Explanations about Image Resolution (DPI or PPI)

Thought of penning this down as a reference for anyone who is confused about image resolution (DPI or PPI), since I think there is a lot of misunderstanding regarding this topic, especially regarding the image resolution setting found in most image processing software when exporting.

To first lay down the basic terms for this discussion

• Image resolution refers to how many dots-per-inch (DPI) or pixels-per-inch (PPI) - effectively meaning how many dots or pixels are represented in 1-inch of display media.

• Image size refers to the dimensions of an image - for example the number of pixels of the length and height of an image.

The term PPI is a newer term, that became common from screen manufacturers’ usage to showcase how sharp (i.e. how fine the pixels were) a screen is. DPI on the other hand was a term used for a long time to discuss the image resolution of an image (either digital or printed) as well as for printer’s resolving power. However, most people or software do use these 2 terms interchangeably when talking about image resolution, so to simplify things we will just use DPI here.

Now a very common statement is: “The higher the DPI the sharper the image.”

There is no argument - this statement is true, but not for the saved “image resolution” setting within the image file, but rather the resultant DPI from printing the image.

What do I mean by that?

The DPI value itself is a ratio of the one dimension of the image size in pixels to its size in inches, and can be used for converting or calculating one or the other.

To put it mathematically,

• Image Size (pixels) / Image Size (inches) = DPI

Any two of these values are needed to calculate the third value, and specifying or locking two values would vary the third.

Here are some examples and comparisons:

1. I print an image with dimensions 3000 x 3000 pixels at 300 DPI, so the resultant print is 10 x 10 inches.
   The same image is printed at 100 DPI, and the resultant print is 30 x 30 inches.
   We know that a higher DPI is sharper, and you can imagine how a same origin image always look more pixelated when printed larger.
   In this case we have Image Size (pixels) and DPI, which controls the final print output.

2. I want to print 10 x 10 inches, and to maintain sharpness I set my image resolution to 300 DPI.
   This means that my images would be saved or exported to 3000 x 3000 pixels.
   A side note from this is that my original image size should be at least 3000 x 3000 pixels, so that it can be resized down to 3000 x 3000 pixels to maintain 300 DPI. Any smaller image (say 2000 x 2000 pixels) would be upscaled to 3000 x 3000 pixels, and thus losing some sharpness and clarity.

In these two examples, we have only the dimensions of the image size in one unit, and use DPI to convert the other unit for input or output.

Now what if we set BOTH the input size and output size?

• I have an image size at 3000 x 3000 pixels, and I know I am going to print it on a 10 x 10 inch paper.
  I have now the following options:

  • I can choose not to set the print resolution, and the resultant print would be in 300 DPI.

  • I can set print resolution at 300 DPI, and the resultant print would be the same as above.

  • I can set print resolution at 600 DPI, at which the software would upscale the image to 6000 x 6000 pixels to maintain 600 DPI for a 10 x 10 inch print, and then finally print the upscaled image at 10 x 10 inch. This image would look very similar, if not identical, to the above two options.

  • I can set the print resolution at 100 DPI, and the software would resize the image to 1000 x 1000 pixels before printing it out at 10 x 10 inches. This would lead to a softer print as compared to the previous examples.

With these explanations in place, now we can continue with the following deductions:

• By setting both the print size and the original image size, the resultant print DPI can just be determined by the ratio relationship. This resultant print DPI is what determines how “sharp” a printed image looks - the higher the better.

• If we set the print DPI to any number that does not correlate perfectly to the image size or print size, a resize (downscale or upscale) would happen and there might be loss of quality involved. While this may be useful when sending it out to a third party printing service, if printing directly from source image it is always better to leave the print DPI field empty if your resultant print DPI from the original size is above your requirements.

And with this comes the next question: “Does the image resolution field of an image mean anything?”

As mentioned, a digital image will always have its image size in pixels, and the stored DPI value is there to convert it to print size if the print size is not specified. Similarly, if you specify a print size, the stored DPI value is there to determine the image size to save as, so that when you print the image as it is it would be at the print size you previously specified.

However, just changing the “image resolution” field within an image file does not change the image at all - as mentioned it is a ratio used for conversion between image size and print size, so unless you specify a print size as well, regardless of what number you put in there the original image would not be affected at all.

