conrad

[conrad]

Future of FCP X and my hopes for Premiere CS5.5

So back in Novemeber when I wrote up the long post predicting what has pretty much happened with the release of FCP X and discontinuation of Final Cut Server, I actually wasn't that confident everything would pan out the way it is.  Yup, I was wow'd a little by the FCP X demo video at the LA User Group meetup.  But then I shook my head, looked a little closer at the screenshots, and asked myself, "Ok, besides looking pretty, what is FCP X?"  I had a lot of questions re: the limitations of background rendering, asset management, and the usability of the entire system.  It looked nice, but so did Motion when it was first demoed.  Touching Motion for 5-minutes gives you deep insight into the Apple philosophy regarding pro apps, and imho, a lot of critical things they do not understand about professional users.  Well, now that I have played with FCP X, all that I feared about background rendering, asset management and usability have come true.  Now what?  To be honest, Apple isn't dumb.  In a few months, they'll release patches for free to include Multicam, some limited import of old FCP projects, EDL support, etc.  By does that mean all is well?  No.  Here are my reasons why:

Apple actually has a very poor track record for pro app updates.  To get a 0.1 update takes a while.  First, compare Apple's pro-photo tool, Aperture to Lightroom.  I don't want to get into a feature war between the two, but people will have to admit that Adobe is much faster with updates and do so more frequently.  When a new camera comes out, count on Adobe to get support out the door ASAP across all its product lines.  So, yes, maybe Apple will have the multicam, etc support in like 2 months (they better!).  But what about after?  An update every 6 months?  Is that enough?  To be honest, in a pro-level environment, I expect significant updates every 2-3 months.  How can you not if you're working daily with your customers and you're iterating based on feedback from them?  <grin>

Even if FCP reaches feature parity to FCP7, we still have the issue of media management.  I fundamentally believe moving to the Project/Event folder model was a wrong move.  I'm a user experience expert, and I would agree with Apple that people think in terms of events.  But do they organize in terms of events?  Here is where I think they interpreted the UX data wrong because professionals do not.  They'll organize by project, or organize by keywords, but not by event.  Whether I shot it Thursday or Friday doesn't matter 5 months down the line.  And so I drag in a bunch of clips, and FCP X separateds it out by day (and I have like one video per day).  So I have this laundry list of video items to scroll through.  Yuck!

I think fundamental to their project/event folder is the inability to transfer or copy what you're working on to someone else.  If you really think deeply about it, your footage is spread across a bunch of events.  Your projects and all the behind-the-scene render files are stored somewhere else.  How can you give that to someone?  Perhaps you can consolidate it and give it to someone else.  Then what do they do?  They import it and it gets dispersed again on their system's project/event folders.  Case in point: how does an iPhoto user transfer 50 albums to you, while preserving meta data and slideshow projects, etc?  Not easily, if not impossible.  As iPhoto users know, they're locked into your own computer.  I'm afraid FCP X adoption of the iPhoto/iMovie media management philosophy does just that, lock you in.  (Btw, I know there's the "Move" project/reference events to another drive dialog thing.  Try to use it.)  To be honest, I'm not sure how Apple can fix this easily.  They're essentially built a baby Digital Asset Management system into FCP X.  But unless your DAM is feature rich and awesome (which they couldn't do right even in Final Cut Server), it's useless and at best, frustrating and wreaks havoc on collaboration. DAMs have to be done right or they create more headache then they're worth.

I'm gonna go out a limb regarding background rendering because I've only played with it here and there. It's interesting to see how Adobe Premiere CS5.5 and FCP X handle your typical H.264 file.  For Adobe, they place all their eggs in the Mercury Playback engine.  It's a playback engine.  It never touches or transcodes your H.264 until you're ready to output.  They've done their due diligence to make sure as many filters and streams can happen while playing back in real-time.  FCP X borrows some of that playback love, and renders ProRes behind-the-scenes.  That sounds good in theory.  But what happens when you just dragged in 500GB of H.264 files into your project?  Background copying, analysis, rendering sequentially kick in.  And it takes forever.  And the horrible part is constantly that popup you get, "Sorry you can't do that while there's background jobs running, bleh!"  I find myself constantly going in there and pausing/canceling background jobs.  Copying btw is dirt slow.  Rendering is even slower. Most editors already did the ingest process, yet FCP X insists on doing it again.  Sigh.  I prefer Adobe's approach ... get the footage in there, give me the tech to work with the footage without rendering, and I'm okay with a little stutter on 5-stream heavy effects shots.  Then you finish your project, hit render, and celebrate, grab a coffee, and come back in 5 minutes (CUDA rendering is blazing!)

