The merits of RAW vs JPG
Discussion
Bacardi said:
Capture One is very slow but, IMHO, produces the best quality. Adobe Raw converter is fast, coverts a 1ds RAW to 32 mb in about 5 sec's (2 1/2 year old Mac dual 1ghz G4). You run batch process with an action so that's 500, give or take, allowing a little for saving to disk time, per hour. Depending on camera/file, your milage may vary. For that extra quality, yes, you will have to do more work and possibly buy some more hard drives. Only you can decide if it's worth it (think I know the answer).)
You do do a batch conversion to JPEG? Isn't that defeating the point somewhat?
That stuff about fractals sounds interesting. Can anyone point me to a website which explains about them?
dcw@pr said:
You do do a batch conversion to JPEG?
Usually no. I convert to TIFF (contrary to Phil's description on the other thread re: '1) TIFF - pointless, outdated format.....', it's still the most widely accepted file, along with EPS, for page layout and print). For RAW conversion, 99% of the time I use Capture One which does allow multiply destination formats. Sometimes a client may want CMYK TIFF for print and sRGB JPEG for web. It's easy to set up the app to process the same RAW file into different folders, with different colour spaces and different resolutions.
dcw@pr said:
Isn't that defeating the point somewhat?
Not if you want to retain detail in your highlights.
dcw@pr said:
That stuff about fractals sounds interesting. Can anyone point me to a website which explains about them?
No doubt Phil can fill us in about b-spline Discrete Cosine Transform inter-nodal kernel wavelength algorithms with a binary multiplier
or some such techno speak. 
(Awaits correct description). But for the rest of us, practically, Fractals most probably has something to do with commercial plug-ins like Genuine Fractals Which enables loss-less compression and better interpolation of files.Bacardi said:
Fractals most probably has something to do with commercial plug-ins like Genuine Fractals Which enables loss-less compression and better interpolation of files.
Interesting. I know about using fractals to make pictures bigger, but I never knew they could make them smaller too.
Eer, no. The fact that I'm shooting RAW and processing all my files from RAW means that I utilise all the qualities from the RAW file for every shot. Not quite sure what you're trying to get at?
That I end up with an 8 bit file is irrelevant. It's how I get there that's important. An 8 bit JPEG from the camera, using the camera's limited software and processing power is not going to be as good as an 8 bit JPEG produced from much better software. Period. Whether the difference is worth it or perceivable to your work flow is down to you. If you only supply work as files for web in sRGB say, then, most probably, don't bother. If you supply work for high quality repro and care about the quality, then bother.
That I end up with an 8 bit file is irrelevant. It's how I get there that's important. An 8 bit JPEG from the camera, using the camera's limited software and processing power is not going to be as good as an 8 bit JPEG produced from much better software. Period. Whether the difference is worth it or perceivable to your work flow is down to you. If you only supply work as files for web in sRGB say, then, most probably, don't bother. If you supply work for high quality repro and care about the quality, then bother.
Bacardi said:
That stuff about fractals sounds interesting. Can anyone point me to a website which explains about them?
No doubt Phil can fill us in about b-spline Discrete Cosine Transform inter-nodal kernel wavelength algorithms with a binary multiplier
The main man for Fractal compression was Michael Barnsley. He was the brains behind Interated Systems (who seem to have going tits up). He never let the top software source code out, but may be showing the maths in white papers now. You may remember a tomorrows World piece he did. 30 seconds of really high quality video played off a 1.2MB floppy disk. Yes, some video compression has got some way to this standard now, but it's still poor in comparison to what he did, over 10 years ago!
The best bit of his system was it's encoding/decoding profile. It took hours to encode it, but a simple 386 machine could decode it in real time. Interated Systems sold some big DSP cards to help with the encoding if I remember correctly.
Yuval Fisher has done some nice work in the same area, but never seem to get the results that Barnsley did.
Shove either name in Google and you'll get a few good hits I'm sure.
Oh, and the correct techincal term for this stuff is 'F'ing hard maths stuff' as you asked
Do not, I repeat DO NOT look at the maths for long. some of the pictures are nice however. TIFF (Totally Idiotic File Format). Nice to know it's still alive....like Chelsea Pensioners
Ah, 'F'ing hard maths stuff'. I new it would have a technical description, thanks for that. 
