Canon native print resolution

[DanK]

I've always thought that the native resolution of Canon printers is 300 dpi and have always set Adobe software to print at that resolution. It turns out that isn't always true.

A discussion of ppi and dpi in another thread led me to review the specs for my Prograf Pro 1000, Canon's competitor to the Epson P800 and P900. Much to my surprise, the online spec sheet for the print head stated 600 dpi in both directions. I called Canon this morning, and the tech was not informed about how to set this in non-Canon software. (Canon software sets it automatically, I think.) However, he said that what I read was correct: the nozzle pitch is 600 dpi. This is not the number of droplets, which is far higher.

I haven't yet had a chance to do comparison prints at 300 and 600 dpi, but I'm curious whether anyone else uses higher-end Canon printers and knows anything about this.

[LenR]

Dan, I frequently use 600 dpi when printing thru LR. See the attached article by Jeff Schewe who was instrumental in developing both LR & PS programs.

https://www.digitalphotopro.com/tech...-resolution/5/

Additional reading on Jeff Schewe.................

http://www.photoshophalloffame.com/jeff-schewe

[DanK]

Len,

thanks. I'll read that.

Dan

[Manfred M]

Dan, I frequently use 600 dpi when printing thru LR. See the attached article by Jeff Schewe who was instrumental in developing both LR & PS programs.

https://www.digitalphotopro.com/tech...-resolution/5/

Carefully said, Schewe was instrumental in the development of Lightroom, not Photoshop. While he was consulted on the software, his main contribution was the sharpening routines (his company developed them and Adobe bought them).

I read somewhere that the 720 / 600 dpi are a bit of a "cheat" as the printers use two nozzles to make a single "dot", so the best we can get out of the printers is 360 / 300 dpi. Now, I read that a number of years ago, so no doubt technology have moved onwards.

[DanK]

I'm still somewhat confused, as I routinely use the "highest quality" setting in the Windows print driver for the Canon printer, which Schewe's article suggests is 600 dpi, but I have been setting LR to 300. From his article, it sounds like switching to 600 is going to produce at best marginal gains in large prints, possibly small gains on smaller prints.

[Manfred M]

I'm still somewhat confused, as I routinely use the "highest quality" setting in the Windows print driver for the Canon printer, which Schewe's article suggests is 600 dpi, but I have been setting LR to 300. From his article, it sounds like switching to 600 is going to produce at best marginal gains in large prints, possibly small gains on smaller prints.

I believe where Schewe is coming from is that while the stated resolution of the printers is 300 or 360 dpi, the printers actually have two different dot sizes. In regular resolution mode, they print with the native 300 / 360 doi. Switch the printer to high resolution mode, the dot size is 600 / 720dpi. BUT the printer will print two 600/720 dpi dots at the same time, so effectively the resolution is still 300 / 360dpi. I have not seen any information concerning dot geometry, so know nothing about this aspect of the printer design. The other "feature" of the high quality mode, at least on Epson printers, is that the print is unidirectional whereas in the lower quality mode it is bi-directional, so the effective print speed is half as fast.

I have asked three good printmakers (i.e. they print for others and make money in the process) if they noticed any quality or ink use difference between the regular and high quality modes; there were online comments that high quality mode used more ink. All three use Epson printers, so I can't say how Canon performs. but suspect it would be similar. They all said they noticed no difference in either print quality or ink use, but this is purely empirical evidence, not lab type tests. All use different types of paper; both cotton rag and different luster papers. What I find interesting in Schewe's article he specifically mentions glossy paper. Perhaps one has to use a surface that is quite reflective to be able to make out the difference?

[DanK]

What puzzles me is that when printing with a Canon printer under Windows from Adobe products, one has two places to set this. The first is in the software itself. The second is in the Windows print dialog. The properties button in the print dialog brings up the firmware for whatever printer you are using. The top two quality options are labeled "highest" and "high".

