Walter,  unless I’m misunderstanding you, I think you’re perhaps not fully aware of how ‘non-destructive’ imaging works.  I still use the original non-destructive software, Live Picture.  It was a revolution in the 90s because images didn’t have ‘that digital look’.  That digital look was, as you say, each and every image edit resulting in a degradation from the original.  For example, in Photoshop, just opening an 8bit image would reduce it to 7bits.  Add sharpening and perhaps a curve adjustment and you are at 5bits!  That’s what you got in those days.  Most photographers were unaware of why this was and just thought it was what you got with digital, but it wasn’t.  Adobe of course didn’t enlighten anyone.  LP is totally 16bit internally but crucially, only ever made 1 adjustment.  This was when your edits were applied to the image when you output in order to do something with the image.  This could mean printing (LP is also a beautiful rip) or making a tiff for repro etc.  This is the same as what happens in Aperture.  Whenever you need to do something with the image, say send it to a client, then Aperture applies your edits to the exported image in a single step.  Since the time PS got layers with flattening, it amounts to much the same thing … but you must do everything on a layer and flatten.  It’s no good just working on the ‘background’ layer.  Also, when computers became fast enough to catch up with LP’s clever maths, photographers started working in 16bit and this makes a gigantic difference, because you have huge latitude.  Aside from that, I’m slightly surprised with your concern for quality when you are shooting to jpeg and later making adjustments in post.  Now that is a recipe for loss of quality.  I’m assuming you’re working in jpeg because you mention applying contrast settings in camera, as obviously these have no bearing on a raw file.

Regarding ETTR  (I had to look up what this is) I don’t know how far back people’s experience go on this site, but I expect many have never shot film, but myself, having spent a very long time in this business, I’ve always taken ETTR to mean working in the same way I used to do on 5x4 and 10x8 ektachrome, which were my daily tools throughout the 80s and 90s.  We used to refer to ‘pegging the highlights’.  This was so important, because you could perhaps push ektachrome 1-stop in processing, but that was really the absolute limit before getting into noticeable degradation.  ’Holding’ the highlights was very important, but equally, we worked with a narrow latitude, hence all the complex lighting we used to use to reduce dynamic range to workable levels.  So we always ‘exposed to the right’ although we didn’t have histograms, but it was the same thing.  Expose for the highlights, which ensured they had detail.  Light or fill into the shadows the reduce the dynamic range of the image.

John …

I meant no offense. I merely pointed out that that no matter your good intentions ETTR, while a good starting point for many situations, may not always be the correct technique for all situations. I only shared my observations based on working as a full time professional photographer for 38 years … most of that time pushing my tools to the extreme to meet my needs. I know all about “crop” sensors … I’ve been using DSLR cameras since the Nikon D1 in 2000 and have yet to purchase a full-frame DSLR.

Noise is the result of the internal camera firmware recording pixel data through pumping up the gain when exposing the sensor with less than adequate light. When this method is applied, the firmware has to sometimes make a best guess for the the tonal and color value recorded by the individual sensor photo sites. The “wrong” or less than accurate guesses result in what is perceived as noise. When you “expose to the right” … often, mid and lower mid tones as well as shadow areas suffer through a loss of detail the sensor ignores because your exposure is clipping those portions.

What I pointed out, when you expose the image by giving the sensor a more correct amount of light for those tones using a more neutral method - maybe sacrificing the highlights and deep shadows just a bit - but preserving the quarter, mid and three-quarter tones … the firmware has to guess less often resulting in less apparent noise. It all boils down to the physics of light. If you push the sensor and firmware to extremes … it is very likely no amount of post-processing is going to save the image and address the noise issue without losing detail, dynamic range and color integrity.

10

That’s a good shot Walter, I’m curious how much post-process might have been applied - so far as NR is concerned. I do not deliberately use ETTR - also had to look it up :) - but I do check for volume of light around the subject, not just on the point of focus, in an attempt to gain a more balanced lighting. By that I mean I expose and focus to the point of interest and then check the readings in the areas around it to gauge variations. If the lighting round the subject is overexposing I might drop my meter a third of a stop or so; underexposing;the opposite. Doesn’t always work but then nothing in this game always works. Skills and techniques aside I think that good low light shooting is highly dependent upon the body unless you’re dealing with long exposure on a tripod - even then the sensor is going to have great influence on the final outcome. I had the opportunity to shoot a professional ballet alongside its credentialed photog not long ago. Rapidly moving subjects lit almost entirely by moving spot lights. He shot a 1Dx using 100-400 mm L f4.5 - 5.6 IS. I shot a 7D using a 70-200 mm L f2.8 IS. He’s a great photog; I’m passable. But we’d done outside daylight event shooting like this prior and the results were comparable. Not this time. Not even close.

10