You can test this out by opening an image, and changing just the “image resolution” field and saving it - the image size (and file size for that matter) should remain exactly the same.

Printing to a specific print size from an image, the “image resolution” field does not mean anything at all, and what matters is if you print it at its native resolution or you set a print DPI on the printer output.

So to summarize everything,

• The “image resolution” field within an image file is there to help convert from pixel size to print size when print size is not specified.

• Just changing the “image resolution” field within an image does not alter the image in any way, unless the print size is also specified (in this case original image would be resized accordingly).

• By specifying a print size from an image’s native pixel dimensions, the resultant resolution of the print can be used to determine how sharp a print is - this is in fact what the first statement, “The higher the DPI the sharper the image.”, refers to.

New Flexible Color Picture Control

I mentioned in my previous entry that I was quite intrigued about the new Flexible Color Picture Control when the Z6III was first announced, and during the touch-and-try event I managed to get some hands-on experience using the picture controls and spoke to some of the staff regarding its development and implementation.

As a brief introduction, picture control profiles are selectable profiles within the camera when shooting (both stills and video), which comes with a set of options that the user can fine tune to their liking for a certain look, and these settings are applied to live view previews (when Apply Settings to Live View/Show Effects of Settings are turned on), applied directly to saved JPG/HEIF files, and also used for the saved JPG previews within the RAW file.

In recent years Nikon has introduced more picture control profiles, such as creative picture controls that gives a more distinct look to the images via color casts and more manipulated tone curves, or multiple monochrome picture controls for different black and white looks - very similar to what other manufacturers do for film simulations or the likes.

And now what they are doing with the new Flexible Color Picture Control is to bring the level of color customization a step further to allow users to fine tune color casts, color blending, color grading on top of tone curve (contrast) and sharpness settings.

So before I continue, no, this is not a unique breakthrough or a game-changer of any sorts (yeah we all dislike this term don’t we), since other manufacturers usually have something similar in their cameras too.

But its implementation is something that I think is intriguing and am keenly interested to see how it will unfold.

As of writing, although the Z6III is the only camera that allows you to select Flexible Color in its picture control settings, I foresee it would come to other Expeed 7 bodies via a firmware update in the near future, like how the extra monochrome or rich tone portrait profiles were added via firmware updates to the Z8.

Even if the body does not support Flexible Color picture controls now, by using the official Nikon RAW development software, NX Studio (version 1.7.0 and above), you can now apply the new picture control to any Nikon RAW files (make sure you choose “Latest Picture Control” in the dropdown menu), and use it to color grade your photos within NX Studio.

Now comes the interesting part - your “recipes” can be then saved and exported, and with the upcoming Nikon Imaging Cloud service, users can share their flexible color picture profiles with others, and also download those created by other independent users or those done in collaboration with Nikon.

More information regarding how to use NX Studio to apply, export or import the new Flexible Color picture control can be viewed on Nikon’s official guide here.

Truth be told, I know NX studio is not the first choice of a RAW developer for a majority of photographers due to speed issues and lack of shortcut key customizations, and I am not in any way trying to convince any one here to move to NX Studio for full RAW development. But it can now be used as a powerful tool to create, tweak or import custom color grading profiles for your camera, and when the in-camera support comes for the other Expeed 7-based bodies, it would be useful in many workflows, especially those of dedicated JPG shooters. On top of that, using the flexible color profiles of other users also gives a chance to experiment with trying out different looks, and also learning how such looks are achieved by looking at how the color settings and tone curves are tweaked.

Hardware Monitor Calibration and ICC Profiles

Just chanced upon this video while looking for some other information regarding ICC profiles earlier on, and thought that it might be an interesting share to post here.

Q&A on Switching ICC Profiles on Hardware Calibrated Displays (opens external YouTube link)

This video talks specifically about why ICC profiles are still required even though a display is hardware calibrated and have LUTs stored within the monitor, and under what circumstances should users change the ICC profile.

While the video uses a hardware calibration capable BenQ display, the concepts and insights shared here are applicable to most, if not all, hardware calibration capable displays (like the EIZOs with Color Navigator).

