Martini Hour 052, In Which Colleen Compliments Our Guest’s Definition of a Martini

At last, a guest who knows a good martini when he is offered one. This week’s guest, Tom Hogarty, besides being discriminating beverage connoisseur, is also the Adobe Project Manager for Lightroom, Camera Raw and the DNG file format. We wanted to talk with Tom specifically about the Lightroom part of his job, because the newest version of the product is currently in public beta (from Adobe Labs), meaning it’s downloadable for you to check out yourself. Yes, you’re welcome.

Tom concisely calls Lightroom his digital assistant, that is, the device which handles all the necessary tasks for managing your digital photography. Then we remind him that it’s “Martini Hour” (even if it’s more like a half-hour) and he can be a little more effusive.

Here are some of Tom’s favorite things about the Lightroom 3 Beta that he graciously shares with us:

Lightroom 3 beta has improved image processing, not only for images you’re shooting now, but upgrades to the raw processor will actually allow you to return to previously processed raw files and get better results. Tom’s other big favorite is the new web connectivity which allows you to post your images to Flickr, and recieve feedback directly in the Lightroom interface. Yes, folks, digital image processing has gone “social.”

For his Toast of the Week (or TOTW as Deke spontaneously likes to call it much to my apparent confusion) Tom raises his triangular glass to the community of Lightroom users, including his MVCs, those most valuable contributors who are helping make Lightroom as great as it can be.

If you’re thinking about checking out the next version of Lightroom while it’s still cooking (and free) then this is the show for you. Here’s the regular-quality (128kbps) audio file. You can stream, or for best results, right-click and choose Download or Save. For you audiophiles out there, here’s the link to the high-quality (320kbps) version. Be sure to download, don’t stream. And don’t forget the usual suggestion to subscribe via iTunes.

Thanks very much to Tom for joining us. You can check out his blog, the Lightroom Journal and follow him on twitter at @LR_Tom. Cheers!

Next entry:Photoshop Top 40, Feature #9: Levels

Previous entry:Photoshop Top 40, Feature #10: Color Settings

  • Deke?  Help please!

    Help! My college neglected to install all CS4 fonts in January as requested, and now that I have started an InDesign CS4 course using Deke’s InDesign CS4 one-on-one as the textbook I have a big problem.

    Can you please identify for me all the fonts used by for the InDesign CS4 one-on-one book exercises? I need to have this ASAP so that a solution can be put in place by ITS at our college. (despite their error, they refuse to make a wholesale lab/network image change).

    Hope you have this info handy!  Thanks.

    Mike Laurence

    Collingwood, ON

  • the math behind ACR and bayer and high iso noise


    Since all camera sensors are really just light sensors and the RGB part has to be faked into the file by the color filters and the bayer equations - (you have mentioned this in prior martini hours) and since we therefore get “digital noise” in our high ISO shots from the random thermal noise in the sensor getting turned into RGorB pixels by the math not by rougue R G or Bs - and I assume Tom is refering to some smarter algorythms that help minmize this - this brings up a point. Can we get access to the luminance values only in ACR before the Bayer math? Can we create our own math through a special Deke trick?

    High ISO just boosts the gain and therefore amplifies the trigger events from thermal and gama ray and other subatomic particles hitting the sensor - just like too much gain on your microphones in martini hour when one of the guests speaks too low or you leave a mic in the other room… Of course many fine sensors are refrigerated to minmize this thermal noise (telescopes and larger format), but not our SLRs. I hesitate to open my Nikon up and epoxy on a thermal electric cooler…so ...

