The RAW vs JPEG Debate

Hello all...

The RAW vs JPEG Debate is one that has been raging for quite some time now. I wish to share with you some of my own observations and how it might aid in your decision as to which to choose. It will not be a technical discourse as I am not in a position to do so. It will be based on my observations and how the decision will affect the final outcome i.e. the print media.

RAW is akin to a recording that is done directly to Pro Tools without compression and JPEG is what that recording might sound like after converting the signal for MP3 listening. That signal has been compressed and as a result has lost some of the high end and low end definition as well as the dynamic range. This analogy can be directly transposed to visual media. In photography RAW is the pure unadulterated signal. Now why would anyone even consider JPEG unless they felt that their image was not worthy of that kind of rendition. It should not come down to a question of memory or cost of storage etc. It is an image that merits the best resolution possible that may in the future be used for a support that needs the kind of resolution that only RAW can provide.

You may think.."Well it is only a snap shot." Well todays snapshot may be tomorrows historical archive. You are leaving a trace of history for future generations to view. Give your image the respect it deserves. Shoot in RAW...

I was going to say, what debate? I convert to JPEG for printing snapshots (they only accept JPEG or TIFF at most labs) and Tiff for anything large.

My K10D is always set to RAW, I'd never consider changing that.

That was also my first thought!

My DS was always set to RAW and and my K10D will always be too (although I have switched to DNG instead of Pentax on that).

It's only a debate for those who can't handle the RAW workflow (for one reason or another... reasons which are diminishing daily!)

For those that use RAW, there is no debate and no question

(BTW, DNG is not necessary. Shoot PEF on the K10D. You can convert to DNG if you absolutely have to use Adobe's stand alone DNG converter. All the major RAW apps handle the K10D PEF format now.... it's smaller than the DNG )
Matt

Another reason for using RAW is that the workflow, using Silkypix, is so fast and efficient, whilst at the same time producing superb quality.

It (the paid version) is designed for fast batch processing. I haven't used any of the high-end Adobe products, but I find it difficult to believe that they do the job better than Silkypix.

Indeed, I believe PeteKD uses Silkypix for processing JPEGs!

G

Same work flow speed with Lightroom with no difficulties at all. I prefer the results from Lightroom (good thing too as I bought it).



At the DNG comment, shoot PEF on the K10 as stated an extra 60% photos is not something that should be ignored. Same quality as DNG, although if they went for DNG support I don't see why they didn't just use compressed DNG to start with.

Because all compressions take serious processor time, and not all compression is the same
Haven't looked at the maths of the DNG, but it probably exceeds the time requirements of the in camera CPU. The PEF compression uses a Huffman tree which ends up being pretty efficient. Canon do a double Huffman compression, and others use different methods. I reckon some of them use custom hardware to do the compression (probably Canon do in their Digic chips).
Matt

anybody have any experience with aperture?
I use it and am very happy.

George is right I use silkypix for my jpegs any PP I use PS but for WB,sharpening,Exposure, etc..etc.. I use silkypix I think it is first class.

I have not tried RAW yet in Silkypix or if PEF files work

Just a point the only problem with RAW is buffering speed if you are shooting lots of shots one after the other. I have a wedding shoot saturday and have been told shoot RAW of course however when trying to capture several shots at once using RAW concerns me having to wait.

Also sometimes it can take an eternity to view the image on the LCD screen while it loads as it is a big file. NOt good if you are in a rush and want to check the histogram quick etc..

Perhaps I am doing something wrong anyone care to tell me ??

Pete

You must have a really "pants" card Pete
I use Sandisk Extreme III and Lexar Pro SD cards (both 133x) and the K10D absolutely flies. Almost immediate replay of images (Under 1 second to view.. histogram shows in another second or so. Scrolling through images is faster than I can see...)
Matt

Shooting RAW you can still take 8 in continuous mode before it slows down.

If you're doing that at a wedding you need a video camera, not a K10D.

Aye, I have no issues at all viewing RAW and I can shoot off as many shots as I like within a reasonable time frame. Instant review, histogram check press shutter to go back and can immediately shoot again. If I know the image is fine or don't have time to check I can shoot straight away without moving the camera.

I have a 2gb card in my k10d, on startup with a formated card it tells me I can take approx:

DNG= 119.
PEF=120.

Not a great deal of difference there but it is an assumption on the camera's part.

however, there is a ~ 6Mb saving for PEF files per image taken. For two identical shots the DNG is 16.2 Mb and the PEF is 10.2 Mb

I've been using DNG as i was under the assumption that there was little or no compression used, whereas PEF has 'some' compression. I thought compression was to be avoided where possible for several reasons.

1. image degradation, even lossless compression would cause some loss or change of data captured.

2. performance issues, saving the raw data must be faster than compressing, THEN saving the data.

I'll do some more research on the PEF format.

The counter does not take into account compression of the PEF files. You can get 180 PEF files on a 2GB card.

1. Lossless compression is just that. There should be no changes.

2. Some time is spent on compression, but less time on writing to the card. I don't know which takes longer; The internal compression of data or the writing of an additional 6M to the card. It is possible that compressing and writing the PEF is faster than writing the DNG itself.
Certainly compressing to a jpeg and writing 1/8 the data is much faster than saving the full raw data.
I haven't done any tests, it would be interesting to hear how many PEFs can be taken compared to DNG before the buffer slows things down.