24 April 2000
(Updated 4 September 2002)

This issue presents a visual comparison between the various print outputs produced by Canon, Epson, HP, Kodak, and Olympus photo-quality inkjet and dye-sublimation printers.

The prints produced by most online photo printers, the latest Canon, Epson and HP 6-color printer, dye-sublimation and Fuji Pictrography printers will appear as good as regular photo prints to most home users.

However, the very pickest photographers and users will be able to see the differences between those produced by photo inkjet printers vs. higher quality prints such as the Fuji Pictrography printers. These differences include points such as lower resolution, sharpness of fine details, minor pixelization, dotty shadow areas and color gradiants, etc.

This article focuses upon these differences in particular to help those who desire to have the utmost highest quality photo prints to make a sound decision. 3 and 4-color inkjet printers are included for comparison.

For inkjet printers, the key features which lead to superior photo prints are: the number of colors used, the size of the ink drops, and resolution. The first two characteristics are the most important, with the number of colors being the most important. Simply increasing the number of colors from 4 to 6 colors yields the greatest improvement in photo print quality of these factors. Second, drop size is most important. The ability of printers to produce fine details is dependent on how small the drop sizes can be made. Finally, while resolution is important, it is the factor that contributes least to the quality of photo inkjet prints vs. the other two.

What this means is that, in general, a printer with lesser qualifications in these three categories can be expected to produce poorer quality prints vs. those with better qualifications.

Keep in mind that while the output of the best 6-color photo printers are practically the same, different features such as panoramic paper roll holders, optional scanner heads, replaceable nozzle heads, and a straighter paper path vs. 180 degree U-feeds which clog and jam easier on heavier/thicker paper stock, print longevity, and the availability of a variety of surfaces and sizes for paper stock, are all additional considerations to take into account besides price.

Most of these printers have the identical parts used in other models, usually wide-body version of the regular width printers, and all printers in the same line have the same print quality.

The scans were made on an Epson 1200S SCSI scanner at the 2400dpi setting. But as you well know, the Epson 1200S only scans slightly better than the 600dpi CCD sensors it is actually based upon due to the mechanical limitations of the arrangement of two such CCD sensors offset by 1/2 pixel to 'fake' 1200dpi. The Epson 1200S is nowhere near the resolution capabilities of a true 1200dpi CCD sensor array scanner, such as the Umax PowerLook III.
(The Epson 1200U/S scanners also have big flaws - they cannot scan the 3mm edge around the inner bevel on the glassbed, they has holes in the bottom you must tape shut to prevent dust from entering, there are no histogram, etc.)

Unsharp was selected in the Epson Twain driver.

Images were assembled and then shrunk 50% vs. the original to allow viewers quick load times. No additional adjustments were done to the scans. Descriptive text was added in black.

All original printed images were made on the respective printer models listed, in their best photo paper setting, on the best photo paper available at that time for that model. Thus, the Epson 780/870 images were printed on the Premium Photo Paper, the others were printed on Epson Photo Paper, and the HP images were printed on the HP Premium Plus Glossy Photo Paper. The dye-sublimation printers used the best (usually only) paper available for that model from the manufacturer.

(Note: The current exception is the Canon 8200 examples -- 3rd party inks were used, and the Canon's best photo paper was not used by the submitter -- only Canon's older film and photo papers were used with the 3rd party inks. Therefore, the current scans of the Canon 8200 prints are not representative of the maximum quality that can be achieved with the Canon 8200 printer.)

The metal ruler is a scan of a Pickett No.33E Japanese made steel pocket ruler, with 1/100th inch rulings. The clear plastic ruler is a scan of a regular classroom ruler with 1 mm rulings.

Longevity of prints has been a long-standing issue and concern with print makers. This problem is under continous research and development, so newer, longer-lasting products are constantly being released. In general, except for the resin-encapsulated ink printers like the Epson 2000P, 2200P and C80, all inkjet printer prints in general will fade in a few months/years in display. Alternatives to lengthening lifespan include storage in a folder, under glass and frame, the use of 3rd party archival inks and papers, etc.

