What do you consider as a high-res file?

[~*~ElmoHolz~*~]

With the megapixels increasing in new cameras, photos (in terms of pixels and file size (mb)) are getting bigger...

What do you consider to be high resolution when providing photos to clients or in general?

Just exported a batch of Jpegs from Lightroom and decided to delete them and adjust the settings for export again as the files were large 10MB+ (in my opinion for Jpegs). And these photos were probably not going to be used anytime soon...

What happened to the days of 1mb = 1 photo

Edit: for photos that are probably just going to sit on my computer I thought the files don't need to be that big...

 

[buckas]

the original processed image is the high res file, before resizing for web etc

usually anything bigger than 2400px longest side is good to go out

 

[DemiLion]

Purely personal, but in my eyes Low Res is for digital use whilst High Res is for print.

Nowt scientific about it.

Being more exact, LR would normally be either 1024 or 1280 and HR 3000-3500.

 

[Nod]

I always shoot in large, fine JPEG and on the D800 that means fairly large files... Not when compared to what raw files would be - I had a play and managed one (1) of them on a 128MB card I have kicking around! I always keep the original full res files since it's easy to shrink them but rather harder to enlarge them. (My favoured final product is prints - and big ones so I need as many pixels as possible to play with!)

 

[martsharm]

With the onset of Retina iPads, I like to export to Flickr 2048 on the long edge, down from 4000 out of camera. This means there's a pixel to pixel correlation on the iPad, and hides a multitude of sins shooting-wise. Surely a retina iPad is about as hi-res as it gets for online.

Printing - always send highest res possible after processing.

 

[Teflon-Mike]

File size is NOT 'resolution'!!!!!!!!!!!
It's not even all that indicative of resolution!
File-Size is to Image-Size what Weight is to Volume!

What do I consider 'Hi-Res'?

FILM!

35mm Halide resolves at least four times the detail of highest pixel count Digi's by my reckoning... a good 35mm film is worth at least a 100MPixel sensor any day. Medium Format is at least two and a half to five times that, and then you have cut-sheet!

Would make pretty much all digital 'low-res' I reckon!

However... once upon a time, anything over a million pixels, was deemed 'Hi-Res' in Digital.... About 1200x800 ish. It's probably still valid enough; sizing to web at around 1000 on the long-side.

 

[PMN]

What would I call a high resolution file?...

...80 megapixel drum scans from my medium format film camera. My 5Dmk2 doesn't even come close.

 

[Teflon-Mike]

What would I call a high resolution file?...
...80 megapixel drum scans from my medium format film camera. My 5Dmk2 doesn't even come close.

And I bet even then, that the scanner is still not getting everything that's non the film!

 

[Steve Smith]

What do I consider 'Hi-Res'?

FILM!

I considered posting something similar then decided not to as I'm sure it will attract many opposing views. Well done!!


Steve.

 

[Pookeyhead]

Hi res for me is a 100MB or more 8bit TIFF or 200MB 16bit TIFF. Anything less is not really good enough to print IMO. There'll be loads of people who will argue this, and say "I can print 15MB TIFFS and they look great"... well, good for them... I clearly have higher standards is all.

That's my 2p worth.

35mm Halide resolves at least four times the detail of highest pixel count Digi's by my reckoning... a good 35mm film is worth at least a 100MPixel sensor any day. Medium Format is at least two and a half to five times that, and then you have cut-sheet!

I have to disagree. I have 35mm films here I've scanned with a Imacon Flextight scanner at maximum res and bit depth and they're not even close to the output of the D800 let alone the IQ180.

If you disagree... post up a 35mm scan at full res... give it all you got... maximum res, bit depth etc... I'll beat it with a D800 file without breaking a sweat. Don't bother linking to academic articles please.. forget numbers and theoretical resolutions... real world stuff please.

35mm sucks. I love film, but to say 35mm is equal to a 100MP sensor is just ludicrous.

 

[DemiLion]

Hi res for me is a 100MB or more 8bit TIFF or 200MB 16bit TIFF. Anything less is not really good enough to print IMO. There'll be loads of people who will argue this, and say "I can print 15MB TIFFS and they look great"... well, good for them... I clearly have higher standards is all.

It's not a question of standards, it's about sending the appropriate sizes file for the job!

 

[Steve Smith]

I love film, but to say 35mm is equal to a 100MP sensor is just ludicrous.

I think 24MP is closer to reality.

That would equate to about 90MP for 6x6 and 300MP for 5x4.


Steve.

 

[Pookeyhead]

It's not a question of standards, it's about sending the appropriate sizes file for the job!

Also depends on whether the files are your archive version, or a specific copy for a specific purpose.

For archive... at least 100MB TIFF for me.

 

[Teflon-Mike]

If you disagree... post up a 35mm scan at full res... give it all you got... maximum res, bit depth etc... I'll beat it with a D800 file without breaking a sweat. Don't bother linking to academic articles please.. forget numbers and theoretical resolutions... real world stuff please.

My opinion is based on pretty crude empirical experimentation. Don't know how you are defining 'resolution', but I deem it to be a measure of how much 'small detail' you can see in an image.