For serious printing, I have always used "highest" (which I now think means 600 dpi) and 300 dpi in the Adobe software. It seems logical to assume that whichever is set lower is the limiting condition, in which case this pair of settings isn't sensible. when I noted that I wasn't sure where the limitation is in this chain, Andrew Rodney replied with this, which I found cryptic:

Resampling (if done) through the OS.

Perhaps I'll test it. It shouldn't be hard, if I can find a suitable image. I'd just print on a fairly smooth coated paper with three combinations: 600/highest, 300/highest, and 300/high. A glossy paper might be best for this, but as I virtually never use one, it might be more realistic to use something like Canson Baryta II, which is about as smooth as I ever go.

[Manfred M]

What puzzles me is that when printing with a Canon printer under Windows from Adobe products, one has two places to set this. The first is in the software itself. The second is in the Windows print dialog.

Interesting; with Epson there is only one place to set this. The printer dialog.

That being said, that is for printing out of Photoshop; I don't know how Lightroom handles this.

[DanK]

Interesting; with Epson there is only one place to set this. The printer dialog.

That being said, that is for printing out of Photoshop; I don't know how Lightroom handles this.

But photoshop sets the resolution automatically, based on the image size and paper size, if you click "scale to fit media". To avoid that, you would have to size the image so that it fits the media at the dpi you want, no? When you print, do you calculate the file size based on 360 dpi and the image size you want to print?

[Manfred M]

But photoshop sets the resolution automatically, based on the image size and paper size, if you click "scale to fit media". To avoid that, you would have to size the image so that it fits the media at the dpi you want, no? When you print, do you calculate the file size based on 360 dpi and the image size you want to print?

That is exactly what I do; I resize to full print dimensions at the native printer resolution.. I do my output sharpening once the image has been resized.

[DanK]

Exactly what I thought. AFAIK, that is exactly what the ppi setting in Lightoom does: it sets the up- or downscaling to get the right dpi for the media size you have selected.

So the logic seems the same. It seems to me that if you have set the image to have 300 dpi, either through the LR setting or by doing the resizing yourself in Photoshop, it can't very well improve print resolution by telling the printer via its firmware to print at twice that resolution. It seems only logical that if one wants that benefit--if there is indeed any appreciable benefit to be had--one has to avoid downgrading the image to the lower resolution at both choke points.

[xpatUSA]

I resize to full print dimensions at the native printer resolution.

I suppose that one advantage of that is that you get to choose the re-sampling algorithm as opposed to what the print driver (or whatever it is) thinks you need?

I couldn't for example imagine a 2268x1512px from my Sigma SD9 upsized Nearest Neighbor to 16x20" - pardon the exaggeration ...

... and not that I'm saying that drivers use the dreaded NN of course.

[Manfred M]

Exactly what I thought. AFAIK, that is exactly what the ppi setting in Lightoom does: it sets the up- or downscaling to get the right dpi for the media size you have selected.

So the logic seems the same. It seems to me that if you have set the image to have 300 dpi, either through the LR setting or by doing the resizing yourself in Photoshop, it can't very well improve print resolution by telling the printer via its firmware to print at twice that resolution. It seems only logical that if one wants that benefit--if there is indeed any appreciable benefit to be had--one has to avoid downgrading the image to the lower resolution at both choke points.

My workflow is such that I ensure that no resampling takes place while I am building the image, but only at the very end where I get ready to create the final output, regardless of its size and output size. That ensures that I do not introduce any unwanted effects from resampling.

Like others here, my prints are based on using a "master file". I duplicate my master file for output work, which is where any upsampling or downsampling occurs, based on final output framing and resolution. I also do my output sharpening at this stage, but I use a somewhat unusual output sharpening step that works well for me, but is not at all what is recommended by many of the "experts". This is also why I cannot use the Lightroom print module, as this does not give me the control I need.

[xpatUSA]

... I also do my output sharpening at this stage, but I use a somewhat unusual output sharpening step that works well for me, but is not at all what is recommended by many of the "experts" ...