Both Inkjetmall.com's Knowledge Base section and Wilhelm Research's web sites have additional information regarding the longevity of prints and tested lifespans. These are constantly changing due to newly discovered factors that cause fading, such as ozone exposure.

Recently, some prints fade rather quickly, often within days or weeks. This rapid orange-fading problem has been noted to occur for a small subset of users of photo inkjet printers. While most users will never encounter this problem, it can occur unexpectedly due to print exposure to various elements and conditions yet to be fully characterized.

Please see POV Image's web site for full background information on the orange-fading problem, as well as a complete history of this problem, along with links to other sites that have tested the inks on other papers and found the same orange-fading problem. This orange-fading effect has been duplicated under certain test conditions on most 6-color inkjet printer prints. Please see the sites linked from P.O.V. above for the most recent findings.

3rd party, archival papers and inks are available for select printers. Proper use of these products can produce prints that will last decades longer than most inkjet photo printer prints available today. Please see Inkjet Mall's web site for product details.

Most buyers should use an online, lowest-price search engine such as shopper.com or pricegrabber.com. (You can find more by searching Yahoo!.) Online prices are often far lower than that found in any retail store, and the current US policy of not collecting any Internet sales tax benefits you greatly vs. the cheaper cost of having the product shipped to you (often waived by online merchants as well).

The savy use of rebates, coupons, and deals found at sites such as Dealcatcher.com, Fatwallet.com, Fatwallet.com Forums - Hot Deals and Free Stuff, and DealNews.com can reduce your prices even further.

Most inkjet printers typically print about ~50 8x10" photo prints before running out of ink. General per 4x6" print costs for inkjet printers are between $0.25 and $0.50 USD. General online photo printer costs run about $0.50 per 4x6" print. General dye-sublimation and Pictography prints run above $0.50 per print.

The purchase or entry cost is simply figured from the purchase price of the 6-color+ photo printers or other photo printers you are considering for purchase. If one 6-color inkjet printer produces about the same quality photo prints as a more expensive model, then the cheaper model is the better buy based on purchase price alone.

The anticipated operating costs based on print volume must be added to the purchase price of the printer to compare cost-effectiveness of various printers.

For example, today, the Epson 820 sells for as low as around $90. The Canon S900 sells for $250. They both produce about the same print quality, and based on purchase price alone, the Epson is $160 cheaper.

If Epson and Canon prints (figuring cost of paper and ink per print) were both $0.50 per page, then the Epson would be cheaper to run in the long-term vs. the Canon. (ie. at any number of prints made, the Epson will still be $160 cheaper) On the other hand, if the Epson prints were $1 vs. Canon $0.50 (or vice versa with the Epson at $0.50 vs. Canon $1.00) per print, then you can calculate when one printer will be more expensive to operate vs. the other.

As an example, if we take the first case, $90 Epson + x * $1 per print = $250 Canon + x * $0.50. Solving for x, the breakeven point, yields 320. In other words, once you make more than 320 prints on the Epson printer, even though it was cheaper to buy the printer, it will cost more to own and operate vs. the Canon. Naturally, if you do not anticipate printing more than 320 prints over the next several years, the cheaper printer would still be the better purchase since it is likely the printer will have broken down by then and will require replacement (ie. cheaper to buy and replace the cheaper printer vs the more expensive one if you don't use them that often).

Naturally, you must use the most recent cartridge and paper prices online along with the yield per cartridge in order to figure out correctly whether one printer will be more cost-effective to own vs. another.

Keep in mind that another factor which results in high operating costs are printers that use all-in-one cartridge designs. These cartridges which contain more than one ink color often waste ink and are more costly to use because when one of the colors runs out, there is no other choice than to throw it away and replace it with a new cartridge.

The Piezeography Black and White inkjet print system, available at Inkjet Mall, has the ability to produce B&W prints which match and can even exceed the quality of traditional B&W photo prints with very high output resolution.

Also available through Inkjet Mall is the CIS continous ink bottle feed system, which allows high-volume users to print hundreds and even thousands of prints before needing to stop to refill inks.

