Print Size Comparison - A7/A7r (andNikon DF)

[Taylor Sherman]

I'd be interested in removing it and *not* adding the cover glass, as long as adapted manual lenses can still focus to infinity.

 

[bensonga]

The only reason I can think of to remove the A7's AA filter is if it somehow reduces astigmatism, but I'm sure the clear cover glass that is put in its place would still cause similar issues, so I doubt it would be worthwhile.

I don't see a need to remove it and so I won't.

Gary

 

[D&A]

Just a few personal comments to this interesting thread even though I don't own/use either the A7 or A7r cameras......yet!

With regards to Epson LF printers, I too would never let the printer up-rez to 360PPI. From the earliest days of the Image Print RIP, I've been extremely impressed with it's algorithms in order to up-rez a file. I have intensely compared it to virtually every alternative up-rezing method and have used most all available scaling programs past and present that I'm aware of. Both LR and PS have gotten so much better over the years and they have recently narrowed the differences significantly when compared to the RIP and other programs designed to scale up a file.

Up until about 2 years ago I kept returning to investigate Fractals in particular, hearing good things about it, but like Tim, I too was disappointed what it did to files. I also occasionally send 180ppi files to the Espon when working with smaller files from lower resolution cameras as I feel they would be better off natively sending them as 180ppi as opposed to uprezing to 360ppi by any method. Files from higher resolution cameras, even thouse from 40MP or bigger sensors, are almost always sent at 360ppi to the printer.

In the very early days when digital cameras were but 2.1 MP or smaller, I'd work with a variety of techniques on each file in order to print as large as possible without creating digital looking images. Printing very large these days from files that have to be uprezed, still benefits from how the file is handled and processed prior to uprezing

Of course starting with quality fatter pixels like those for the Nikon Df, often times makes it much easier to scale up and print larger than from the native file itself as compared to a higher resolution/smaller pixel camera that too is uprezed the same proportionally. Pixel quality is just as important element as camera resolution, in my opinion.

I feel like many, with all else being equal in terms of identical subject matter, that an increase from say 16MP to 24MP files, has a greater impact on subsequent print size chosen and the quality of details seen, than the math suggests. People tend to view prints at a similar distance whether it be a 16 x 24" or 24" x 35" for example. There are of course many variables to consider, so its hard to make quick and fast generalizations.

Dave (D&A)

 

[lambert]

Just a few personal comments to this interesting thread even though I don't own/use either the A7 or A7r cameras......yet!

With regards to Epson LF printers, I too would never let the printer up rez to 360PPI. From the earliest days of the Image Print RIP I've been extremely impressed with it's algorithms to up rez, having intensely compared it to virtually every alternative method and have used virtually all available scaling programs. Both LR and PS have gotten so much better over the years and they have recently narrowed the differences significantly when compared to the RIP.

Up until about 2 years ago I kept returning to investigate Fractals but like Tim, I too was disappointed what it did to files. I also occasionally send 180ppi files to the Espon when working with smaller files from lower resolution cameras as I feel they would be better off natively sending them as 180ppi as opposed to uprezing to 360ppi by any method. Files from higher resolution cameras, even thouse from 40MP or bigger sensors, are almost always sent at 360ppi to the printer.

In the very early days when digital cameras were but 2.1 MP or smaller, I'd work with a variety of techniques on each file in order to print as large as possible without creating digital looking images. Printing very large these days from files that have to be uprezed, still benefits from how the file is handled and processed prior to uprezing

Of course starting with quality fatter pixels like those for the Nikon Df, often times makes it much easier to scale up and print larger than from the native file itself as compared to a higher resolution/smaller pixel camera that too is uprezed the same proportionally. Pixel quality is just as important element as camera resolution, in my opinion.

I feel like many, with all else being equal in terms of identical subject matter, that an increase from say 16MP to 24MP files, has a greater impact on subsequent print size chosen and the quality of details seen, than the math suggests. People tend to view prints at a similar distance whether it be a 16 x 24" or 24" x 35" for example. There are of course many variables to consider, so its hard to make quick and fast generalizations.

Dave (D&A)

I couldn't agree more. The comparison posted above assumes that the pixels from these different cameras are of equal quality, which is not the case. A 16mp file from a large sensor camera with no AA will be able to achieve a higher resolution print than a 16mp file from a point and shoot. In the case of the A7 / A7R, one has an AA, one does not. This will have some bearing on print resolution/quality over and above the difference in resolution between the two.

 

[jonoslack]

HI There Dave
Interesting remarks most of which I agree with.

However - this remark is exactly why I started this thread in the first place:

I feel like many, with all else being equal in terms of identical subject matter, that an increase from say 16MP to 24MP files, has a greater impact on subsequent print size chosen and the quality of details seen, than the math suggests. People tend to view prints at a similar distance whether it be a 16 x 24" or 24" x 35" for example. There are of course many variables to consider, so its hard to make quick and fast generalizations.