Please take note that the principles and concepts shared are not applicable to software based ICC profiling with LUTs by using a hardware display calibration device, but rather for a display that is capable of saving LUTs within the display itself.

To briefly summarize the main points in the video,

• A hardware calibrated monitor should have an accompanying ICC profile, which usually contains only the gamut information of the calibrated display.

• Changing just the brightness of the display while all other gamut/color settings are kept identical do not require a change in ICC profile.

• Changing to a calibration slot with a different gamut (like from sRGB to AdobeRGB), the corresponding ICC profile has to be selected and loaded into the OS.

For EIZO users, if Color Navigator is active when you change the calibration slots, the proper ICC profile would be loaded into the OS automatically.

So to extrapolate the information from this video a little,

• When using a hardware calibrated monitor (not same model, but the exact same monitor) on multiple PCs, it is advised to load the ICC profile on all the PCs so that the OS and software know how to map the colors to the display’s gamut.

• It is simpler to just calibrate to the widest gamut the display can show (native in most cases), and let the ICC-aware software you are using handle the conversion of colors when the proper corresponding ICC profile for the native gamut is loaded.

I think Art, the author behind this video, did a wonderful job explaining and demonstrating, and this video is definitely worth your time taking a look at if you are running a color managed workflow.

Thoughts on the Synology BeeStation

Earlier this year Synology announced the BeeStation, a simple Network Attached Storage (NAS) solution to create personal cloud servers.

(For more detailed information, the official press release of the product can be accessed here, with its corresponding product page here.)

While using NAS to create a personal cloud and backup solution is definitely not a new idea, BeeStation seems to simplify the process as much as it could to appeal to a wider audience.

With your typical NAS systems, it does require one to have certain technical knowledge (like port forwarding etc), and most NAS devices come without drives so hardware installation skills are required too. In this particular aspect, I personally do think that the BeeStation is going in the right direction by simplifying the whole device set-up process so that users can just get it up and running with minimal fuss.

As of writing there seems to be only a single model available - a single bay BeeStation with a 4TB HDD pre-installed.

From what I have read online, the set-up process is pretty straightforward - after plugging it into a router, create a Synology account and activate the device via the BeeStation portal site (I wonder if it is called BeePortal?), install and sign-in to the apps on the devices (mobile/desktop) you want to have access to the BeeStation, and that’s it.

No fighting with port-forwarding rules, no fighting with expired SSL certs.

More advanced users who require specific features or higher levels of customization may find this a bit too simplistic and lack many options that the usual DSM offers, but for the target audience of this product I think that its simplicity is definitely very appealing.

I personally own a couple of Synology NAS devices and while I have no issues about them with respect to my use case, I do find it hard to recommend to friends and family who are not very comfortable with tech as initial set-up and the occasional troubleshooting do require some knowledge of how things work.

The BeeStation does look like something that I could recommend to them for a simple personal cloud solution, though I do have certain reservations about it right now:

• Which exact 4TB HDD is pre-installed?

• Can the HDD be changed to a higher capacity or a different brand by the end-user?

• Will there be RAID-1 (or SHR-1) 2-bay models down the road?

That said, I am definitely interested to see how this product line will develop over time and am keeping my eyes on it. I can see myself adding it to run alongside my current NAS’es as a family shared storage of sorts, though I would like to see how the product line evolves first.

Tresorit Send - Secure File Sharing Service

Before I get started, let me just say this upfront: no this is not a sponsored article, nor are there any paid ads or affiliate links in here.

Some time back, I was looking for a simple, secure file sharing service to deliver files to my clients. I might have multiple file deliveries with overlapping download durations, so typical methods like using Google Drive or Dropbox to generate shared links were not feasible unless I upgraded to a paid plan with way more storage.

In Japan the de-facto file sharing service is called Gigafile, which I shall not link here as I personally am totally not fond of it - ads are everywhere on the site, and I have never managed to find any notes on how uploaded files are encrypted or secured so I just assume it is non-existent.

That was when I stumbled upon Tresorit Send. Tresorit provides secure cloud storage solutions, with extra emphasis on security, and Tresorit Send is a free standalone product provided by them that uses similar underlying technologies in their secure cloud storage solutions.