Furthermore, some cartridges like the Intelliedge chipped cartridges in Epson printers are more troublesome and difficult to refill with cheaper 3rd party inks, often sold in large quantities by the bottle. In my opinion, the manufacturer's decision to prevent the use of 3rd party inks (archival, cheaper, neon, etc.) by attaching useless protection chips to cartridges which provide no significant advantage in measuring ink levels vs. prior and alternative methods for the sole reason of increasing their profits is a very bad decision for consumers. It prevents the users from using alternative inks, forces the consumer to waste money on original ink cartridges that often sell for a 20-50% profit, and prevents the use of continuous ink bottle feed systems for high-volume use. The Epson Intelliedge chipped-cartridges are among the worst photo printer ideas ever.

Because time is money for many commerical users, print speeds can make or break a job.

In general, the S900/S9000 printers are the world's fastest photo printers based on numerous US and Japanese review tests. Expect about 3 1/2 minutes per letter sized print. Other printers from Epson and HP print slower, up to 5x slower, depending upon the model and image used. Refer to the various benchmark results available elsewhere for exact numbers.

Some printer brands offer a wider range of photo papers. Epson currently offers the widest range of photo papers than the other brands, including panoramic papers sized from 4" rolls to 8.3x23" sheets to 13" rolls which allow you to print 13" x 32 feet long panoramics on their widebody printers! (For prints longer than the standard 44" maximum length, a Windows driver-limit, use Linux & GimpPrint.) Other stock include matte, sticker, transparencies, cheaper general use inkjet, direct view and backlit films, post and greeting cards, adhesive sheets, and iron-ons. Larger sizes up to 13" x 19" pre-cut photo papers are available. All other brands offer a more limited selection of photo papers at this time.

Currently, the best deal on photo paper is the Epson Photo Paper sold at Costco stores for $19.99 / 100 sheets. (OEM paper for OEM printers. 3rd party papers for OEM printers sell cheaper with varying results.)

The only printers that can produce superior output and are generally available and affordable for consumer purchase are dye-sub printers (such as the Olympus P400 for <$1000 USD and Kodak for >$2000), the Fuji Pictography printers ($1500+), and commerical online photo printers such as Ofoto.com ($0.25+ per print).

In general, the ability to produce smooth gradiants between colors will result in the best prints - dye-sublimation and Pictography printers inherently produces prints without any visible dots due to their continous tone output. Smaller inkjet dots and more colors are more important than resolution for inkjet prints in producing the best photographic output. The number of colors used is crucial to the quality produced by photo inkjet printers. This is the reason 6-color+ photo inkjet printer can produce far better prints vs. 4-color inkjets.

As a guage of quality, an actual photograph is 100%, the online photo printers and Fuji Pictography printers are at 95-100%, the dye-subs are at 95-97%, the photo inkjets at 90-95%. These will all look like regular photographs in viewing tests given to the majority of viewers who are not told they are not regular photos.

I have noticed that due to the use of a rectangular dithering matrix, dots are laid down in a more grid-like manner on the Canon S800 and 8200. This results in two potential defects under magnification - a grid-like dithering pattern that is slightly more noticable than an error diffusion dithering pattern (and non-complete coverage of 100% black areas only on the 8200/8500. You can have dots with white, uncovered areas at the corners of the grid these dots are in.) Actually, some research has been done and reveals that a 2x higher resolution in one dimension increases the ability of round dots at filling square pixels on paper; see my Dotty Spotty Patents issue for the patent reference.

On the Epson, the problem of microbanding occurs, but rarely, where occasional horizontal 'lines' or stripes appear in color graidents. This can be attributed in part to the mechanical stepping not being perfectly exact on each line, as well as the contributing factors from the dithering algorithms used.

On the HP RET 4 printers, the microbanding/linear banding effect occurs vertically, much similar to the Epson. It is more noticable when using the 4800x1200dpi mode vs. the default RET 4 mode.