The effects management in FCP X is something I will never get used to.  It's the UI, the feel of it, to be honest.  It feels like Motion.  Feels clunky.  I want it to feel like After Effects or Shake (I hoped the new Motion would have been a re-skinned Shake ...sigh).

Outputting on FCP X is like iMovie but worse.  You can't even export a down-res'd, low data rate Quicktime without resorting to Compressor (and I'm not going to pay $50 for that!)

Now I have a few things I wish Adobe would do (hopefully Todd Kopriva is listening).

Adobe Dynamic Link should be just available.  Don't force people to buy the whole suite to get it.  There are cases where they only want After Effects and Premiere together, and they have them as separate licenses, but they don't have the interconnectivity between unless they're from the same suite.

The monthly subscription price is a little hefty.  The math just doesn't make sense to me.  I know for many Adobe users like myself, I skip upgrades.  I never went to CS4 because the benefits were not that great to warrant the expense.  CS5 however, deserved it.  I think the subscription model should be priced assuming that kinda of behavior is the norm.  Basically, slash the subscription price in half.

I really wish Adobe would offer competitive cross-over pricing.  I think there's a ripe opportunity for Adobe to welcome a whole bunch of unhappy FCP users.  This one facility manager posted on Creative Cow how he can't expand his facility because Apple discontinued FCP7.  To be honest, the $299 price for FCP X was irresistible for pro-users.   With compressor, it was more like $350.  I think if Adobe offered Premiere Pro CS5.5 for that price to existing FCP users, it would do much to gain entry in the video editing pro-level market that was once dominated by Apple.  Come on, where's the "Switch to Premiere CS 5.5" campaign?

I wish Premiere CS5.5 could refresh some of it's layout UI.  The myriad of in/out type buttons below the viewer/canvas is just too much.  FCP X does look slick because of it's minimal design, and I think Premiere could use a little layout clean up to make it a more space efficient.

[conrad]

30-minutes with Final Cut Pro X

30-minutes is not a lot of time with software.  I edited a few things in FCP X to get a hang of it, and here are my findings:

It's the big question on ppl's mind.  Yes, it feels like a glorified iMovie.  I wish it weren't so, but a lot of it has that feel.  When I started using it, it was the exact same feeling I had going from After Effects to Motion.  Yes, technically Motion has a lot of features of AE, but there's the  subtle cues that what you're using doesn't have a pro edge to it and that is what FCP X exhibits.

The interface is clean, though I don't consider it intuitive.  Apple really believes in showing you what you need to see, so often things are hidden away.  Hiding the Project window is awkward.  Do not think going from FCP7 to X will not have a learning curve.  Actually, it's a pretty dramatic shift in things, especially in media management.

It feels very slow.  Primarily due to the background render management.  When I first heard of it, BG rendering sounded nice.  K, I admit, I was getting jealous.  Now, I don't want it.  It's too slow, and hinders the zippiness that I need.  As a reference point, I have a 17" MBP 2.66Ghz Core i7, 7200RPM HD, 8GB RAM.  Everything feels sluggish.  I tried it on a 16-core Mac Pro and it was better.  But it was still gonna take hours to transcode all my H264 to ProRes. Sorta like Motion if you've experienced that before.  If you look what's going on in the background, FCP X is doing ingest, analysis, and rendering simultaneously via the new multiprocessing Grand Dispatch tech.  That's way too much imho and my laptop was on the verge of setting aflame.  Maybe it would be smooth on the 16-core Mac Pro, but not on my teeny quad-core MBP.

No Media Manager .. so Im not sure how we'll archive projects.  Project consolidation feature wasn't very helpful here.

Can't open previous FCP projects.  Rumor is they're working on a utility of sorts, but I don't expect it to be perfect.  How can it be?  If it was easy to do, they would have done it.  But because it's not in FCP X, it's therefore hard to do.  Therefore when they do release it, it won't be perfect.  Things that are hard to make never produce perfect results; they're not supposed to.

Media Metadata Management is kinda nice.  Auto balancing is good, as is the face detection.  But the Events thing is horrendous.  Forces you to use iPhoto-like thinking when managing assets.  I know it's a good way to think, but in all honesty, the way we work on projects at church doesn't really fit this model.

Bad! importing footage is copied to your local drive!  I don't know how FCP X can work effectively in a networked, shared environment.  I was watching how it was doing all the symlinking of footage in the background, and honestly, it felt very brittle.   You can disable this, but having a local copy of FCP handle organization-wide digital asset management is kinda scary.  Until I look into this more, this is the major deal breaker for us as we need a very seamless, reliable integration to our server assets.  The same above goes for organizing Projects (it's all inherently local, like an iMovie project).  I know you can disable the copy, but now the project/events/assets are beginning to look quite dispersed.  I think the major mistake of FCP X is insisting on using their own folder structure akin to iPhoto.  They should have taken tips from Adobe Lightroom in this regarad, which does a great job of handling metadata without touching your folders.