While you do not need an account with Tresorit to use Tresorit Send, a valid e-mail address is required during upload (for admin control, more on this later).

The way to use it is pretty straightforward - simply select your files through their website or browser plugin, fill in your e-mail address, add password protection if desired, and start the upload process.

(Side note: I do recommend the browser plugin - makes the entire process more seamless)

The files are encrypted locally on-the-fly as they are uploaded, and when the upload process is completed, a URL is generated for sharing. Both the path to the files and the decryption key are embedded in the generated URL, and as a user accesses the shared URL, the files are decrypted on-the-fly as they are being downloaded.

If password protection was selected during upload, the user accessing the shared URL must key in the password before being able to start the decryption process.

An admin link is also sent to your e-mail address which allows you to check how many times the file has been accessed, as well as to revoke the link any time. It is important to note that the admin link does not contain the link to the actual file in any manner (for security reasons), and as such if you lose the original generated URL your only way is to just re-upload and get a new URL.

More in depth information on the service and its technologies can be read in their FAQ here.

Being a free service it does have certain limitations and restrictions:

• Maximum total file size is 5GB per link.

• Maximum number of files is 100 per link.

• Each link can only be accessed 10 times.

• Each link is only active for 7 days.

• Admin links only show number of downloads, not detailed statistics like IP addresses.

• The service is only available in English, which may matter to some users.

While you can circumvent the maximum file number limit by zipping up your files prior to upload, removing the other restrictions and limitations (except for the English language only restriction) would require you to subscribe to their cloud storage solutions and using it to generate a secure link.

To be fair, 5GB/link is really generous for a free service, and the duration of the active link is decent too since most of my clients would download it within a few days upon receipt.

However, a certain OS (from a certain fruit) does sometimes not show when a file is downloading, especially when using an embedded browser, so there are instances whereby my clients just spammed the download button and exceeded the 10-times download limit unknowingly within seconds.

Anyway, something to take note of is that since encryption and decryption happens locally on-the-fly, for large files it is actually a pretty resource intensive process, so depending on your machine’s/device’s specs it may take much longer than anticipated to upload or download a file.

Wireless Keyboard Woes

Just thought of sharing this sort of hilarious incident I had with my recent wireless keyboard purchase.

At home when I am using my desk-bound laptop, as it is elevated on a laptop stand I had a wired keyboard attached to it - nothing fanciful, just a simple full-sized US-layout membrane keyboard.

In the sea of JIS-layout keyboards available here, it was seriously difficult to find a proper US-layout keyboard that did not cost an arm and a leg, so I was happy to be able to grab that wired keyboard at a very reasonable price.

As the little human is getting more active, when she’s sitting with me in front of the PC she tends to want to type on my keyboard too - I am sure anyone can imagine what sort of interesting text strings can arise from that.

So I thought of a simple solution - if I got a wireless keyboard, when she’s sitting with me I can just turn it off, so even if she hits on the keys nothing would happen. Perfect!

Looked through Amazon and found a keyboard that suits my requirements - US-layout, long lasting battery life, properly sized chiclets-styled keys.

Enter the HP 230 Wireless Keyboard. It is a simple wireless membrane keyboard that promises 16 months of battery life, and cost about JPY 2600. Perfect balance of everything for me.

Upon receiving my new keyboard I took it out of the box, plugged the USB receiver into my PC’s USB port, and searched for the On/Off switch on the keyboard.

Wait a minute..

There is no On/Off switch on the keyboard.

Anyway I just booted up my PC to see if it works, and the USB receiver and keyboard were pre-paired and could recognize each other instantly - and a tap of any key on the keyboard would bring it out of sleep mode as long as the USB receiver is active.

I do not have any issues with the keyboard itself - it works as advertised, being very comfortable to type on for a membrane keyboard, and is responsive to wake up from sleep and hibernates automatically when not in use.

The issue is with myself - I bought a wireless keyboard wanting an option to quickly turn off the keys to prevent accidental typing by the little human, but this keyboard though ticks all the other requirements for me, and just misses out on this one little feature that I was truly after.