I too have an odd one for quick-and-dirty work in my Viewer (FastStone). I do three-step sharpening first at about 2.3px next at 0.7px lastly at 0.3px, normally increasing the amount at each step. Sometimes gives surprisingly good results provided one keeps halos at bay.

I am aware that it can't be done in a "parametric" editor.

In a parametric editor, I usually combine Contrast By Detail Levels, Impulse NR and de-convolution at about 0.57px.

Barely relevant but yesterday I read a post on a Sigma forum by a guy who never sharpens at all. Any shot that's soft or has blur of any kind gets trashed! I quite admire the view. More power to his elbow ...

[DanK]

My workflow is such that I ensure that no resampling takes place while I am building the image, but only at the very end where I get ready to create the final output, regardless of its size and output size. That ensures that I do not introduce any unwanted effects from resampling.

Like others here, my prints are based on using a "master file". I duplicate my master file for output work, which is where any upsampling or downsampling occurs, based on final output framing and resolution. I also do my output sharpening at this stage, but I use a somewhat unusual output sharpening step that works well for me, but is not at all what is recommended by many of the "experts". This is also why I cannot use the Lightroom print module, as this does not give me the control I need.

Interesting, but off topic. I certainly wasn't trying to persuade you to print from Lightroom. I was simply trying to confirm that the process has two choke points, regardless of the software choice. In both cases, the first is when you set the resolution in software--by doing the resizing arithmetic yourself in photoshop, or by telling Lightroom to do it. The second is when the OS passes the image to the printer's firmware.

It seems to me that the answer has to be that leaving aside imputation methods, it shouldn't help much to set one of these parameters to double the value of the other. That is, I can't see how setting the printer software to 600/720 dpi can help much if one has told the software to produce an image that is 300/360 dpi on each dimension. And since imputation, when it's done at all, is done at the software stage, I'm guessing that setting the firmware to 720/600 isn't going to improve an image that is outputted by the softare at half that resolution.

This despite the fact that this is precisely what I have been doing, it seems: setting resolution to 300 dpi at the software stage but setting the firmware to highest quality.

At some point, I'll test this, but I probably won't get around to it for a bit. I feel less pressure now, as I recent reverted to connecting my printer by USB rather than via wireless, and that has sped up printing.

[xpatUSA]

... I was simply trying to confirm that the process has two choke points, regardless of the software choice. In both cases, the first is when you set the resolution in software--by doing the resizing arithmetic yourself in photoshop, or by telling Lightroom to do it. The second is when the OS passes the image to the printer's firmware ...

Dan, I looked up "choke point" on-line but, not being one who prints, I couldn't quite see it's meaning with reference to the selection of resolution ... other than the lower the selection, the lower the resolution. Am I missing something?

[DanK]

Ted,

Here's what I think is going on. I could be entirely wrong, of course.

While you are still in editing software, you can change resolution easily, by deleting data or imputing more data. However, at the end of the editing process, you output a file with a specific number of pixels. Manfred does that manually in photoshop. I do it semi-automatically in Lightoom by giving it a print size and a ppi value. Either way, once it goes out the door, it is what it is. He has been setting his to 360 dpi. I've been using 300, which is the Canon equivalent.

The printer firmware, in contrast, has to take the resolution it's given. If you tell it to print at 720 dpi, I think it will just use two dots per pixel, as Manfred stated earlier on. that's not going to increase resolution.

So perhaps a more accurate way of saying what I think is going on is that the resolution produced by printer is less than or equal to the resolution fed to it. So, by that logic, if you want to see improvements from setting the firmware to 720 dpi (assuming one can see improvements), it stands to reason that you need to size it to 720 dpi in the software.

[xpatUSA]

Ted,

Here's what I think is going on. I could be entirely wrong, of course.

Thanks for taking the time to clarify, Dan. and please pardon my excessive interest in the minutiae of printing even though I don't print. I'll express my understanding below, irritatingly para by para.

