Digital latitude: improving?

[msbarnes]

I don't really follow digital but I have an old Nikon D70, and since going film, dealing with the narrow latitude is annoying.

1. Do newer sensors have noticeably better latitude? I know there are other improvements, such as in resolutation and dynamic range but I'm not so sure about latitude. I'm not going to go out and buy a Nikon D800 but I'm wondering if it is any betterr in this regard?

Thanks

 

[Speedfreak]

Exposure latitute is linked to dynamic range.

And yes, there are vast improvements.

http://www.dxomark.com/index.php/Ca...brand)/Nikon/(appareil2)/197|0/(brand2)/Nikon

 

[Photo_Smith]

I guess there has been some improvement though not as great as you'd think.
From the same site as the link above comparing two 1.5 crop APS sensors in Nikon Prosumer range (which is what your D70 was when released)
http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/664|0/(brand)/Nikon/(appareil2)/197|0/(brand2)/Nikon

No so cut and dry, better but not massively so.

 

[msbarnes]

Thank you, that was very helpful.

I won't be giving up my film anytime soon.

 

[chikubi]

I guess there has been some improvement though not as great as you'd think.
From the same site as the link above comparing two 1.5 crop APS sensors in Nikon Prosumer range (which is what your D70 was when released)
http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/664|0/(brand)/Nikon/(appareil2)/197|0/(brand2)/Nikon

No so cut and dry, better but not massively so.

D3100 isn't even close to prosumer, it's the lowest entry-level Nikon there is. Do the same comparison with something like a Pentax K5 and the results are much more telling. 14.1 stops is nothing to sneeze at.

 

[Photo_Smith]

It is the rough equivalent of the D70 in the Nikon line-up though which was the entry level DSLR in their range before the D50 came out.
I would compare 1.5 crop Nikons rather than other makes as the OP will probably but that brand because of existing lenses.
The reason I pointed it out is it's not fair to compare crop with FF when dealing with noise a DR, digital is improving just not as massively as you'd be lead to believe, to get the 14 stops of DR from the Pentax is impossible outside of a test bench, a real world situation will yield nearer 11 stops for most cameras less with jpg
imatest rates the Pentax Dr 7 stops min and 10 maximum.
http://www.imaging-resource.com/PRODS/K5/K5IMATEST.HTM

Total dynamic range is good at 10.5 f-stops

Don't believe the hype
We still have a way to go......

 

[chikubi]

Even using that data, 10.5 is significantly higher than the 5.8 of the D70; and 7.2 for its jpgs is a fair increase as well.

And while I now understand your logic behind choosing the D3100, it's somewhat of a moot point in that the OP was asking about digital in general, and if you're going to make a general comparison you really should be basing it on the best capability that the current crop of cameras provides; that's not found in an entry level camera.

In any case, improvements have been made, but whether it's enough for the OP is up to him.

 

[mdarnton]

In general, the situation is much better now than when the D70 was made, because the higher end has moved up quite a bit. Thus, the statistical average is for higher dynamic range. What range you will get with a new camera. . . . well, obviously it very much depends where you choose to put your money. If dynamic range is important to you, you'd better shop carefully.

Me, if I get a new digital it's going to be either a Nikon D800, or a Sony Nex-7, because I really want that range, and those two cameras will take my Nikon lenses easily.

 

[clayne]

Exposure latitude is not really about dynamic range at all - it's entirely about how the shoulder handles saturation. You could have 100 stops of DR but if the shoulder hard-limits/cliffs your signal - you have no latitude.

Compression, post-expansion via contrast, etc. Kinda hard to do that with digital when past the 0-point there's nothing there due to no inherent signal compression, aka LATITUDE.

 

[Speedfreak]

Exposure latitude is not really about dynamic range at all - it's entirely about how the shoulder handles saturation. You could have 100 stops of DR but if the shoulder hard-limits/cliffs your signal - you have no latitude.

I dont know what you are talking about. DXO doesnt seem to measure that either.