You are converting what is in effect a measure of area (16mp / 24 mp) to a linear measurement (16" / 24") It's almost universally done, and it isn't right.

if you're talking about a 16" x 24" print with 16mp, then the equivalent pixel density for 24mp will create a 19.6" x 29.5" print NOT 24" x 35"

This is because the length relates directly to the number of pixels on the long side: i.e. 16*6000/4900 where 6000 is the number of pixels on the long side in a 24mp sensor and 4900 the number in a 16mp sensor.

Of course, this is thoroughly relevant when considering which camera to get - and as you point out, (other things being equal - thanks Marc) you'll get a better class of pixel with less of them assuming the same size sensor.



All the best

 

[jonoslack]

In the case of the A7 / A7R, one has an AA, one does not. This will have some bearing on print resolution/quality over and above the difference in resolution between the two.

Indeed, but theoretically the large pixels of the A7 should also have a bearing? Whether one cancels out the other is a moot point

 

[Guy Mancuso]

Just FYI this is not easy to read hungover. Happy New Year everyone. I'm going back to sleep. LOL

 

[fotografz]

Of course, this is thoroughly relevant when considering which camera to get - and as you point out, you'll get a better class of pixel with less of them assuming the same size sensor.



All the best

Jono, doesn't this assume that the pixel technology of the A7 is the same as that of the A7R? I have read that the 36meg sensor employs a new "gapless" pixel array, but have not seen that referenced regarding the A7. I cannot seem to locate the actual pixel size specifications for each version of the A7 cameras.

I've also read comments referencing the 36 meg sensor as being the same as the D800 ... but I do not think that is true. Distance to sensor presented new challenges that required different engineering to accommodate the native FE FF lenses I think.

Sony is on the forefront of sensor design and has breeched pixel size barriers using new ideas and designs ... like this sensor referenced below.

New Sony sensor achieves industry smallest pixel size

I don't think we have enough information to reach any conclusions yet. However, practical experience of users that will using both cameras (like Guy) may help shed more light as time goes on.

- Marc

 

[Deleted member 7792]

HI There Dave
Interesting remarks most of which I agree with.

However - this remark is exactly why I started this thread in the first place:



You are converting what is in effect a measure of area (16mp / 24 mp) to a linear measurement (16" / 24") It's almost universally done, and it isn't right.

if you're talking about a 16" x 24" print with 16mp, then the equivalent pixel density for 24mp will create a 19.6" x 29.5" print NOT 24" x 35"

This is because the length relates directly to the number of pixels on the long side: i.e. 16*6000/4900 where 6000 is the number of pixels on the long side in a 24mp sensor and 4900 the number in a 16mp sensor.

Of course, this is thoroughly relevant when considering which camera to get - and as you point out, you'll get a better class of pixel with less of them assuming the same size sensor.



All the best

Exactly Jono. You [once again] have hit the nail on the head.

Furthermore, when printing (using LR and/or PS) I frequently refer to an article written by Jeff Shewe in 2011 where he says:

The bottom line is, if the image you're printing to a high-end inkjet printer has a native resolution at the print size of less than the printer resolution, upsample to the printer's dpi. In the case of Epson, that's 360 dpi, and it's 300 dpi with Canon and other printers with similar print heads. If the native resolution is above the resolution, upsample the image to the higher reported resolution of the printers (720 ppi for Epson and 600 ppi for Canon).
​

Jeff is the acknowledged expert in printing with Epson printers using either LR or PS, having worked with the developers in both Adobe and Epson to "get it right". His book "The Digital Print" is worth reading and using as a guide if one prints his/her own images.

And yes, pixel "quality" matters. Superior tonal gradations from so-called "fat pixels" usually make for better prints at any resolution. As Marc pointed out though, I haven't seen any evidence yet that suggests the pixels from the A7 lead to better prints than those from the A7R.

Joe

 

[jonoslack]

Hi Marc
Happy New Year
You're quite right, - I should have said 'other things being equal' and now I have done so.

Of course, there are lots of other variables which might also effect print size - but the basic principle stands - no good muddling up areas with linear measurements!

Jono, doesn't this assume that the pixel technology of the A7 is the same as that of the A7R? I have read that the 36meg sensor employs a new "gapless" pixel array, but have not seen that referenced regarding the A7. I cannot seem to locate the actual pixel size specifications for each version of the A7 cameras.

 

[Guy Mancuso]

I'm not sure I would put money down on a bet the D800e is exactly the same as the A7r. They seem different to me. But that's just a feeling I have so far. I think the Sony actually has better DR but I have to shoot a lot more to be firm in any if that. I think this answer would have to come from Sony

 

[douglasf13]

Exactly Jono. You [once again] have hit the nail on the head.

Furthermore, when printing (using LR and/or PS) I frequently refer to an article written by Jeff Shewe in 2011 where he says:

The bottom line is, if the image you're printing to a high-end inkjet printer has a native resolution at the print size of less than the printer resolution, upsample to the printer's dpi. In the case of Epson, that's 360 dpi, and it's 300 dpi with Canon and other printers with similar print heads. If the native resolution is above the resolution, upsample the image to the higher reported resolution of the printers (720 ppi for Epson and 600 ppi for Canon).
​

Jeff is the acknowledged expert in printing with Epson printers using either LR or PS, having worked with the developers in both Adobe and Epson to "get it right". His book "The Digital Print" is worth reading and using as a guide if one prints his/her own images.