First principles; more pixels you get, more detail you get.

60x40 pixel image (resized to 600x400 for screen) you dont get bog all detail

600x400, and you have 100 squares instead of one, so can see detail in that block rather than just an average blurr.

Here's a quickie.... side by side scans at 17Mpix and 100Mpix (800x500 crop section 1:1 from 100MPix . 17Mpix enlarged to scale)

Top scanned at 17Mpix, bottom at 100. which has 'resolved' more discernable detail?

And, does a 36x24mm 35mm film frame have more or less detail, than can be resolved by 24,000,000 pixels?

So I have a slide.
Pop it in the enlarger, and project it onto my D3200 laid on its back & make an exposure. I get a 24Mpix reproduction of the entire image.
Crop that, to look at detail in the image, and at some point all I see is pixilated mosaic.
Pop camera back under enlarger. Crank the head up; take exposure of the crop section I am interested in. I now have 24Mpix in the area the camera would have only captured perhaps 2000 pixels, and I can see more 'resolved' detail.

Ie, there has to be more 'resolution' on the 35mm frame than can be captured by a 24Mpix sensor.

How much more? Well.. so far I have had reasonable success taking 24Mpix crops from enlargements, aprox 5x4".. and seeing detail, without getting down to film grain, not resolved in single 24Mpix image. Implies that I would need a 100-150Mpix sensor, to capture the same level of detail as can be seen on the film.

Taking the enlarger head up, 10x8", 24x16 under it, is seeing grain structure; but it's only seeing 100th of the full image area!

Back down, and back down, and you start to see more image detail, before the grain dissapears... here....

That's 800x500 pixels, taken from 6000x4000 crop, of a full image that would have needed 48,000x3200 pixels. so, 8x8 : 1x1. Pixilated at 24Mpix resolution, that section then would be just 100 pixels wide by 68 tall, and look like this, enlarged to the same view-size.

Pretty crude experiment; BUT... at the extremity, film appears to my eyes to be 'resolving' more detail, more smoothly than digital.

And that, as it seems to have turned out, is fro a not particularly wonderful original image, a bit muggy around the edges, that could be lens definition or motion blurr, from camera shake or fact its an action shot.... but doesn't really matter.

May be a crude experiment, and your comment about colour depth might open the debate, and take us off into the symantics between resolution, defanition & clarity....... But it is a fair enough test.

How far can you magnify the halide image before you stop seeing more detail than the sensor is resolving into pixels? Answer, by this test, is somewhere between four and sixteen times.

Take it back to first principles; more pixels = mpre possible detail.... and there is detail on the film to fill those pixels, up to around 100MPix.

Whether the detail is WORTH resolving... entirely open to debate, but... it is definitely there!

 

[PMN]

Mike, with the greatest respect if there's one person on this forum you don't need to explain the concept of resolution to it's Pookeyhead.

 

[Pookeyhead]

Mike.. all you're doing there is showing the difference between the analogue structure of the film, and the pixel based nature of the digital camera. You are using the SAME image for both tests. The detail is identical in both A low speed film will have smaller grains than a D3200 has pixels, yes, but the image resolution is independent from that. If grains get large enough they will impact upon it, as will pixels, but the image resolution is actually independent from the SENSOR resolution in the real world, as lenses start to play a role. You are projecting the image directly onto the sensor via an enlarger Mike... that's not how we take images in the real world

Can I suggest if you want to do this properly, you shoot a test image on film, then shoot the same image on a digital camera. Also... scan it... as that is the only real way film images will be disseminated in the real world unless you invite the world to your house to see your finely crafted hand prints Even then.... read on.

The grain structure is NOT image detail Mike. You may need to enlarge more to show grain in a low speed 35mm image than you do to show pixelation from a D3200, but that is all. The noise levels are pretty intrusive in either case at those magnifications, but the random analogue nature of the film "noise" will always appear smoother tonally.

And that, as it seems to have turned out, is fro a not particularly wonderful original image, a bit muggy around the edges, that could be lens definition or motion blurr, from camera shake or fact its an action shot.... but doesn't really matter.

Of course it matters. Lens definition is what limits small formats Mike. The base structure of the film appears smoother if you use low speed, fine grain emulsions yes. and taken with care, a low speed 35mm emulsion can rival a 24MP image.. however.....

There's also one issue you're not addressing. The ONLY way to view a film shot properly is either a projected slide, or decent quality print enlarged (projected with a great lens via the negative in an enlarger.. not scanned) from the negative. In this day and age, that isn't going to happen is it. If you need to disseminate your images on 35mm film, you'll have to scan them, and that adds another layer of deterioration as well. I scan using a Imacon Flextight X5 Mike... it's as good as it gets unless you want to start with incredibly high end drum scanners. The results are noticeably worse than when viewing the transparency using a decent projector using a decent lens.

