more involved. This article covers the full answer – a cup of coffee is recommended before taking it on.
Lies, Damn Lies, and PPI
First, before we talk about what PPI our systems are capable of we have to acknowledge that not all PPIs are created equally. In fact most digitization systems that claim to scan at a certain PPI fall short of that claim when tested. Legacy systems may use interpolation (basically, making up pixels), planetary scanners slightly blur the image because they capture while moving, commodity cameras use lenses that fall off in sharpness at the corners and often use blurring filters over their sensors. Often such systems then mask this softness by sharpening the image until it appears visual sharp again. This is why FADGI, Metamorfoze, and ISO guidelines don’t just specify PPI; they specify PPI and sampling efficiency, and limit the amount of sharpening that can be applied to achieve it, and then importantly define a set of targets and methods to check whether a system complies.
Moreover there are many different specifications that are important when judging image quality; PPI (aka resolution) is only one of them. The FADGI guidelines for example call for an image to be evaluated on sampling frequency, tone response, white balance error, illumination non-uniformity, color accuracy, color channel mis-registration, MTF/SFR, noise, and lack of abusive sharpening. Our systems have been designed from the ground up to provide top quality
For more technical details on this please see Overview of FADGI & METAMORFOZE in our Digitization Program Planning Guide. But for the purposes of this article just understand:
1) Total image quality depends on a lot more than PPI.
2) At the same stated PPI our systems will likely outperform whatever you’re comparing them to.
Yeah, Yeah, but what PPI is your system?
Some of our systems can capture any object at up to around 10,000 ppi (even higher ppi is possible with specialized accessories; contact us if interested). Let’s just stop here and take a moment to appreciate just how high of resolution that is. At 10,000ppi each pixel represents only 0.0001 inches, or around 3 microns – that means a single human hair would be around 30 pixels across. The typical driver of a Corvette Z06 only occasionally drives at 200mph; likewise, most users of our systems run them at 300ppi to 600ppi for reflective scanning and 1000ppi to 5000ppi for film scanning. But, like with the Corvette, when you occasionally need to kick it up to high gear, our system is glad to oblige.
Factors that determine PPI
There are four main factors that determine the PPI our systems can achieve.
Camera: We use a few different cameras on our systems, and not all of them have the same maximum PPI. For example the DT RCam has no absolute maximum, and is very comfortable at 10,000ppi, while the Phase One iXG is limited to around 6000ppi. Note that the maximum PPI of a camera is not a stand in for “how good is that camera”; the absolute maximum only matters if you’ll need resolutions that high.
Sensor: In addition to having a couple different camera bodies, we also use a variety of sensors. A given sensor may have as few as 40 megapixels or as high as 100 megapixels. All else being equal a sensor with more megapixels yields a higher PPI.
Lens: Lenses are designed for a specific range of focusing. A lens designed for capturing large objects just does not perform as well when capturing small objects. For this reason our systems have a standard lens (used for anything roughly A4 and larger) and a macro lens (used for anything smaller than roughly A4). This means that the maximum PPI of our normal lens is limited to around 1000ppi (depending on the sensor and other factors); on some of our systems you can use it at higher PPIs than that, and the results are surprisingly good, but for best results the macro lens should be used.
Stitching: This is the largest factor. For any given camera in our lineup there is a maximum size it can capture at a given PPI with a single capture; to capture objects larger than that multiple images are taken of the different parts of the object and then stitched together. It sounds more difficult than it is (at least, with the right hardware and software) and is a technique used by many of our clients for oversized materials. We’ve even developed a tool called DT BatchStitch for doing those stitches in large batches so the post processing is as efficient as possible. Still, a good rule of thumb for stitching is to avoid needing to stitch wherever otherwise possible.
So the question “what PPI is your system?” requires a bit more context.
Okay, got it! So with Camera X, Lens Y, and no stitching, what PPI is your system?
Now we’re talking. Answering that question is straight forward. If math scares you, or you haven’t had your coffee yet, you can skip to the example resolution charts below and find the answer to many common combinations, or contact us and we’ll step you through it personally. But the math is really easy. Your goal is to determine the PPI – the pixels per inch which can also be stated as “how many pixels does the camera provide per inch of the object?” So you need to know how many pixels the camera provides and the number of inches of the object. Then you divide. Yes, it’s really that simple. Let’s take a look at an example.
Say you have an Phase One iXG 100mp you have (very roughly*) 12,000 pixels on the long side and want to capture an object that is 12 inches on the long side. That means you have 12,000 pixels per 12 inches of subject. Dividing that out (12,000 pixels / 12 inches) yields 1,000 pixels per inch, or 1,000ppi.
Using that same camera, let’s say you wanted to capture a long object like a scroll that measures 30 inches by 9 inches, and you wanted the capture to be at 1200ppi. You would need to capture three images; one each for the left, center, and right section of the scroll. Each frame would capture 12 inches of the subject, which means you’d have 3 inches of overlap between each section (stitching works best if you have overlap so the software knows, without assistance, exactly where the images fit together).