    I was thinking if we could get just the luminace data out of raw, then shoot 3 shots with R G and B filters create an alpha channnel from each and recreate RGB in photoshop (but the color stripes would still be in the camera- don’t want to pry that off my Nikon either) - or maybe even just extract only the R G and B channels from multiple rapid shots without using lens filters and use difference (or some other great deke secret handshake) to highlight the thermal noise pixels and remove or mask them and therefore greatly reduce noise (noise pixels would have high frequency and stand out in difference from shot to shot)

    ....And so we could create a new photomerge to reduce noise by comparing multiple shots of the same scene in high iso looking for and removing the random high luminance pixels - what a stack in bridge and action in photoshop that could be my Deke man! Lets work it out then you could call Tranberry for a custom panel (aka palette). Oh - and Russell could even apply it to video because the multiple shots would just be the adjacent frames

    Sorry Colleen -

    uh…if this is too much perhaps deke and i were part of the same late 50s DARPA gene experiment “Corpus Callosum excessum” where mothers were secretly given early fetal viruses carring the expermental gene - Deke did your mom have a mysterious cold in early pregnancy? could explain alot…


  • Installing Camera Raw

    I had a stroke so go SLOW and CLEAR!

    I installed the upgrade to CS/4 and the interface between lightroom and photoshop still seems to have 4.6 interface. HELP!

  • Hmm, that’s strange

    Try going to Help menu and choosing Updates. It’s time-consuming and requires you to quit all Adobe apps, but it should correct the Camera Raw problem.

    If not, write back and let me know.

  • All fonts are listed in Figure 5

    On page xvi of the book, Adobe InDesign CS4 One-on-One.

  • Well, there are a couple of problems here

    First, the red, green, and blue resin is baked onto the chip. There’s no way to turn it on and off. Which means two things: 1) An unfiltered back-and-white shot would vary from one pixel to the next b/c the luminance of pixel A1 would be red only, that of A2 and B1 green only, and B2 blue only. 2) If you slapped on a physical red, green, or blue lens filter, you would cancel the otherwise filtered pixels. In my example, a red filter would leave pixels A2, B1, and B2 black.

    Some color noise is a function of the demosaicing that resolves the filtered pixels into RGB. But most of the noise Tom and I are discussing is a function of the gamma boost required to convert the linear luminance information into an image that looks right to our eyes. This stretches the shadows and compresses the highlights. So shadow noise grows more apparent and highlight noise less.

    Next versions of Lightroom and Camera Raw process detail information differently and offer more controls in the hopes of quelling some of the noise without sacrificing detail. But you can take heart in two things: Noise is far more obvious at 100% on screen than it is in print. (In fact, as a rule, downsampling is about the best noise-reduction solution there is.) And a great image looks great noisy or not.

    Noise is to a digital image what brushstrokes are to paintings—fundamental and sometimes welcome elements of the medium.

    (You were so cheeky, and there I go being so serious . . . in a Martini Hour comment, no less.)

  • Thanks!

    oh dear, and i was trying so cheeky hard to make you laugh, my humble apologies.


    I am serious though, about using merge to try to average away those pesky sparkles in the shadows, i just am not sure what method would work best…


    BTW…sorry I am so old i remember when the mosaic was part of the IR cut filter and we could remove it from the sensor and make super night vision from cheap sensors - forgot it is baked on now and guess that cooked my goose in the last post (- i’ll try puns now) - i think you could see the pixels then too with mild magnification, but it was a very dark room in that optics lab and they did not let me out much

  • more simply put.. sort of

    ok if that was too complicated - more simply put -can we use photomerge and some Deke tricks to average away high ISO noise from repeated shots of the same subject in rapid time sequence?

    adding a bit more techy background…

    a couple of white papers from Bruce Fraser on the math and sensors and the fact that different math has different effects on noise and how the noise causes problems for the math:

    makes me wish for 3 sensors and 3 separate filters…hmm like a high end video camera… the noise would not occur on the 3 sensors at the same pixel at the same time and could be statistically removed

    - can ACR 5.x read the 3 separate sensors of a high end video frame 10 bit 4:4:4 rgb?

    oh well since the video resolution lags behind the stills I guess Deke’s downsampling still wins as best immediate approach

Be the first to drop some wisdom...