While you are still in editing software, you can change resolution easily, by deleting data or inputing more data.

Generally by re-sampling upward or downward, occasionally downward with just plain cropping.

However, at the end of the editing process, you output a file with a specific number of pixels.

With both aspect ratio and number of pixels chosen to match a target print size and resolution.

Manfred does that manually in photoshop. I do it semi-automatically in Lightoom by giving it a print size and a ppi value. Either way, once it goes out the door, it is what it is. He has been setting his to 360 dpi. I've been using 300, which is the Canon equivalent.

OK, say 8x10" is the target, then aspect ratio is 4 is to 5 paper-wise so your source image for say 300dpi would be 3000x2400px and nothing else.

The printer firmware, in contrast, has to take the resolution it's given. If you tell it to print at 720 dpi, I think it will just use two dots per pixel, as Manfred stated earlier on. that's not going to increase resolution.

While I realize that "dot" here doesn't mean a squirted droplet - I'm floundering with "two dots per pixel" because my humble Canon all-in-one doesn't mention that action. Could we go deeper?

So perhaps a more accurate way of saying what I think is going on is that the resolution produced by the printer is less than or equal to the resolution fed to it. So, by that logic, if you want to see improvements from setting the firmware to 720 dpi (assuming one can see improvements), it stands to reason that you need to size it to 720 dpi in the software.

OK (I theenk).

Thanks again,

[DanK]

Sorry, I shouldn't have used "dots", since it has several meanings. What I meant is using two nozzles for one pixel. If you go back to the beginning of this, what I found puzzling is that everyone says that the "native" resolution of Canon printers is 300 dpi. However, I found documentation that the nozzle pitch is 600 dpi. (For brevity, I won't keep inserting the comparable numbers for Epson, but the same issue applies, however.)

Manfred responded:

as the printers use two nozzles to make a single "dot", so the best we can get out of the printers is 360 / 300 dpi.

So, my point is that if one is feeding a 300 ppi file into a printer that has a nozzle pitch of 600 dpi, it can't produce 600 different dots per inch. My hypothesis is that if there is anything to be gained by setting the firmware to 600 dpi (720 in the Epson world), it would only be by outputting 600 ppi files from the software.

Therefore, if I get around to testing this, I'll run only three prints:

1. Output at the standard 300, firmware set to 300. (I think that is "high", but I have to check this.)
2. Output at the standard 300, firmware set to 600. (I think this is "highest", but again, I have to check.) If my understanding of "highest" is correct, this is what I have been doing since I bought the printer.)
3. Output at 600, firmware set to 600.

My hypothesis is that if there is a difference in resolution, it will only be #3, and if there is a difference in quality between #1 and #2, it won't be resolution per se.

[xpatUSA]

Sorry, I shouldn't have used "dots", since it has several meanings. What I meant is using two nozzles for one pixel. If you go back to the beginning of this, what I found puzzling is that everyone says that the "native" resolution of Canon printers is 300 dpi. However, I found documentation that the nozzle pitch is 600 dpi. (For brevity, I won't keep inserting the comparable numbers for Epson, but the same issue applies, however.)

Manfred responded:



So, my point is that if one is feeding a 300 ppi file into a printer that has a nozzle pitch of 600 dpi, it can't produce 600 different dots per inch. My hypothesis is that if there is anything to be gained by setting the firmware to 600 dpi (720 in the Epson world), it would only be by outputting 600 ppi files from the software.

Therefore, if I get around to testing this, I'll run only three prints:

1. Output at the standard 300, firmware set to 300. (I think that is "high", but I have to check this.)
2. Output at the standard 300, firmware set to 600. (I think this is "highest", but again, I have to check.) If my understanding of "highest" is correct, this is what I have been doing since I bought the printer.)
3. Output at 600, firmware set to 600.

My hypothesis is that if there is a difference in resolution, it will only be #3, and if there is a difference in quality between #1 and #2, it won't be resolution per se.

thanks again. Would be interested in the results when you get a round tuit ...