Effects handling was clunky.  Keyframing is not easy to use, and hard to edit (they use the timeline itself to show you the keyframes, which I like, but it's tiny!)   A lot of the effects like ken burns and stabilization is very nice.  A lot of good out of the box effects are available.  In the past, I did a lot of hackery to get some decent motion graphic effects in FCP7 without having to go to After Effects.  I think you'll have a much harder job doing that now in FCPX.  There are a lot of nice effect in FCP X, period.  However, using them felt like iMovie, rather than the power of After Effects.    It's as I feared.  They reduced options to make it look easy, but in the end, it takes control away from the editor.

Titling is pretty nice.  Nothing to call home about; it was about time.

Magnetic Timeline was annoying.  I think it's good for the person who doesn't understand clip handles and blank space, but one move, and the whole thing starts moving. Like trying to create some blank space between clips to insert a new clip is quite a challenge.  There are no slugs also :(  Requires me to insert a gap?  Really?

It couldn't import my AF100 AVCHD footage that I copied on my HD.  FCP 7 was able to.  The directory structure was intact.  Dunno why ... boo!

Color Correction... I really thought FCP X would shine here.  They have masking tools and grade saving, just like in Color.  But the color board UI is just hard to use and is confusing.  No control of numerical values either, so it's all done by hand.  I think ppl will have trouble using this... and as much as I hated the 3-way color corrector, I'm hating this FCP X's color table more.  Match Color is pretty amazing btw.

No 3rd party plugin support.  Sorry, seems like we'll have to re-buy some of our plugins...argh!

No multi-cam support.  Let's hope they add this in before our next big show.

Tape support is bad.  I know we're past tape, but capturing a lot of clips of an old DV tape is gonna be very tedious.  No more log and capture folks... it's all "Capture Now" button, ala FCP Express.  Sigh.

When exporting there are no options to resize your video, etc.  Suppose your project is 1080p, there's no way to set your own data rates or down res it, without sending to Compressor (which is another issue altogether imho).  When considering price, expect to spend $350, not $299.  You have to buy Compressor to finish your footage out of FCP X.  To be honest, I'm fine in a way (as this was the old Export via Quicktime Conversion).  But I wish they didn't force you to buy Compressor to get any decent settings.  Compressor just isn't easy to use.  Workaround: export as 1080p ProRes (the default) and bring into QuickTime 7 to convert it.  Ah, old school!

Final Cut Pro X crashed a few times on me.  Launching FCP X made all the videos in Keynote go black, so be sure to install the Pro Apps update.  Just feels a little finicky.  As expected in a 1.0 of course

Motion still sucks.  They just don't understand compositing and motion graphics.  Only benefits seems they are moving to become a motion graphic publishing platform, where ppl can share their work.  If community contributions start to pop up, this can be interesting!

Compressor still looks buggy, and now it's gonna try to do auto-qmaster.. uh oh.  I'm looking to move into Adobe Media Encoder, now as they added watch folders to it.  If nothing fundamentally changed about compressor, it's time to look for something else imho.

No DVD Studio Pro.  Bye bye, DVD mastering.  All digital distribution is the new way (not sure how subtitling would fit into this).

[conrad]

PunyPNG hits the 10k mark

PunyPNG finally surpassed it's 10,000th user.  I guess not such a big milestone for most web services out there.  Two years later, almost 2 million puny images later, on a shoe-string budget, it kinda hit me there are that many developer/designers out there that rely on PunyPNG in their daily workflows.  That's awesome!  Going into the future, we'll be moving to a HTML5 backend soon and offer drag-and-drop support, so stayed tuned.

[conrad]

I'm through with Final Cut Pro

I never though I'd say this, but I'm finally through with FCP and I'm moving over to Adobe Premiere CS5.  Farewell, Final Cut Pro, it's been nice knowing you.  So why am I switching?  I was really skeptical at first, but I'm now convinced that Premiere gives the best results and fastest editing for prosumer and independent editors.

Here are my reasons for switching:

Apple recently announced they are discontinuing the XServe, which means Final Cut Server won't be seeing any major upgrades.  It signals to me that Apple isn't that invested in Pro Apps any more since they will no longer have server-level architecture to support the Pro Apps (like a Final Cut Server or QMaster nodes).  Which means, bye bye render farms and video archiving.

