RAW vs JPG, searching for a thread.

<p>Last year (?) there was a thread on photo.net about RAW vs. JPG. Actually there are quite a few threads about this subject but this one had an elaborate discussion about the quality of JPGs, stating among others (if memory serves me well) that JPGs are 8 bit but their bits are distributed smarter than the 12/14 bits of a RAW image where the bits are distributed linearly. And more information'.<br /> Does anyone recall this thread and it's URL? I tried the search engine but the terms RAW and JPG are too general and yield too many results. Most are about what people use and not about the differences between these formats. BTW, I shoot in RAW almost 100% of the time but I'm also interested in the characteristics of other file formats. No need to start a new discussion because all arguments are laid out already - they're just hard to find.</p>

 

<blockquote>

<p>JPGs are 8 bit but their bits are distributed smarter than the 12/14 bits of a RAW image where the bits are distributed linearly.</p>

</blockquote>

<p>JPEG's are 8-bit per color, gamma corrected (not linear) and have been '<em>baked</em>' by the camera to produce a color appearance. Raw data is high bit (more than 8-bits per color, 10-12 and maybe in rare cases real 16-bits per color), it's linear encoded (gamma 1.0) and not processed to appear anything like an image you'd use. Unbaked (see: http://www.digitaldog.net/files/raw.jpg).<br>

Not sure exactly what article you're referring to but this one covers the concepts quite well:<br>

http://www.lumita.com/site_media/work/whitepapers/files/pscs3_rendering_image.pdf</p>

<p>Is it this one http://www.photo.net/beginner-photography-questions-forum/00cSWc</p>

<p>@dan: thanks, but alas, that's not the one I was thinking about. I'm looking for one from the beginninng of 2014 or older, which puts JPGs in perspective, not just the 8 vs. 12/14 bits debate but more positive about JPG.</p>

<p>I'ld like to read that thread as well especially on how they explain "smarter" distribution of internal Raw sensor data cooked to 8 bit jpgs.</p>

<p>As an aside I'm also curious if incamera processing of its own Raw sensor data applies their own Clarity adjustment or whether they use their own custom algorithms tuned to whatever the lens data communicates to achieve definition and clarity.</p>

<p>This from seeing posted unedited jpeg samples from newer cameras that seem to produce renderings with better clarity over my own 2006 DSLR jpegs.</p>

<p> </p>

<p>Stick with RAW! A picture that is developed in camera as a JPG may have blown highlights or just too dark. Repairs on a JPG area a lot more difficult. If the file is in RAW, any raw developer will allow you to recover highlights and open up the blacks. It is also a snap to get rid of color aberration before doing a finished output--as a TIFF, which is lossless.</p>

<p>Mathew, I believe that's not what the OP is asking about.</p>

<p>Hi Jon - Could not find the particular post to which you referred. Here is a link to a post I made on the topic a few weeks ago. http://www.fredmiranda.com/forum/topic/1293749/1#12346921<br>

Not sure that covers what you needed yet thought it was worth mentioning.</p>

<blockquote>

<p>JPGs are 8 bit but their bits are distributed smarter than the 12/14 bits of a RAW image where the bits are distributed linearly</p>

</blockquote>

<p>JPEG is gamma corrected (for example, a gamma or TRC of 2.2) and <strong>only</strong> 8-bits per color. Raw data is at least 10-bits per color and linearly encoded (1.0 gamma). There's nothing smarter about the JPEG, it's just different data. Baked, less bit depth and smaller potential color gamut. </p>

<p>Yes, but is anyone here noticing straight out of the camera jpegs having more snap and clarity than the jpegs produced by much older cameras?</p>

<p>Don't these camera manufacturers notice all the before and after/jpeg vs Raw work posted online and think it might do them good competitively to improve their own internal jpeg rendering algorithms? </p>

To the OP: there's nothing in a JPG that isn't in the RAW. The RAW stores all the data that is used to create the JPG. That said, a

camera's JPG engine may create a better JPG than the user using a RAW converter is able to. That says nothing about the formats.

 

Andrew, Matthew, do you know of good resources explaining what one can do with a raw file, and how? I know of no single, easy to read

resource.

<p>Thanks for your responses. The thread I'm looking for is not mentioned so I guess I'll have to keep looking out and reading :-)<br>

I'm aware of the main differences between these formats and I shoot RAW in the majority of cases since a long time; that's not the question.</p>

Hi Jos,

 

That's ok, but in that case what is the question? Since you seemed to be saying JPG could do something that RAW

couldn't.

Maybe you'll not need that old thread, Jos. The only positive thing about JPG's is that they can pack a large image in a small file (but don't ask how well, because that would no longer be a positive thing).

