[TDvN57]

After shooting with the 645z for six-plus years, my work-flow was simple, shoot in RAW only, process with ACR, output to jpg in Bridge, physical prints with Photoshop.

Earlier this year I started to take an interest in some of the built-in image processing abilities in the camera. To my disappointment i found that none of the settings were captured in the RAW file in a way that ACR could read it. The only way I could access those settings were in Pentax's version of SilkyPix. I tried that as a workflow and it just didn't work for me.

My next stop was processing the RAW files in-camera into TIFF files. I selected TIFF because it is a 16bit file as opposed to JPG in 8 bits. My aim was not to produce a finished image in-camera, but rather for the camera to do some initial processing in-camera that I could not not easily achieve in post processing, using other apps (ACR, Topaz, DXO etc).

I was pleasantly surprised. I started to compare results from editing in ACR versus processing in-camera. The camera was much better in dealing with noise in low light, compared to what I could achieve in ACR; but the in-camera processing would more easily clip the high-ends. So I had to learn to be less heavy-handed with the sliders in the camera.

One evening I wanted to see what the camera would do with really low light conditions. I found that (a) if you reduce the saturation; (b) be light-handed in correcting the shadows; (c) select the "Muted" in-camera profile; (d) add a Cyan toning; (e) set the white balance to "AWB"; (f) set the exposure bias to +/- 0. This will create a RAW file that the in-camera processor can interpret and correct the WB much easier.

I recommend the output to be in TIFF which will give you a 16 bit file you can edit in your normal workflow, plus of course you still have the RAW file.

To really pull the maximum out of this process will take time to learn what works best for a given circumstance. Unfortunately the screen on the camera is not color calibrated, nor is the resolution high enough to confirm critical focusing. Thus, just accept what you see on the camera screen is not what you will see on a high resolution color calibrated computer screen. Given this limitation it took me a while to setup a workflow of in-camera processing saving the output on SD2, check it on the computer, adjust processing settings in-camera and repeat. You can repeat the processing in-camera for the same RAW file as many times as you wish.

I don't think these capabilities are revolutionary for Pentax, but it does add some extra value in some circumstances.

If we ever get a new version of the 645z, I certainly hope Pentax will include these processing settings in an embedded profile in the RAW file, readable with ACR/Photoshop. An upgraded flippy back-screen will be a must as well, with much higher resolution to able to check critical focus in-camera and the ability to color calibrate the screen.

I recommend anybody interested in making their camera work smarter, to experiment with these settings and in-camera processing. I can see much more capability in this feature than what I have been able to get so far.

In future I will most likely switch to this as a work-flow for any low light (high ISO) pictures. Thus far I have found the most added value with the low light/high ISO images. Having said that I know there is much more available.

If you have tried in-camera processing, what has been your experience?

These two pictures were processed in-camera:

#1 was without the Cyan tint and

#2 with the Cyan tint. ISO 25k and exposure bias -1, which I only saw later, it was a left-over setting I used during the day. At +/-0 the results would have been a bit better.

#3 I include this one as a reference: the RAW file of #2, unprocessed.

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[biz-engineer]

I don't have a 645z, but I also found that in-camera processing deliver the best results relative to effort (zero effort). Of course I can achieve better processing the raw without pre-defined Pentax styles, but it takes me some effort to bring my raw image look to the level of straight OOC Jpeg. Camera image processor is super fast, it's been already optimized by Pentax/Ricoh engineer. I've found that the fastest way to get to the best looking results is by pre-processing raws in camera into -1ev, 0ev , +1ev TIFF (converting a complete folder at once, using the efficiency of hardware processor) and then using masks for global/local light/contrast adjustments => cloning out / cleaning unwanted spots => sharpening. In addition, one a certain look is achieved for one image, it can be replicated to many other images, in camera. Also I'd wish the Pentax TIFF be 16bits deep instead on 8bits deep, as 16bits would still give some more room to tweak the files after raw conversion, whereas 8bits not so much. But PDCU can export to TIFF 16bits, albeit at a much slower speed than camera processor.

[TDvN57]

I don't have a 645z, but I also found that in-camera processing deliver the best results relative to effort (zero effort). Of course I can achieve better processing the raw without pre-defined Pentax styles, but it takes me some effort to bring my raw image look to the level of straight OOC Jpeg. Camera image processor is super fast, it's been already optimized by Pentax/Ricoh engineer. I've found that the fastest way to get to the best looking results is by pre-processing raws in camera into -1ev, 0ev , +1ev TIFF (converting a complete folder at once, using the efficiency of hardware processor) and then using masks for global/local light/contrast adjustments => cloning out / cleaning unwanted spots => sharpening. In addition, one a certain look is achieved for one image, it can be replicated to many other images, in camera. Also I'd wish the Pentax TIFF be 16bits deep instead on 8bits deep, as 16bits would still give some more room to tweak the files after raw conversion, whereas 8bits not so much. But PDCU can export to TIFF 16bits, albeit at a much slower speed than camera processor.