There were rumors have it that the Pro Apps team at Apple were given a directive to focus on bringing more newbie users from iMovie into the Pro Apps, and cater less to the power users.  Basically, all the cool features are coming into iMovie first (like in iLife '11), and then later onto FCP.  If you look at the last FCP7 release, it was a lame release with virtually no really improvements and I never felt any real benefit from it.  iChat integration?  Motion has fundamentally been flawed.  Color 1.5 is as hard to use as Color 1.0.  Soundtrack crashes the same when round-tripping with FCP.  Apple knows the money is in iLife apps not Pro Apps.  Look at the latest iMovie '11, with it's fancy new rolling shutter fix for DSLRs, which was missing in Final Cut Pro 7.  So, iMovie will be in the lead for product innovation?  Please.

Premiere's rendering/playback engine is superb.  I can have 7 layers of text, animated blur, blending modes, transparency composited over moving video.  You preview close to 20fps with no rendering (that in just software, without leveraging the NVIDIA GPU CUDA).   If you do want to render, rendering is multiprocessor and 64-bit, same goes for exporting.

Premiere's text tool is far better than FCPs.  You can have multiple text per title, with fine control of kerning.  Font preview actually works and wrapping is all done for you.  You can preview titles over the video, and create templates.

Color Correction tools are better.  Though it's not as good as Color, the UI however is more accessible.  FCP's 3-way color corrector felt clunky. Moving the controls on the color wheels always felt imprecise.

Premiere CS5's UI is great.  Just being to adjust UI brightness is a plus as I need to work in dim environments when editing to maintain color accuracy.  Values are precisely set similar to the After Effects UI.  Setting and moving keyframes is like After Effects.  Keyframe editor that actually feels usable and unconstrained to a little palette window.  Coming from a UX background, CS5 (and I would guess CS4 as well) inherits the UI investments that Adobe has made across all their Creative Suite.  Compare to FCP, where none of the applications in the Final Cut Studio look, feel or work the same.

FCP import/export will allow decent interoperability.

This is a big one.  Premiere CS5 allows editing of DSLR footage at various formats without having to use ProRes.  On FCP, we convert to ProRes in order to make editing easier since applying effects to H264 footage means tons of rendering in FCP.   Well, because of Premiere's awesome playback engine, you can use H264 native footage from DSLRs without having to convert ahead of time.  This means we save like 4 times the hard drive space and time!  Plus, this is nerdy, but when converting H264 footage to ProRes there are huge gamma bugs and your almost always ends up dark and saturated.  You get none of that in Premiere.

Premiere is cross-platform, so we can have more editors and edit stations.

There are cons to Adobe Premiere CS5 as well:

There's about a second or 2 delay when starting a playback (I assume it's caching).  JKL works but not as responsive as in FCP.  I'm willing to pay for this to avoid rendering and I heard is solved when you have a more powerful graphics card.

Premiere is expensive at $800.  Final Cut Studio is $1000.  Adobe Production Premium which includes Premiere, Photoshop, After Effects, etc is $1,700.  Hopefully student/nonprofit discounts can help here.

There are some silly bugs on the timeline when dragging clips in and targeting the proper track.  For example, when unlinking a video from it's audio, you have to deselect and re-select it for the link to actually break so you can roll edit one without affecting the other.  Razor tool hotkey is funky, requiring you to enable tracks that you want to cut.  You're forced to use the razor tool (c) to do any cutting, and I hate lifting my hands off the keyboard when I edit.  Perhaps I'm a newbie, but it does have some annoying hiccups and there's less hotkeys to selecting edit cuts, extracting empty space before a clip (hotkey 'x' in FCP), etc.

I've read some reviews and comparisons between Adobe Premiere CS5 and Final Cut Pro, but they tend to be pretty superficial.  The Adobe Premiere users tend to be PC users that have a gripe against Apple in general, and the level of editing expertise is a little underwhelming from the points they raise against FCP.  Well, as a Mac users and admittedly an Apple fanboy, I'm finally leaving Apple pro apps.  First, it was my decision to use Adobe Lightroom over Apple Aperture,which I never regret doing.  Now, it's FCP.  I love OS X and it's the best platform to code, edit, design, and manage files.  I just can't say I have the same love for Apple's pro apps.  They need to learn a few lessons from Adobe here.

[conrad]

Disappointed with the new iPod Nano

I think there comes a point where size of a product does somewhat dictate price.  $150 for something that small seems too expensive.  I don't recall ever owning anything that small yet that expensive.

[conrad]

What Makes a Good Creative Director

Just ate some humble pie this morning reading this article, What Makes a Good Creative Director from Design Taxi.  Honestly, when I took up the design director role at Ask.com, I'm not sure I knew what I was doing.  It's nice to finally read something to tell me what I'm doing right and what I'm doing wrong.  At least, all the things I am doing right, I owe to either the mentorship afforded to me from my old boss, Dahveed, or just cut my teeth learning from painful mistakes I made in the past when leading designers. Anyways, it's a good read for anyone aspiring to be a design lead, and better, what you should be look from your design lead.