I won’t be returning this keyboard as I like it much more than the previous wired one I was using (and not having an on/off switch sounds like a really bad reason to return the product), so I guess I’ll be putting up with more random text strings by the little human for a while more.

Capture One Pro Lens Profiles

When images are loaded or imported into Capture One Pro, lens profiles are automatically applied if they are available.

While the profiles made by the Capture One team are pretty accurate and work well, one shortcoming of using their profiles (or the Generic Pincushion/Generic Distortion profiles) is that any form of distortion correction will cause some pixel loss, so even if no resizing is done and the images are exported at full size, there is a chance that you will end up with images with slightly varying dimensions (depending on lens used and amount of distortion correction applied).

I have found that one way to circumvent this, if your camera body allows for it, is to use the “Manufacturer Profile” option, so even if distortion corrections are applied, the dimensions would remain the same.

Surface Pro 4

I’ve been looking for a solution to get data backup and movie streaming for shoots out of town, as my XPS 15 isn't exactly lightweight to lug around.

And also that I'm kind of lazy to plug and unplug all my peripherals and display cables when I need it for the trip.

My requirements were pretty straightforward - at least 8GB RAM, at least 256GB SSD, at least 1080p display, and if possible, 100% sRGB coverage.

Turns out that Surface Pros do meet that specs, and refurbished Surface Pro 4s are quite a nice bargain now so I snagged one recently.

Being my first Windows tablet, it’s quite an interesting experience per se to try to balance it on my lap when using it with the keypad.

And to be honest I'm pretty impressed with the hardware. The system itself is snappy and responsive, screen is bright enough and covers 100% sRGB, the keypad is very tactile and nice to type on, the body looks and feels polished and I don’t feel insecure about the kickstand at all. The machine itself is perfect for my needs.

But for whatever reason, it feels like a disastrous marriage between good hardware and badly written code - it feels as though the software is bogging the product's greatness down.

There were a few quirks that I had to sort out ever since I got the system.

1: Windows Update kind of needs to be a little smarter in terms of installation. My machine came with an older build of Windows 10 preinstalled, and upon connecting to WiFi for the first time, it just downloaded ALL the updates available and installed them ALL at one shot. I'm pretty sure certain updates require restarts, and certain updates should be installed prior to later ones.. but it was so haphazardous in terms of the order of updates installed that I could not reboot the system halfway while installation/downloading was ongoing. And voila, a couple of cases of corrupted .dll and missing directories emerged.

2: Surface NVMe firmware drivers seem to be bottlenecking the SSD. Yes I don't expect WD Black or Samsung 980 Pro performance, but 300MB/s write speed for a NVMe SSD is just sad. I could live with this speed for my current uses for it, but knowing that the drive could perform better (apparently it’s rated at 1GB/s write speed) just feels frustrating. And nope, never managed to find a fix for this.

3: Ah.. connected standby. Why does it even exist? It felt strange that the battery drain was so bad - even when my machine was shut down about 50% was drained overnight. I went to read about it and found out that connected standby was the culprit, but latest build of Windows removed the option to disable connected standby via registry. I had to use another registry key to permanently disable Sleep mode instead - so now I only have Hibernate and Shut Down options. But the battery drain, while improved, is still there.. perhaps about 5%/day even when shut down?

I really enjoy using the machine, and am really glad that the hardware itself suits my purpose perfectly, but having to work around or find solutions for these software kinks do irritate me every now and then.

Nikon Z6 vs Z7 Colors

Most of my assignments shot under the various studios are done with the Z6 + Z6II combo, as the combination of cleaner high ISO images, higher AF sensitivity, and more manageable 24mp file sizes are the more crucial considerations.

However in my personal time (and the occasional odd assignments where high resolution files are needed) I use the Z7 quite frequently.

One thing that has always intrigued me is that the files coming out of the Z7 are richer in color and easier to post process to my taste, as compared to the Z6/Z6II, which usually have a hint of quirky magenta/red tint that I would have to remove.

Not that the Z6/Z6II are slouches; it’s just that I just have a much easier time dealing with the colors of the Z7 files.

While I know that they use totally different sensors, it’s still interesting that even within the same generation of cameras and boasting the same CPUs, the images produced are quite distinctly different.