In a direct comparison of identical photos printed, the photo print quality produced by the 6-color Epsons and HPs are among the best samples seen thus far vs. the other inkjet printers tested. The differences between these two are slight and only visible at 6" or closer viewing distances under controlled, bright lighting during extended viewing sessions. In other words, most home consumers will never notice any differences and only the pickiest people wanting the very best photo prints available from inkjet photo printers will.

Epson printers tend to produce prints with almost no noticable linear banding but their dithering patterns tend to introduce a very slightly grainier patterning of light color tones and shading. Examples of this in the Photodisc test target print are egg shells and faces that feel a touch rougher vs. HP 6-color photo prints. On the other hand, HP 6-color photo prints possess vertical banding in photo prints, but at a scale less than that which occurs in Canon printers. This artifact in HP prints are very slight, and again, only noticable at 6" or closer under extended view conditions. Examples of this in the Photodisc print are patches of color chips that possess vertical banding, and darker areas of shadow. Whether you prefer one or the other print is almost completely subjective at this level of difference - the artifacts occur at a level which most people will never detect, and will appear almost inconsequential, initially, to those that do. Naturally, the pickest user will need to carefully compare to see which one is prefered, but you could literally swap prints and not know which brand produced what at first glance.

Keep in mind that at this level of artifacting, head misalignments, nozzle clogs, and using the wrong paper thickness settings can magnify these artifacts or even cause them to appear.

Additional scans of these test target prints will be added as time allows so you can better see these minor differences.

However, the initial impression one gets when looking at the Epson and HP prints vs. the other inkjets tested here is significant, with the former appearing photo-like and the latter appearing far grainier. A simple side-by-side comparison between prints from the older HP RET 3 printers vs. the latest 6-color RET 4 printers reveal a significant improvement in print quality in the 6-color prints, and so much so that these older printers are not even worth consideration vs. the latest 6-color photo printers.

Keep in mind that as good as these 6-color photo printers are, the 7-color, 2 picolitre (smallest dots of any inkjet printer made today) Epson PM-950C has been reviewed by many Japanese computer magazines and picked as the best photo inkjet printer available today. People who live in Japan or can export both the printer and ink cartridges who desire the best photo inkjet print quality available should look at this printer. I'll hold my judgement on this printer until print samples can be obtained, but based on specifications alone, one can assume that this printer has a good shot of actually being so.

Most other inkjet printers made today will only produce worse looking prints.

Nevertheless, no photo inkjet printer made today will look as good as a photo print made from the Fuji Pictrography printers. The use of the Pictrography print process results in absolutely no dots at all at any distance, and the appearance in a direct side-by-side comparison at 6" and closer is one of superior, smoother, dot-free prints.

This difference is immediately noticable. The pickiest users will not be completely satisfied with the output from even the best photo inkjet printers, although they may concede that most viewers looking at prints from greater than one foot away will never see anything less than a perfectly fine photo print. Even when they do examine prints at 6" and closer, most viewers will simply not even care about the presence of the finest grain pattern on the best photo prints produced by Epson and HP's 6-color printers. You can easily put the Fuji Pictrography prints as being at least 25% better than those produced by inkjets, more so if you are a pickier person, less so if you are not.

Until I acquire photo print samples from the Canon S900/S9000 series, I can't say if it performs as good as the 6-color Epson and HP printers. I will update this article as soon as I do. However, based on current user reports, one can assume that the Canon S900/S9000 will not be that far in quality, if at all.

My scans my not reflect exactly what you'd see under a higher power microscope -- ie. you cannot see each and every inkjet dot clearly and distinctly.

Images of dots taken with high-magnification microscopes have been done before:
Extreme Magnification
Extreme Magnification 2

and my own wall chart of inkjet dots from various inkjet printers:
Dotty Spotty: Issue 3
I only had use of a poor 30x pocket microscope for my wall chart of ~50 magnified dots samples.

However, the scans in this article do reflect what an average viewer would see upon closer examination of prints, as well as what they may look like farther away if he sits farther from the monitor. Kinda pointless to be looking at 60x microscope magnifications of the dots themselves when it doesn't help most average people understand and see what the actual visual differences are with the unaided eye, does it?