[conrad]

PunyPNG on a New Server

I finally found some time to move PunyPNG to its own dedicated slice (previous it lived on Gracepoint After Five's slice). You can now go directly to PunyPNG via this new URL: http://www.punypng.com (http://www.gracepointafterfive.com/punypng will now be redirected to the new URL, but any API calls will not be redirected)

[conrad]

Alltop is Psychic

I heard about Alltop on Smashing Magazine so I decided to check it out. This is what I saw on it's homepage:

Smashing Magazine

Strobist, DPreview, Joe McNally's Blog, Digital Photography School

Mac Rumors, Mac World

Lifehacker, Seth Godin, Guy Kawasaki

Wow!  How did Alltop anticipate all of my interests?  It was like it knew what all my subscribed RSS feeds were.  Perhaps some new fancy Web tech where they got into my Google Reader feeds or something.  So I fired up another browser, cleared the cookies and went back to Alltop, only to see the exact same feeds! And so, it hit me.  Alltop is psychic. That or perhaps all who are part of the my middle-age web techie demographic are all predictably into the exact same things. Sigh. Anyways, check out Alltop, it's pretty cool.

[conrad]

Shooting Video on the Canon 7D

[vimeo]http://www.vimeo.com/10584736[/vimeo] Recently, our church purchased a sweet Canon EOS 7D for the express purpose of shooting video.  It's an awesome camera, but most of us didn't know how to use it.  After spending a few months with this camera, I decided to record this tutorial on using the 7D to shoot video.  Though having photography fundamentals is really useful, the tutorial just assumes you've shot on a consumer DV camera in the past and introduces the basics and not-so basics of the 7D -- all from the perspective of shooting video. Now, this video is really, really rough.  I threw my own personal 7D on a tripod and hit record.  My son was just born when I made this recording so I know I look like a zombie in this footage.  Focus on the camera, not me ;) Tutorial Chapters:

Introduction (11 min)

Lens Overview (19 min)

More Lenses (14 min)

Camera / ISO / Exposure (15 min)

Exposure Triangle (20 min)

White Balance (8 min)

Presets (7 min)

Production Techniques (12 min)

More Production Techniques (8 min)

Tutorial Notes (PDF, 2MB) You can also watch see the entire tutorial as a playlist album.

[conrad]

New Features in PunyPNG

We just pushed out some hot new features to PunyPNG today:

Added option to preserve EXIF data, to maintain copyright and other image metadata

Added option to skip bit reduction when compressing.  This improved IE6 compatibility (see PunyPNG's IE6 support)

REST-based API support (beta)

We also fixed a handful of outstanding issues:

Improved performance of dirty transparency compression

Improved simultaneous uploading of multiple files

Fixed bug where the total savings being reported were incorrect

You can now sign up for PunyPNG to save your custom compression options and get assigned an API key.  We're also excited to release the long-awaited PunyPNG API.  This will give you the ability to optimize multiple files in batch using a REST-based API (responses are in JSON).  Smusher is a great example of how you can use the command-line to invoke PunyPNG without sacrificing your own CPU cycles.  The API is currently in limited beta, with some restrictions to ensure nobody decides to optimize their entire photo archive library over winter break or something. Look out for more updates in the coming weeks.  Your feedback and feature requests are invaluable to making PunyPNG the awesome compression tool available.   Keep it coming. Merry Christmas!!

[conrad]

My Password File

Ok, I confess, I keep all my passwords in a single text file, and I keep it stored online. How am I supposed to remember my username and password to the Oracle iExpense thingy at work?  Or the Member ID for my health insurance?  I gotta keep them online so that info is available whether I'm at work or at home. Maybe I'll be the next cyber-theft victim like the poor folks at Twitter who had their corporate documents compromised since they stored it all online using Google for Domains.  Tsk tsk. Well, the way I've been able to pull if off and sleep peacefully at night is using a combination of some awesome Mac tools (I'm sure there's a similar PC equivalent):

Dropbox: Online storage volume, backend is actually Amazon's S3 service

TrueCrypt: Open-source on-the-fly volume encryption -- allows to you quickly mount and unmount secure volumes.