I know that current sensors hold a lot of detail both in highlights and shadows so that I can easily push / pull without clipping. Much more so than with scans made from negatives. And I dont even shoot raw which holds even more data.

If your workflow is analog from exposure to print than all that doesnt matter anyway. But as soon as a scanner is involved film loses. There is hardly any scanner on the market that reaches the "latitude" of an entry level dslr.

 

[Photo_Smith]

But as soon as a scanner is involved film loses. There is hardly any scanner on the market that reaches the "latitude" of an entry level dslr.

This just plainly isn't true there are plenty of scanners that can do that. If I have a wider scale single negative then I can make one scan for the shadows, one for the highlights and one for the mid tones–then merge them in PS.
The D-max the scanner is capable of set a ceiling in that respect normally negative film would have a D-max of less than 2-2.5 slides are higher but have lower DR.
It's not a that hard to take a SBR of 10,000:1 and with the right film (say HP5) judicious metering, increased exposure and a compensating development record the whole scene in order to present a printable negative which will also be easy to scan.

In fact scanning and printing gives a wider range than traditional printing where you need to be an expert to get more than 10 zones.
So film capture allows for compression to take place in the highlight shoulder region of the curve at capture, that compression could separate a diffuse highlight from a specular one where the difference (SBR) could be five stops in the actual scene, but the print their relationship may be less than one.
Its a non linear relationship that the eye finds an acceptable illusion of reality.

 

[kanzlr]

you can just tell your scanner to do a HDR scan of the negative 🙂

 

[clayne]

Anyone who has spent any amount of time in recording engineering and/or working with reel to reel or other similar analog media and/or signal compressors knows what's going on here. It's no where near as simple as a simple dynamic range measurement. Sure, sound vs light, but it's the same thing for this context. Scanning an already compressed signal is not a problem at all for the scanner nor is it a problem with traditional printing and using a higher contrast grade to perform expansion.

Digital dudes usually have a severe problem coming to grips with this as it's not a strict cut and dry a:b measurement they can quantify. Welcome to analog. Suffice to say, film has non-linear response to input and will saturate when pushed. It will *not* clip, but only compress. Come to terms with that.

Photo_Smith knows what I'm talking about.

 

[Speedfreak]

This just plainly isn't true there are plenty of scanners that can do that. If I have a wider scale single negative then I can make one scan for the shadows, one for the highlights and one for the mid tones–then merge them in PS.
The D-max the scanner is capable of set a ceiling in that respect normally negative film would have a D-max of less than 2-2.5 slides are higher but have lower DR.
It's not a that hard to take a SBR of 10,000:1 and with the right film (say HP5) judicious metering, increased exposure and a compensating development record the whole scene in order to present a printable negative which will also be easy to scan.

In fact scanning and printing gives a wider range than traditional printing where you need to be an expert to get more than 10 zones.
So film capture allows for compression to take place in the highlight shoulder region of the curve at capture, that compression could separate a diffuse highlight from a specular one where the difference (SBR) could be five stops in the actual scene, but the print their relationship may be less than one.
Its a non linear relationship that the eye finds an acceptable illusion of reality.

Please post a sample picture demonstrating the effects you describe.

Triple-scanning + triple dust / scratch removal would´nt be worth the hassle for me anyway but I guess film shooters would want to learn more about the subject.

Also pleast list "plenty scanners" that are capable of the technique you mention.

Thanks

 

[Photo_Smith]

Its in quite a few images, a well known method film photographers use to compress or expand the tonal range of a print.
If I want the specualr highlight to be paper white it will be exposed on the last possible area of the shoulder of the curve. Depending on how steep the curves shoulder is will determine how many stops you can fit into that shoulder area, compressing them in other words.
If you lower the effective speed of your film and change your development time you can lower the slope of the H&D curve Kodak suggest 18 stops is not beyond B&W film.
Of course you can go the other way and on a dull day compress tones to give 'snap' by underexposing (rating higher) and making the slope higher.

Its a common film photographers technique called expansion and contraction.
All scanners are capable of multiple exposures it's pretty simple I use Vuescan.