Chuck:
I knew that JPG was lossy, and should have thought more about how the
encoding works to "lose" the data. Thanks for the explanation. Since
these negs were monochrome, I assume the same is true for
grayscale/shades as well as B#1, B#2, etc.
Once I received the order back and noticed the output was JPG, I thought
"Hmmm. Maybe I should have asked for TIFF." Oh well.
I understand how this sort of stuff (poor handling by a commercial
outfit) drives one to digital. But after seeing my Kodachromes...
Earl
Chuck Norcutt wrote:
>Earl Dunbar wrote:
>
>
>Sorry to hear about the scratches. It's part of what drives me to
>digital. I haven't processed my own B&W in many years but could if I
>wanted to. But I really can't do color.
>
>
>
>>1. What accounts for the variable file sizes of the XP2 scans? I mean,
>>every frame is a full 35mm frame. Is this related to overall negative
>>density, or with scanning technique or settings?
>>
>>
>
>The variable size is just the nature of the JPEG image format. JPEG is
>a "compressed" format and, further, is a "lossy" format. The actual
>processing is quite complex but a simplified example would be something
>like this:
>
>Across the 3000 pixel width of a single row of blue sky pixels let us
>say that 2995 are all the identical color of blue which, for simplicity,
>I will designate B#1. Randomly scattered across that line are 5 pixels
>that are different which I'll call B#2. A simple compression algorigthm
>is called run length encoding. So the software records this line of
>pixels as something like (1000, B#1), (2, B#2), (800, B#1, (1, B#2),
>(600, B#1), (2, B#2), (595, B#1).
>
>As you can see, the representation of the data and the space required to
>store it is significantly reduced from simpy repeating B#1, B#1, B#1,
>etc, etc, 3000 times. You can also see that, if the pixels in a given
>area of the image are of very uniform color and brightness that the data
>can be encoded in a small space. If there is a lot of variation the
>space required will be much larger.
>
>This type of space reducing data encoding is common to all compression
>algorithms. The final twist in "lossy" algorithms such as JPEG is that
>the software might determine (based on the JPEG "quality" level) that
>B#1 and B#2 are really so close in color and brightness that no one will
>ever be able to detect that it just ignored the difference and stored
>the entire line of pixels encoded as the very short line (3000, B#1).
> From this you can also see that as the quality level is reduced (the
>"lossy" part) the required storage space is also reduced.
>
>
>
>
>>2. I assume that since I got JPGs, and from a non-custom service, I
>>really can't judge the dynamic range of the files as being best
>>possible. One some of the shots with wide tonal range, the highlights
>>are blown. Even on the negs they look like they may be blocked under a
>>low power loupe, but I can't be totally sure. Any insight?
>>
>>
>
>If the highlights are blown on the JPEG file they're probably blown on
>the negative as well but that's certainly not a given. Depends on the
>quality of the scanning. The JPEG file is not capable of reproducing
>the full dynamic range of the film but the values should be adjusted so
>as to preserve the highlight and shadow detail as best as possible
>within the range of brightness values allowed by the JPEG data. In
>short, you shouldn't have blown hightlights.
>
>
>
>>3. If I find that the highlights are actually blown on the negative,
>>might this be an indication of exposure error? Does anyone have
>>experience of XP2 doing better when rated higher than 400, which is what
>>I used?
>>
>>
>
>That's certainly how I would interpret it.
>
>Chuck Norcutt
>
>
>>TIA, Earl
>>Now practicing the zone system, as in "no political sniping zone"
>>
>>
>>==============================================
>>List usage info: http://www.zuikoholic.com
>>List nannies: olympusadmin@xxxxxxxxxx
>>==============================================
>>
>>
>>
>
>
>==============================================
>List usage info: http://www.zuikoholic.com
>List nannies: olympusadmin@xxxxxxxxxx
>==============================================
>
>
>
>
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|