Interesting to hear your, slightly negative, comments about TIFFs, from a programers point of view. In my little world it's an 'Industry standard'; still, looks fine printed on a page, so I don't give a monkeys....
thatphilbrettguy said:
TIFF (Totally Idiotic File Format). Nice to know it's still alive....like Chelsea Pensioners
Interesting to hear your, slightly negative,
comments about TIFFs, from a programers point of view. In my little world it's an 'Industry standard'; still, looks fine printed on a page, so I don't give a monkeys.... Bacardi said:
Eer, no. The fact that I'm shooting RAW and processing all my files from RAW means that I utilise all the qualities from the RAW file for every shot. Not quite sure what you're trying to get at?
sorry, not trying tp "get at" anything. I'm just curious about how your workflow goes. From what I can tell, you take the photos in RAW, then batch convert them into 8 bit TIFFs. At which point do you utilise the abilities of the RAW format?
Bacardi said:
Interesting to hear your, slightly negative,
Yup. Couldn't agree more. Industry standards are good. Did you know that the IEEE standard for floppy discs up until not long ago was 720K 5-1/4"....
I've got this great idea for an almost infinite resolution / bit depth file format. It's even got it's own data storage system that just needs a torch to decode it. Hhhmmm I've got a bad feeling about this.....
thatphilbrettguy said:
I've got this great idea for an almost infinite resolution / bit depth file format. It's even got it's own data storage system that just needs a torch to decode it. Hhhmmm I've got a bad feeling about this.....
I'm working on software emulation of hardware. If it works, I won't be able to find my computer though...
dcw@pr said:
I'm just curious about how your workflow goes. From what I can tell, you take the photos in RAW, then batch convert them into 8 bit TIFFs. At which point do you utilise the abilities of the RAW format?
Sorry, I'll try to be clearer. This is my workflow.
Depending on the job in hand, I usually check each file individually and adjust curves and levels as necessary. Although in Capture One I usually just click the 'Auto' button, nine times out of ten it works just fine. However, If the job has many captures shot under the same lighting/exposure condition, I will make adjustments to the first file and then batch apply these settings to all the other captures. I also 'Tag' only the files I want to process.
I usually shoot a Macbeth color/gray card in the first of a set of captures and again, adjust the gray balance for the first capture and batch apply to the rest. Depending on whether I'm shooting with the Phase One back or 1ds and depending on the subject matter I can select off the shelf camera profiles such as 'Daylight', 'Portrait', 'Product Flash' etc. to further improve the end result (I could make custom profiles for a given situation but, there is a limit
). I can also use built in, pre-set, film curves for more contrast or more shadow detail. Then I set the sharpening, depending on camera used and final use. If I have shot on the 1ds at higher ASA I apply noise and banding suppression as necessary. Sharpening and noise reduction settings are applied to all files as they are processed. Finally I set up the destination, multiple, if needed, resolutions, file formats and desired colour spaces. Then select all captures, click 'Process' and go to lunch. Or, work on something else. Capture One works doesn't hamper other computer work so I may be doing some PS work in the foreground whilst it chomps away in the background.
Sounds like a load of work? Maybe, but, I wouldn't do it if I didn't think it gave better results. The software itself is like having the next camera model up from using Nikon or Canon's own RAW conversion software. You can see an overview here and to see the difference in quality open up the pictures of the portrait at the bottom of the page. (Right or control click the yellow text under the pics to open up the full res files into separate windows).
Chalk & cheese.
Another review can be found here
I sound like a C1 salesman. If you want to look at doing batch RAW conversions with PS CS go here: www.russellbrown.com/body.html and check out 'Dr. Brown's Image Processor 2.0 Java script utility'. It will process faster and give nearly as good a result as C1, but, IMHO, it's a messy solution.
Hope that's a little clearer now?
Again, your milage may vary and it might not be worth the extra effort with your particular work practices.
HTH, cheers
Paul
(edited because I can't smell priperly)
>> Edited by Bacardi on Tuesday 31st August 19:46
Message Board | Photography & Video | Top of Page | What's New | My Stuff