I think I am on a similar path to end up where you are with the work flow. I didn't realize that the in-camera TIFF is 8 bit. That is a pity. When I open it in ACR it is set to open everything in 16bit, so it does not really tell me what it was when I loaded it, just what ACR is doing from that point onwards. If I check in the "Details" section in Windows, the tiff bit depth is 24, which I presume is Microsoft's way of saying 8bit times 3 for RBG. However it also reports the DNG files to be 24 bit. So MS must be using some other interpretation of bit depth.

Is there a way to check? because the TIFF files from the 645z are just under 150 MB each, whilst the JPG files are much smaller. BTW the 645z is super slow in processing these tiff files, although I am OK to let it run while I get another coffee. Perhaps I am just spoiled and try to compare it to my computer's processing speed, which is also getting slower by the day. Nothing is fast enough these days, so I wait patiently.

[biz-engineer]

Is there a way to check? because the TIFF files from the 645z are just under 150 MB each

I could tell if a TIFF (uncompressed) is 8bits and 16bits just by looking at the file size, 16bits TIFF are 300MB. Sometimes I just use JPEG if the exposure wasn't perfect, the details aren't very good so TIFF doesn't bring much advantage. What's really important for me is to export in camera raws with various amount of light levels (not sure if I use the proper word), so that shifting levels isn't necessary on exported files (as layers), so that I can mix layers later. In fact, that's a trick I learned from another member here, and contrary to bracketed exposures, there is no need for aligning layers when each layer come from the same raw file.

[phat_bog]

645Z's sensor -> 14bit. just like Fuji's GFX50.

(Hasselblad claims on some recent models sensor : 16bit. but.. the 2+ bits of data don't seem hardware related).
So, i understand that for some applications a photographer wishes 16bit TIFFS instead of 8bit ones when shooting with a 14bit depth sensor, and would be afraid of banding issues in color tones.
But also, actually, most printing is done taking 8bit data.

So, what seems to me more important it's the color gamut size.
Good gamut & 8bit work perfectly fine for 99% of the cases, unless of course some rare very rare cases when dealing with special, really special & complex color tones or if you need to make some huge tonal adjustments.

[TDvN57]

---------- Post added 2021-12-17 at 07:49 PM ----------

[/COLOR]

I could tell if a TIFF (uncompressed) is 8bits and 16bits just by looking at the file size, 16bits TIFF are 300MB. Sometimes I just use JPEG if the exposure wasn't perfect, the details aren't very good so TIFF doesn't bring much advantage. What's really important for me is to export in camera raws with various amount of light levels (not sure if I use the proper word), so that shifting levels isn't necessary on exported files (as layers), so that I can mix layers later. In fact, that's a trick I learned from another member here, and contrary to bracketed exposures, there is no need for aligning layers when each layer come from the same raw file.

That's a brilliant idea. Thanks for sharing.

645Z's sensor -> 14bit. just like Fuji's GFX50.

(Hasselblad claims on some recent models sensor : 16bit. but.. the 2+ bits of data don't seem hardware related).
So, i understand that for some applications a photographer wishes 16bit TIFFS instead of 8bit ones when shooting with a 14bit depth sensor, and would be afraid of banding issues in color tones.
But also, actually, most printing is done taking 8bit data.

So, what seems to me more important it's the color gamut size.
Good gamut & 8bit work perfectly fine for 99% of the cases, unless of course some rare very rare cases when dealing with special, really special & complex color tones or if you need to make some huge tonal adjustments.

Point taken, and I yes I agree the gamut has more influence. If one intends to do a lot of post work, then it might be best to convert the 8-tiff to a 16-dng in ACR before editing starts. My camera is set to AdobeRGB and not sRGB, which only comes into play when doing in-camera processing, in this case yes it would be an important setting to consider. Case-based and preference issues count of course.

Adding to the story:
I took another picture in the house just now and made a +/- 60% crop to show the differences when using in-camera processing in high ISO settings. This picture was ISO 50k, bias +/- 0. The aperture at f/4.5 results in a very shallow DoF and I focused on some of the leaves (fake yes) just to the upper left of the gold ball in the tree.

The purpose is to find the best outcome re-noise and colors, so my focus and not-so steady hand contributed to the super sharp image :-)

Picture #1: is the RAW file developed in ACR with a lot of editing to get rid of the noise.

Picture #2: is the in-camera processed TIFF file, shot on "Landscape profile", but developed with "Muted" and Cyan tint. The resulting file was very flat and un-appealing. However in ACR it took me less than a minute to tweak the saturation and get the colors back. The image also appears sharper than the developed RAW file.

If you compare the two images, also look at the black on the background picture, the color is black and not much noise. On the edge for acceptable, but I am OK with that as an out of focus background. In the RAW file Picture #1, The noise and artifacts are plentiful. Not acceptable to me at all, and I will dump that image without thinking about it.

The colors in the RAW file developed not bad, but not crisp at all. The "flat" TIFF file easily gave the colors back and I would say is much more realistic.