Dropbox is a great tool available for Windows and OS X that allows you to keep a virtual disk online, available everywhere.  I know there are many other similar services and I've tried my share.  However, I'm a big believer in Dropbox because it never crashes and never misses a sync.  You can be moving files around in it, copying a large file in there, and then for kicks, yank the internet connection.  Next time you log in, it syncs flawlessly.  Beauteous! Inside Dropbox, I store an encrypted file container created by TrueCrypt.  I can mount that file container like a USB drive, and I can in turn store sensitive files in there. I know there are password websites out there but I just don't trust some third party to store my passwords.  "Store all your passwords in a single place!"  Something about that value proposition gives me the creeps!  Well, those sites are dime a dozen, and all startupy.  Not my idea of real security.  Ultimately, I want to be the only one who has the keys to the safe.  There's a similar technique which uses Disk Utility to create a password protected AES-encrypted .DMG file, but that requires you to remember to never click "Save Password" when decrypting it and the disk image itself is read-only so it's a pain to make changes to its contents. Creating your Encrypted File Container After you install Dropbox, create a folder called Secure which will have a file called Secure Files (in case it isn't any clearer):

Secure Files is actually an encrypted file container created by TrueCrypt.  That's basically fancy lingo for a .DMG disk image volume that has strong encryption (I'm using Serpent-Twofish-AES ... three ciphers in cascade).  You can easily create one using TrueCypt by clicking on Create Volume:

Save your encrypted file container in your Dropbox's Secure folder or save it to your desktop and copy to Dropbox later.  For the Encryption algorithm, it's up to you.  I chose the Serpent-Twofish-AES since it's basically impossible to break.  Next, set a volume size of 50MB (more if you need to store lots of stuff in there).  Create a volume password (a very long one preferably and one you don't use elsewhere) and a filesystem type (I'm using FAT for highest compatibility) and format the volume. Now, drag your encrypted file container from Dropbox to TrueCrypt and mount it.  Volia!  Super-secure disk image to go.  You can drag important documents and password files directly into the mounted volume (shows up like a USB disk in Finder):

When you unmount the volume (either through TrueCrypt or Finder), it is automatically re-encrypted.  Even if your Dropbox account is somehow compromised, your secure volume files remain encrypted and protected. I also keep a copy of the TrueCrypt application inside the Secure Files folder in case I'm on a Mac that doesn't have it. Now you can keep your password file guilt-free!

[conrad]

GraceList - Craigslist-style classifieds app for small communities

I originally created GraceList as an internal application for our church, Gracepoint Fellowship Church, Berkeley.  With a close-knit community that emphasizes open homes and lives, you can imagine how many things get lost and found.  I wanted a dead-simple classified apps so people could make postings about the small Bible they found at last week's prayer meeting to the new analyst opening at their work.  It has to be easy to use: from our techniest geeks to Kelly, our pastor's wife ;) In the past, we would just email blast our Gracepoint member alias and that would just clutter hundreds of inboxes.  We needed our own little private Craigslist with a smart digest email to highlight new postings. View Demo Features:

Custom categories for posting

Simple UI for making posts.  All you need is an email!

Support for Textfile

Send digest HTML emails periodically to an alias with new postings highlighted

GraceList requires Rails 2.2 or above.  Download and installation notes are available on GitHub: http://github.com/conradchu/GraceList We hope you enjoy using this simple little app that has made life a lot easier for us.  We're glad we can finally share it with the rest of the world.

[conrad]

punypng Benchmarks

I recently ran some benchmarks, comparing punypng to a lot of the popular tools out there. I selected tools I felt were decently strong and versatile in handling images:

punypng: the new kid of the block.

smush.it: Uses pngcrush as the main PNG optimizer.  Currently, available in Yahoo's YSlow Firefox plug-in.  I believe it uses the -brute option for pngcrush.

OptiPNG: A slighlty better algorithm, compared to pngcrush.  In case you're wondering, Google's PageSpeed plugin also uses OptiPNG for it's compression library.

ImageOptim: The heavyweight contender.  Available for OS X, ImageOptim uses every major library out there: advdef, pngcrush, optipng, pngcrush, jpegoptim, jpegtran, and optionally pngout.  I ran this the benchmarks with pngout enabled.