And yes, pixel "quality" matters. Superior tonal gradations from so-called "fat pixels" usually make for better prints at any resolution. As Marc pointed out though, I haven't seen any evidence yet that suggests the pixels from the A7 lead to better prints than those from the A7R.

Joe

There are tests out there that show that the actual native resolution of Epson printers is 288 ppi, rather than 360.

 

[davemillier]

So, with Epson printers, what is the maximum uninterpolated input ppi that will improve printed detail?

For example, what is the optimum print size from a 16MP sensor (assuming perfect technique)? In other words the just right print size, the size at which going bigger will lessen quality and going smaller will not show an improvement.

There must be a maximum input ppi which represents the absolute sharpest the printer can deliver....

 

[tashley]

And there are other experts who swear blind that you should send to the printer at native resolution and tell the printer to upsample to its maximum resolution. Others who say you absolutely HAVE to use a RIP. Others who say you should only ever up or downsample by factors of 100% or 50% and others who insist that if you want to upsample, you should do it in incremental steps.

In the end there is an awful lot of juju involved, the science is either beyond most of us or we don't have enough facts to work by it. Use a widely recommended method and see f you like it. Try a different method and see if it is better with your printer and your paper and your types of image… and be aware that in 18 months time the 'rules' might be different; in The Digital Print, Schewe says that the interpolation routines in LR have improved greatly in recent versions, for example, so that now you should indeed use LR to upsample to the the printer's highest resolution.

It's a forest. I try to look at the trees.

 

[turtle]

....

It's a forest. I try to look at the trees.

Yes, but we are trying to make the trees as sharp as possible.

 

[douglasf13]

And there are other experts who swear blind that you should send to the printer at native resolution and tell the printer to upsample to its maximum resolution. Others who say you absolutely HAVE to use a RIP. Others who say you should only ever up or downsample by factors of 100% or 50% and others who insist that if you want to upsample, you should do it in incremental steps.

In the end there is an awful lot of juju involved, the science is either beyond most of us or we don't have enough facts to work by it. Use a widely recommended method and see f you like it. Try a different method and see if it is better with your printer and your paper and your types of image… and be aware that in 18 months time the 'rules' might be different; in The Digital Print, Schewe says that the interpolation routines in LR have improved greatly in recent versions, for example, so that now you should indeed use LR to upsample to the the printer's highest resolution.

It's a forest. I try to look at the trees.

Agreed. I've read so much about this over the years, and it really isn't conclusive. I just print right out of LR set at 288ppi to my Epson and don't worry about it anymore.

 

[Deleted member 7792]

There are tests out there that show that the actual native resolution of Epson printers is 288 ppi, rather than 360.

I don't doubt that there are tests and statistics that "show" just about anything. Jeff Schewe is one of the insiders who worked with Adobe and Epson on the basic design and coding of the print interface, including sharpening, for Lightroom, Photoshop and the Epson inkjet printers. He's as expert as I need to advise me on native resolution and upsampling. He's active on LuLa if you want to debate the issue with him.

And there are other experts who swear blind that you should send to the printer at native resolution and tell the printer to upsample to its maximum resolution. Others who say you absolutely HAVE to use a RIP. Others who say you should only ever up or downsample by factors of 100% or 50% and others who insist that if you want to upsample, you should do it in incremental steps.

In the end there is an awful lot of juju involved, the science is either beyond most of us or we don't have enough facts to work by it. Use a widely recommended method and see f you like it. Try a different method and see if it is better with your printer and your paper and your types of image… and be aware that in 18 months time the 'rules' might be different; in The Digital Print, Schewe says that the interpolation routines in LR have improved greatly in recent versions, for example, so that now you should indeed use LR to upsample to the the printer's highest resolution.

It's a forest. I try to look at the trees.

I agree that there is "juju" involved. Since printing is an expensive process, in ink, paper and time, for my Epson 7900, I've used Schewe's expert opinion as my starting point. I find I waste fewer supplies that way, and end up with a better result (for me) in less time than by using trial and error.

It's my 30 years of engineering discipline that favors fact over "juju". YMMV

Joe

 

[Deleted member 7792]

Agreed. I've read so much about this over the years, and it really isn't conclusive. I just print right out of LR set at 288ppi to my Epson and don't worry about it anymore.

And like many who do their own printing, your results likely aren't as good as they could be. Just sayin'.

Joe

 

[vjbelle]

=tashley; Schewe says that the interpolation routines in LR have improved greatly in recent versions, for example, so that now you should indeed use LR to upsample to the the printer's highest resolution.

It's a forest. I try to look at the trees.

I have said it before and will say it again..... I would never use Lightroom for my printing if it involved using Lightroom's interpolation. As Tim says there a lots of ways to approach this... in the end use what works for you.

Victor

 

[vjbelle]

There are tests out there that show that the actual native resolution of Epson printers is 288 ppi, rather than 360.

First time I've ever heard of this..... maybe you could point me to that source.

Victor