*The exact number is 11,608 pixels on the long side. Also, normally you’d want to anticipate a slight border around the object. We just wanted to keep the math super simple for our example.
Below is a table with the pixel dimensions of each of our cameras to assist you with the math described above.
|Sensor Type||Examples||Short Dimension||Long Dimension|
|150mp Systems||iXH 150mp
RCam with IQ4 150mp
XF with IQ4 150mp
|14204 pixels||10652 pixels|
|100mp Systems||iXG 100mp
RCam with IQ3 100mp
RCam with IQ1 100mp
XF with IQ3 100mp
XF with IQ1 100mp
|8708 pixels||11608 pixels|
|80mp Systems||RCam with IQ3 80mp
RCam with IQ2 80mp
RCam with IQ1 80mp
XF with IQ3 80mp
XF with IQ2 80mp
XF with IQ1 80mp
|7760 pixels||10328 pixels|
|60mp Systems||RCam with IQ3 60mp
RCam with IQ2 60mp
RCam with IQ1 60mp
RCam with P65+
XF with IQ3 60mp
XF with IQ2 60mp
XF with IQ1 60mp
|6732 pixels||8984 pixels|
|50mp Systems||iXG 50mp
RCam with IQ3 50mp
RCam with IQ2 50mp
RCam with IQ1 50mp
XF with IQ3 50mp
XF with IQ2 50mp
XF with IQ1 50mp
|6208 pixels||8280 pixels|
|40mp Systems||RCam with IQ1 40mp
RCam with P40+
XF with IQ1 40mp
|5484 pixels||7320 pixels|
Example Resolution Charts with Single-Shot Captures
The below chart provides the largest object size that can be captured with a single capture from one of our systems. It provides the same answer as the basic math above, but saves you the calculations.
|300ppi||47″ x 36″||39″ x 29″||34″ x 26″||30″ x 22″||28″ x 21″||24″ x 18″|
|400ppi||36″ x 27″||29″ x 22″||26″ x 19″||22″ x 17″||21″ x 16″||18″ x 14″|
|600ppi||24″ x 18″||19″ x 15″||17″ x 13″||15″ x 11″||14″ x 10″||12″ x 9″|
|1200ppi||11.8″ x 8.9″||9.7″ x 7.3″||8.6″ x 6.5″||7.5″ x 5.6″||6.9″ x 5.2″||6.1″ x 4.6″|
|2000ppi||7.1″ x 5.3″||5.8″ x 4.4″||5.2″ x 3.9″||4.5″ x 3.4″||4.1″ x 3.1″||3.7″ x 2.7″|
|3000ppi||4.7″ x 3.6″||3.9″ x 2.9″||3.4″ x 2.6″||3.0″ x 2.2″||2.8″ x 2.1″||2.4″ x 1.8″|
|4000ppi||3.6″ x 2.7″||2.9″ x 2.2″||2.6″ x 1.9″||2.2″ x 1.7″||2.1″ x 1.6″||1.8″ x 1.4″|
|5000ppi||2.8″ x 2.1″||2.3″ x 1.7″||2.1 x 1.5″||1.8″ x 1.3″||1.7″ x 1.2″||1.5″ x 1.1″|
Example Resolution Charts with Stitching
|300ppi [3 x 1]||124″ x 48″||101″ x 39″||90″ x 26″||78″ x 22″||72″ x 21″||63″ x 18″|
|300ppi [2 x 2]||86″ x 64″||70″ x 52″||62″ x 46″||54″ x 40″||50″ x 37″||44″ x 33″|
|300ppi [3 x 3]||124″ x 92″||101″ x 75″||90″ x 67″||78″ x 58″||72″ x 54″||63″ x 48″|
|300ppi [5 x 5]||200″ x 149″||163″ x 122″||145″ x 108″||126″ x 94″||116″ x 87″||102″ x 77″|
|600ppi [3 x 1]||61″ x 18″||50″ x 15″||45″ x 13″||39″ x 11″||36″ x 10″||32″ x 9″|
|600ppi [2 x 2]||43″ x 32″||35″ x 26″||31″ x 23″||27″ x 20″||25″ x 19″||22″ x 16″|
|600ppi [3 x 3]||61″ x 47″||50″ x 38″||45″ x 34″||39″ x 29″||36″ x 27″||32″ x 24″|
|3000ppi [3 x 1]||12.4″ x 3.6″||10.1″ x 2.9″||9.0″ x 2.6″||7.8″ x 2.2″||7.2″ x 2.1″||6.3″ x 1.8″|
|3000ppi [2 x 2]||8.6″ x 6.4″||7.0″ x 5.2″||6.2″ x 4.6″||5.4″ x 4.0″||5.0″ x 3.7″||4.4″ x 3.3″|
|3000ppi [3 x 3]||12.4″ x 9.2″||10.1″ x 7.5″||9.0″ x 6.7″||7.8″ x 5.8″||7.2″ x 5.4″||6.3″ x 4.8|
Don’t worry! We’re very glad to help. If you contact us we’ll be glad to listen to your needs, get to know your institution, and help you formulate both your PPI requirements and guide you on what equipment will fulfill those needs.