Keysticks + Gamepad for Image Culling

A while ago I had this idea about using my gamepad for image culling.

And after experimenting with it with various software, I have to say I am really pleased that it works really well, and has sped up my work flow quite a bit!

The key-mapping software I am using is called Keysticks, a free open-source software, and am coupling it with my Xbox One wireless controller.

My current layout is like this:

• D-pad mimics the directional arrow keys

• Right joystick mimics the mouse cursor movement

• "A" button for 1-star rating

• "B" button for 0-star rating

• "X" button for 5-star rating

• "Y" button for "Hand" tool

• Right trigger for mouse double-click

• Right shoulder button for mouse single-click

• Left trigger for mouse scroll-up

• Left shoulder button for mouse scroll-down

And my workflow for culling is as such:

• After ingesting the files into Capture One Pro, first I press "Y" to make sure the hand tool is selected

• As I go through the images using the D-pad, I use the joystick to move my mouse cursor to roughly where I want to check for critical focus. A tap of the right trigger button will zoom in to 100% at the mouse location, and another tap would send it back to fit screen mode, making it easy to quickly zoom in and out to check critical focus.

• In the first round of image culling, I tag the images that have misfocused shots/blur shots/closed eyes with 1-star rating by tapping the "A" button. In the event I accidentally tag something wrongly, I cancel it using the "B" button (0-star rating).

• These 1-star rating images are moved to a separate folder, and I go through the remaining images in the original folder to select images I want to post process. This is done by using the "X" button to give them 5-star ratings, and cancelling accidental tags with the "B" button as well.

• If I need to check on some small details around the image (without the need for 100% zoom), I can do so with the left trigger button and left shoulder button (which mimics mouse scroll functions), and panning around can be done by pressing and holding the right shoulder button and moving the joystick.

• And finally by filtering to only display the 5-star rated images, I then proceed back to my keyboard and mouse for processing the images (yea I know I know..am just not a huge fan of pen tablets)

Using the gamepad to do the culling is an enjoyable relaxing experience so to speak, as I can literally lay back with a bad posture on my chair while going through the images.

I really want to give a shoutout to Keysticks for the wonderful software!

Now I wonder if a portable version that could run off a USB drive without any installation could be developed......

EIZO CS2410 Thoughts

While I have been using an software color calibrated displays for a long time now, I have been wanting to upgrade to a color accurate monitor that can be hardware calibrated, but had to put it off till we moved to somewhere more permanent. And after moving in to our new place in the late ‘19, COVID reared its ugly head and it was really difficult to justify a purchase as we tightened our belts.

But, ironic as it is, being at home pretty much the whole time due to COVID got me thinking about the purchase of a display again, as I spend more time in front of the PC reprocessing older images that I had previously archived (that's a story for another day), and some new assignments that I am receiving would receive a nice productivity boost from having more real estate for viewing.

And after a month of deliberation with myself, I went on ahead and purchased my first EIZO display - the basic CS2410 sRGB model. While not as slim or light as the wife's BenQ GW2480T (which she also recently purchased for working from home, and is absolutely awesome for its intended purpose and price-performance ratio), setting it up alone was not much of a trouble and within several minutes the display was hooked up to my XPS15 and running.

The build is solid - no rough edges, the base plate locks into place intuitively, all the tilt/height adjustment/pivot actions are smooth, and the base plate swivels around easily. And having an extra 11% of vertical real estate is simply fantastic.

Linked up my Spyder 5 calibration unit next, ran the calibration software Color Navigator 7, compared the calibrated CS2410 and my XPS15's display side by side, and it confirmed what I had been suspecting for a while - my XPS15's display (4K touch version) does have a slight reddish tint to it, even if I tried setting the white point to a higher 7000K rather than D65.

The common consensus seem to agree that the older Spyder units (or Spyder units in general?) are not good at calibrating wide gamut white-LED displays, and perhaps that is the cause of my red tint? I guess I might pick up an i1 Display Pro unit some time down the road..?

Well, at least now that my suspicions are confirmed and I know that my eyes were not playing tricks with me, I have shifted all color sensitive work to the CS2410, which is really a breeze to work with as I no longer have to fight with the red tint (cheers!) or glare from the windows (double cheers!).