Now, it's tricky business to try to compare compressors side-by-side against a single image.  So instead, I selected some typical images that I found on the web, like sprites from some popular websites (Facebook, Yahoo's icon set, and of course Ask.com, hehe).  Images vary from 24-bit transparent PNGs to 8-bit indexed GIFs.  Amount of transparency among the PNGs also varies. I also included a blocky-styled JPEG to see how the different tools handled it.  I know some might cry foul because I included a JPEG with solid colors (since you could have just saved it as an indexed GIF or PNG).  However, I've seen hundreds of JPEGs that people have been uploading to punypng and it just shows that it's often hard to know which format was certainly be the best one. OptiPNG and ImageOptim also cannot process JPEGs and GIFs respectively.  Before benchmarking, I converted them to the more digestable PNG format using ImageMagick to ensure no metadata was added In total, the test images in total weighed in at 276KB. Download the Test Images Here are the results: punypng was able to achieve 41% savings, almost double the competition  The most surprising result was how much punypng trounced ImageOptim, especially on the NASCAR image.  I was actually thinking that ImageOptim would beat punypng on some of the tests, since it uses everything under the sky.  In terms of processing time, punypng was a wee faster than smush.it, about the same as OptiPNG, and light-years ahead of ImageOptim, which took eons to compress as it's CPU-intensive. One caveat I want to make is that this benchmark is not meant to say punypng is going to beat out the competition every time.  However, on average, it does an excellent job, especially on images with 24-bit transparency.  Libraries like pngout (which is part of ImageOptim's toolbox) does a good job as a standalone library, but it's very slow.  I've often found that pngout only saves a few bytes here or there, and isn't worth the wait. I think one thing that favors punypng (as well as smush.it)  is that it handles every kind of image you would use on web page: PNGs, GIFs, JPEGs, and Animated GIFs.  I was kinda bummed with the other tools because they primarily processed only PNGs and had weak support for other formats.  Personally, I just want a tool I can use to compress all my images using the latest compression technologies and techniques, and for me, that's why I recommend punypng to all.

[conrad]

punypng now supports Dirty Transparency

Gracepoint After Five's super-ninja compression tool, punypng, got a new upgrade to help slice and dice the size of your website images.  punypng's compression algorithm now automatically supports Dirty Transparency as described by Sergey Chikuyonok's awesome article on Clever PNG Optimization Techniques that was recently published on Smashing Mazagine.  This unique and innovative technique offers additional compression to any PNG or GIF as long as it has either 8-bit or 1-bit transparency.  On average, you can save about 10% more, and in some cases, you will save even as much as 35%! As I wrote previously regarding dirty transparency, you can use our Photoshop action to create a more optimized PNG.  If you ever find yourself outputting a layer that has been masked, you can look to get some serious savings. [caption id="attachment_124" align="aligncenter" width="300" caption="Typical use case where dirty transparency will give you the most savings"][/caption] This butterfly image above was reduced by 36% when I ran it through the updated dirty transparency compression algorithms in punypng.  This image was only reduced by 7% when dirty transparency was disabled. One of my Ask.com UX Design buddies was a little skeptical in terms of how dirty transparency would help our team.  He mentioned that our sprite preparation process already optimizes the transparency of our assets;  transparencies are already cut and matted ... no hidden image data.  Right? Well, I wasn't sure either way , so I decided to run a test on our Ask.com sprites.  Our sprites span a good variety of use cases, ranging from full 24-bit color PNGs to 8-bit GIFs to 24-bit hand-quantized, indexed PNGs.  In terms of image data, everything is as tight and pixel-pushed as they possibly can be ... transparencies are matted and all masks applied. Here are the results: [caption id="attachment_129" align="aligncenter" width="300" caption="punypng's new support for dirty transparency "][/caption]

The above results speak for themselves.  The results are very surprising!  punypng was still able to squeeze out another 3% of compression on average on top of the existing compression algorithm! So what's the catch.  Well, in some minor cases, the dirty transparency technique produced images that were 1-2% larger than usual (though they were still smaller than anything that pngcrush or optipng could ever produce).  However, these were rare cases and the savings as a whole in fact did add up to 10% or more.  In terms of performance, using dirty transparency to compress 15 or more files adds only added 1-2 seconds to the process. In general, I think it's safe to say that dirty transparency is a proven technique that warrants inclusion in everyone's arsenal.  It's my hope that with this latest addition, punypng can start emerging from the myriad of compression tools on the market as the one-stop tool for designers and web geeks who are serious about making the web more puny.  Give a try and post your comparison results. Thanks again to Sergey Chikuyonok for his original contributions.  I look forward to adding more of his techniques into punypng. Update on 10/12/09: PunyPNG supports IE6 using the dd_belatedpng plugin.

[conrad]

Dirty Transparency in PNG Optimization

Smashing Magazine has a great article on PNG Optimization techniques.  The dirty transparency technique is awesome, and I started running tests on our internal assets at Ask.com and Gracepoint Berkeley, and we're seeing some significant savings.  Just when I thought PNG compression was all there was to it, Sergey Chikuyonok (the author of the article), blows us away with some really innovative optimization techniques. To make dirty transparency easier, I created a Photoshop Action that you can use to run on your PNGs.  Make sure to run it on single-layer PNGs that have already been saved off via Photoshop's Save for Web.  Don't try to run it on your multi-layer PSD file as the script is looking for a layer called "Layer 0". Download: Dirty Transparency Photoshop Action (1KB) I also like his usage of posterization as a way of reducing colors.  At Ask.com, I developed a similar technique except using Photoshop's color-indexing algorithms (perceptual, adaptive, etc) which I think offers the compressibility of posterization but does a better job at maintaining smoother gradients/detail.   Posterization's diffusion characteristics are a little rough for my taste.  One of these days, I'll post a tutorial on how we do this technique, which isn't terribly innovative imho, but I haven't seen anyone talking about it so it'll be worth writing up at some point. Meanwhile, be sure to check out punypng, as simply the best way to optimize your PNGs on the net.