<blockquote>

<p>The thread I'm looking for is not mentioned so I guess I'll have to keep looking out and reading :-)</p>

</blockquote>

<p>What specific information are you looking for? What is the question? </p>

<p>Mostly - like in a few of the above threads - people mention the 8 vs 14 bit difference, more or less suggesting that JPGs offer about half the quality of the RAW image. One of the replies in the thread I'm looking for mentioned something about the distribution of the information in the JPG, it should be more intelligent that RAW in that respect. That's the info I'm looking for, mainly out of curiosity.<br>

Of course all information be it JPG or something else is derived from the RAW image, so nothing adds information to the RAW, it's all less info, to a certain degree.<br>

In my question I mentioned "and more information", perhaps this was confusing, I meant that there are more differences between RAW and JPG. I believe it had something to do with the linearity of the information. RAWs are linear, leading to the old expose to the right practice, maybe JPGs follow a different way of storing the data - maybe not. That's basically the information I'm after.<br>

In the old threat someone was enthousiastically defending the JPG format, seemingly well motivated. Anyhow, it won't change my shooting. I shoot RAW until I have only one card left, then I may switch to JPG only if there is no other way to keep shooting :-)</p>

<p> </p>

<blockquote>

<p>like in a few of the above threads - people mention the 8 vs 14 bit difference, more or less suggesting that JPGs offer about half the quality of the RAW image.</p>

</blockquote>

<p>It doesn't work that way but yes, more bits can mean more data to edit such that at some point, you reduce or eliminate banding in the image due to insufficient bits. That's true of high bit non raw files too. This might be a good start in terms of what high bit data, raw or otherwise bring to the party:<br>

http://www.digitalphotopro.com/gear/imaging-tech/the-bit-depth-decision.html</p>

<blockquote>

<p>One of the replies in the thread I'm looking for mentioned something about the distribution of the information in the JPG, it should be more intelligent that RAW in that respect. That's the info I'm looking for, mainly out of curiosity.</p>

</blockquote>

<p>That doesn't make much sense however. The JPEG is processed by the camera and how the camera maker thinks you'd like the image to appear. Just like we saw with transparency film. Some loved Velvia, some loved Ektachrome film. They rendered the scene differently. Is that more intelligent than raw data which expects the photographer to do the rendering? The URL above I posted addresses that topic. <br>

Raw is raw, unrendered and it doesn't look as you wish it to appear, just like a color negative. Is a transparency more intelligent than a color neg? It depends on how you look at it and what your final goal is. Then there's the difference in the amount of data and gamut potential. </p>

<blockquote>

<p>In the old threat someone was enthousiastically defending the JPG format, seemingly well motivated.</p>

</blockquote>

<p>The first question is, do you want the camera to render the scene or do you want to render the scene? That URL covers that topic. Once you make the decision, you'll pick one or the other (or both I guess). If you prefer to have the camera render the scene, then it's quite easy to defend the JPEG. I have no problem with that but like you, I'm always going to capture raw data. </p>

<p>I've been asking myself the same question re: JPG, RAW, TIFF etc, and Resolution and Bit Depth. This link explained it well ...<br>

http://www.graphics.com/article-old/digital-photography-fundamentals-understanding-resolution-and-bit-depth</p>

<p>Ken</p>

<p>if you can't find it searching, just go backwards in the archive. January 2014 wasn't that long ago...</p>

<p>my curiosity on this topic was ended after I read Bruce Fraser's (rip) Camera Raw book 10 years ago, mine is old copy.</p>

<p>I don't always print or save articles I want to refer to later, but if I do I print to pdf...</p>

<p>I'm not sure what you are looking for, but maybe it would be easier to consult the wikipedia page on JPG? (I haven't, so I don't know what's there, but it may be what you're looking for.)<br>

JPG doesn't store pixel data as such. It divides an image into blocks (8x8 pixels, I think), and then stores a mathematical function (a Fourier transform, I think) that will approximate the contents of the block. The higher the approximation, the more complex the function will be, so the more space it will take. It's indeed a smart approach, but not maybe in the way you were thinking. The fact that it's always an approximation also means that it is never lossless, even if it takes up more space than the original, so lossless JPG must use something else (maybe an approximation plus the correction data?). But at the end of the day, there is nothing you can express via JPG that you can't express via RAW - it's not a matter of the JPG being 'made from' the data in the RAW, it's a matter of it not offering anything else even if there were something else.<br>

By contrast, formats such as PNG, BMP, etc, alll are ways of encoding the pixels. JPG is not - it is an encoding of the image, which when decoded may or may not correspond to the original pixels.</p>

<p>Here area few reasonable easy reads for what you may seek (same web site):<br>

Gamma encoding that you mentioned: http://www.cambridgeincolour.com/tutorials/gamma-correction.htm<br>

Bit Depth: http://www.cambridgeincolour.com/tutorials/bit-depth.htm<br>

Posterization (from low bit depth): http://www.cambridgeincolour.com/tutorials/posterization.htm<br>

Raw file format: http://www.cambridgeincolour.com/tutorials/RAW-file-format.htm<br>

Tiff and JPEG file format: http://www.cambridgeincolour.com/tutorials/imagetypes.htm<br>

Each has a good number of embedded links for further details.</p>

<p>John</p>

<p>Excellent - thanks for posting those links. I've bookmarked all</p>

<p>You're welcome Ken.</p>