I've only realised the potential in the 36x24mm format since using the D800, and I've been shooting 35mm for over quarter of a century.. I've got a massive archive of stuff here shot on all formats from 110 to 10x8" film, and I can confidently tell you that the output from the D800 is easily as good as 6x4.5cm film if you use primes and use it properly. It's not quite as good as 6x7, but close enough for me to have rarely touched my RB67 since I bought the Nikon. I may not shoot much commercially any more, but that doesn't mean my standards have slipped, as I still occasionally exhibit work, and when I do, I usually print at A2 minimum, and often as large as 2 metres across. I could never have done that shooting on 35mm film... not to the same standards anyway. That's 6x7 or higher territory.

I've just found a negative of a scene fairly local to me. It was shot using a FM3A and a standard prime lens on ISO100 film. I'll scan it on the X5, and then go to the same scene and reshoot it on the D800. We'll see what's what then shall we?

 

[Teflon-Mike]

This is probably going to come down to a matter of symantics, where we ultimately feel agreaved at arguing over stuff we probably mostly agree on! But, lets get on with it!

Mike.. all you're doing there is showing the difference between the analogue structure of the film, and the pixel based nature of the digital camera.

Pretty much, yeah. I'm going back to first principles; without clouding the issue, compounding simple detail 'resolution' with overall 'Image Quality'; they are not one and the same, and you know it.

What is the smallest detail of the 'scene' that can be resolved to image?

Analogue nature of film, you have already admitted:-

You may need to enlarge more to show grain in a low speed 35mm image than you do to show pixelation from a D3200, but that is all. The noise levels are pretty intrusive in either case at those magnifications, but the random analogue nature of the film "noise" will always appear smoother tonally.

So how much more enlargement, before you see pixilation or noise? I think that is the question here, and then; When you agree the more random nature of film grain, provides smoother tonal transitions... how much more 'fine detail' might that allow to be seen?

By my rather crude empirical estimation, I take content of a 35mm frame, and enlarge it; well, its on screen at about 10x7" at the moment... original frame was 36x24mm, crop is aprox 4.5x3mm.... something in the order of 50xmagnification.

At that magnification 24Mpix full-frame digitisation has pixilated out, and detail is indistinct, edges & lines leggo-bricked.

Same magnification, 1,500Mpix full frame digitisation, is showing grain, but rendering edges & line in that grain, sufficiently smoothly to be able to pick out 'detail'.

Similar magnification applied to 100Mpix & 17MPix. again, there is some grain still in the higher MPix scan at that magnification, but, it is still rendering edges and transitions more 'discernably'.

Looks Horrible? The grain is 'intrusive'? YEAH! I dont dispute you. But you cant argue that the edges and shapes ARE rendered within it, the 'detail' has been 'resolved'!

Of course it matters.Lens definition is what limits small formats Mike.

I think it limits ALL formats... if we are honest; it is merely a question of to what degree.

Larger format, the less multiplication needed to put the scene on the sensor. Less multiplication, less any error is going to be multiplied. Back to basics, back to first principles.

But that's to take the topic off, again, into tangential argument of image quality, not simple image resolution, and a sub debate on what is more important, lens resolution or camera resolution... to wit cutting to the chase, the answer is both; Have a lens that can resolve more detail than the sensor, you ent going to get it. Have a sensor that can resolve more detail then the lens, it ent there for you to get.

Which takes us to a suggestion we are likely to agree on, which is that overall image quality is not dependent on the sum of the parts, but the balence between them, limited by the weakest link.

Can I suggest if you want to do this properly, you shoot a test image on film, then shoot the same image on a digital camera.

Can I suggest that that is the right answer to another question?

Its the 'real world' answer to test which delivers better 'image quality'. On Bag, everything thrown into the pot, especially if you add..

Also... scan it... as that is the only real way film images will be disseminated in the real world

Bringing us to a point of probable agreement.... 'in the real world' displayed on a 1080ish pixel screen.... any resolution above that is probably redundant ANYWAY. Any detail that cant be cought around 1.3Mpix is unlikely to be seen by very many people.

The OP's question, was; "What do you consider 'High Res'?" and I answered "Film"

You asked me to explain and support my opinion that the 'resolution' of 35mm film was around 100Mpix equivilent, and do it with 'real world' explanation.

I have done that; I have shown the empirical evidence I looked at; 24Mpix scan of full frame image... pixilating out, at high magnification, where higher Mpix scan, up around 100Mpix I could still see detail not captured at 24Mpix.

And now you are challenging me to ultimate IQ tests comparing film to digital?

OP's intent might have been to ascertain what 'resolution' people felt gave better image quality, or what was an adequete sensor resolution, but question asked wasn't about image quality, but simple detail resolution.

 

[Steve Smith]

There's also one issue you're not addressing. The ONLY way to view a film shot properly is either a projected slide, or decent quality print enlarged (projected with a great lens via the negative in an enlarger.. not scanned) from the negative. In this day and age, that isn't going to happen is it.

Why not? It's the way I do it.

In any digital vs. film discussion scanning is always the weakest link.


Steve.

 

[Pookeyhead]

I thought I'd explained that already:

Because no one will ever see your work except you and whoever else happens to be in the room. If you win a competition, or get your work exhibited in a gallery, they're not going to project your slides.. they'll be scanned digitally, and then printed.