The highlights in the TIFF file blew out and the detail below the lamp is gone, even though it is out of focus, it is still visible. I should perhaps bring down the exposure during the in-camera development, since the RAW file show some of the detail is still there. Alternatively I could just set the bias to -0.5 or -1 and move forward.

My conclusion so far:
The in-camera processing is much better in dealing with noise reduction and color rendering, especially the low and mid ranges.

EDIT:

I did another in-camera edit of the same RAW file: Reduced the exposure one stop, added shadow correction, and gave the tiff file the same treatment in ACR. Now the blacks are even better and the details previously clipped are better. BTW the aperture is at f/5.6 and not f/4.5 as I mentioned above.

[BigMackCam]

This is an interesting approach. Thanks for sharing

My one concern with this (aside from the bit depth) is that when you convert to TIFF, anything below the black point is crushed to black, while anything above the white point is clipped to white. You lose that data and with it the opportunity to recover shadows and highlights to their full extent. For reasonably balanced scenes captured with accurate exposure, that shouldn't be an issue - but for higher dynamic range scenes or captures that are under- or over-exposed, it could be a problem.

Even so, it's an interesting workflow, and if it delivers the results you want, I'm all for it

EDIT: I'm not a huge fan of Pentax DCU, but it seems to give very similar results to in-camera JPEGs when processing raw files, at least on the cameras I use (K-5, K-3, K-3II). I wonder if that would carry over to the 645Z? Might be worth trying...

[biz-engineer]

The problem with camera processing is with capture sharpening, it creates hallows around the edges, even if not obvious when looking at image from the camera, it limits how much output sharpening can be added for printing large. If you use camera to convert RAWs to TIFF I recommend tuning down camera sharpening slider. Something to evaluate on 645z.

[jgnfld]

Have you tried the full up version of Silkypix? I use 10 pro and am quite happy. Of course I only have KP-sized RAW files. Silkypix says they support both 645s.

It's a bit of a quirky programme and the manuals are written in Japanglic, but once you get by the initial learning curve it is very capable.

[TDvN57]

Have you tried the full up version of Silkypix? I use 10 pro and am quite happy. Of course I only have KP-sized RAW files. Silkypix says they support both 645s.

It's a bit of a quirky programme and the manuals are written in Japanglic, but once you get by the initial learning curve it is very capable.

I haven't tried it. At the moment I am entrenched in the Adobe suite, for better or worse. I would be willing to use another program as long as they can output into the raw file, or sidecar. As far as I know silkypix also only outputs to tiff or jpg.

Perhaps one day I will change my mind. :-)

[EssJayEff]

To my disappointment i found that none of the settings were captured in the RAW file in a way that ACR could read it.

By RAW I am presuming you mean .PEF files. If so, what bearing, if any, would there be in saving raw files as .DNG files? Wouldn’t ACR have a way to read your settings? (My line of questioning is one of my own ignorance.)

[TDvN57]

By RAW I am presuming you mean .PEF files. If so, what bearing, if any, would there be in saving raw files as .DNG files? Wouldn’t ACR have a way to read your settings? (My line of questioning is one of my own ignorance.)

I assume your question is related to DNG vs PEF. Personally I prefer DNG. I once took two test pictures of a staged scene with bright light and very dark shadows. The PEF file clipped the extreme darks and lights. The DNG did not. BUT, my test is unconfirmed. Apart from that I could not see any quality or speed difference between the two formats.

In ACR you can select to ignore sidecar files and save the settings in the DNG file, not baked in, but rather incorporating the sidecar into the DNG file. As far as I remember that option is not available with PEF. I prefer to not have an extra file to manage. Just my personal preference.

[texandrews]

Hmmmm....Besides the time involved at capture, I'm just not sure I like the idea of a cooked file coming out of the camera. On the Z, I kind of feel like raw is de rigueur. I'd rather improve my PP skills than try this workflow, myself, and one of the things that so appeals to me about the Z is the malleability of the files. On other cameras I might find this appealing, especially a FF or apsc camera I was using for events or something.

[EssJayEff]

I assume your question is related to DNG vs PEF.

Yes it was, because DNG is a [mostly] open format developed by Adobe, so I thought that might offer something that ACR could leverage that it could not obtain from PEFs. I was wondering that because, as noted by your comment, one can store embedded metadata as part of the DNG file.

FYI: I only shoot DNG so as not to have my images locked into a proprietary format. It is one of the reasons I choose Pentax. When I have to shoot Canon, for example, I convert to DNG. Also, I treat my DNGs like they are negatives. I maintain a master file of DNGs copied straight to a hard drive. I never use them for post-processing. I duplicate image file masters on which I want to work, then assign a prefix (essentially name/date) to the camera-generated file name. I would be very reluctant to have a camera process my images UNLESS I saved those files separately on a second memory card.

Both of those practices stem from my being a photographic archivist for nearly forty years.

[texandrews]

Both of those practices stem from my being a photographic archivist for nearly forty years.

And this fact makes me want to emulate you. People don't take archivists as seriously as they should!