[conrad]

punypng: making the web more puny, one png at a time

I finally released punypng to the world last week.  It's a free png compression service that intelligently leverages multiple open-source png compression algorithms in the hopes of making the web  more puny, one png at a time. Try out punypng (as a short url: http://www.punypng.com works as well) How punypng came about punypng actually started not as a need at Gracepoint Fellowship Church, but rather at my day job.  At the Ask.com User Experience Team, we believe down to our toes that design is about craftsmanship.  Our designers are obsessive about detail, and though few notice, everything is "pixel-pushed" -- no stray pixels, no unnecessary colors.  A big part of search is fast page load times.  Every search experience team (Google, Yahoo, etc) knows that the faster the page loads, the more loyal your users will be over time (if you don't by now, well, now you know).  The Ask.com team previously relied on Yahoo's smush.it service (no shame, Yahoo dev tools are great and a huge asset to everyone), which was very impressive, but for our day-to-day work, we wanted something that was made for designers by designers.  So after a few sleepless nights thinking how I can make my PNGs smaller, I decided to start on the punypng project.  Fast forward a month later, punypng is now the bread and butter tool among the Ask.com UX designers (and the Gracepoint designers as well) The specialness of punypng I don't claim that punypng is for everyone.  I wanted a tool made for designers, and so I also left out a lot features (for now) such as an API or fetching images via URLs (I assume all original assets are on your hard drive not on some website).  But though lacking in these small ways, it does boast some great features:

Fully supports for PNG, .GIF, .JPG

Clear affordances for # of bytes saved (as well as being pretty bar graphs)

JPEG Compression -- punypng doesn't leave JPEGs out in the cold.  JPEGs are analyzed to see if a compressed PNG format is better (ex: JPEGs with heavy solid areas benefit from this).  But if not, don't despair, punypng is backed with jpeg-tran and jpegoptim for further JPEG optimzation.

"Fire-and-forget" batch processing: You can upload up to 50 files in a single session.  Optimized versions are clearly labeled, and if no further optimization can be made to the uploaded file, you get the original back untouched.  After you upload a batch, you can go ahead and upload another batch without having to reload the page.

Download batch jobs as a single time-stamped ZIP.

See the changelog for future updates

The future I don't know what the future holds in store for punypng.  Please let me know if there's some enhancement or new feature that would help your day to day work.  punypng has relentless commitment to making every png as small as possible, so barring enormous CPU requirements, let us know if there's some experimental binary out there that we can include. The great thing of an online tool vs a desktop tool is that we can constantly improve the performance and efficiency of the compression as new algorithms are made available.  Just today I found a way to squeeze out 3-5% more.

Help support punypng Well, every Gracepoint After Five project is free to the world, and is our belief, that all the little tools we build will end up benefiting the church at large (as well as my team at Ask.com of course).  Running a CPU-intensive site like punypng isn't cheap.  Please help support the cause.  If not, please help spread the word.  Our gratitude to you is non-puny. Update on 10/12/09: PunyPNG supports IE6 using the dd_belatedpng plugin.

[conrad]

Thou shall not rip off

Last week, Otty my trusty senior designer that leads the Dictionary.com design work sent me a website I never seen before, http://mydictionary.myresources.com.  The shocking thing is that bears an awful resemblance to our Dictionary.com homepage!

Errr, rip off?  The funny thing is that they tried to change the colors around and removed some content to lessen the rip-off index.  Maybe Otty has a screenshot of what it looked like two weeks ago because it was even more blatant. If that wasn't enough, yesterday, my old colleague Dahveed Gomez-Rosado, who's VP of User Experience at Lithium, twitter'd this: "The most flagrant and cynical act of plagiarism I have ever witnessed in the Web. Ever: http://ow.ly/gByA" Perhaps Kevan at his Turnit.com gig should figure out how to do anti-design-plagiarism.

Wow, two plagiarism offenders in two weeks for me! Gracepoint After Five of course uses a modified version of the Agregado theme, which was masterfully created by Darren Hoyt and Matt Dawson of Category4. I couldn't swallow the original pink feel of Agregado, so I tweaked the Hues in the Photoshop file they provided everyone and bam, we were good to go.  Agregado is awesome and Darren/Matt deserve all the credit.  Not us.