By Antony Bolante Macworld.com | on May 12, 2010
Price: £675 plus VAT . upgrade £238 plus VAT
Pros: Ultra keying effect; Improved speech-to-text; Improved performance and stability; Native support of popular formats; GPU accelerated effects; 64-bit Adobe Media Encoder; New Adobe Story.
Cons: Two GPUs qualified for hardware acceleration; Encore not 64-bit; Transcriptions difficult to edit; No playback frame rate display; High-end system needed for best performance.
Premiere Pro CS5 sports the Mercury Playback Engine, a software mechanism that improves the program’s performance -- especially when paired with a qualified GPU. And like some other programs in CS5, Premiere Pro has become a native 64-bit application, shattering the RAM limitation imposed by 32-bit programs.
Other features include expanded tapeless format support, scalable playback resolution, the Ultra chromakey effect, better still frame export, and the ability to export directly from Premiere Pro. But apart from enhanced performance and stability, the most notable new feature is how Premiere Pro CS5 fits into a workflow that integrates new software and services.
CS4 expanded the Premiere Pro package to include OnLocation and Adobe Media Encoder -- programs dedicated to the beginning and the end of the video creation process. CS5 extends its reach into the pre-production phase with a script creation and collaboration program called Adobe Story, which is still in beta.
As an Adobe Air-enabled online service, Story allows multiple writers to edit or comment on the same document. Visible tags identify who makes a change and when. You can also work in offline mode, locking the online version of the script to prevent collaborators from making changes before you upload your updates. Collaborating and saving scripts online is free. But after a year of complimentary service, Adobe plans to charge for its various online services. The fee -- and whether users will be willing to pay it -- remains to be seen.
Story’s interface is spare but intuitive. After selecting a template for, say, a screenplay or two-column A/V script, you can use a context menu to select formatting options: a scene header, dialog, and so on. Story checks spelling and prompts you to save unfamiliar words into a custom dictionary.
Clearly, Story wasn’t designed to compete with a full-featured program, such as the venerable and expensive Final Draft. Instead, it’s aimed at writers who want a simple scriptwriting program that permits remote collaboration. And Story succeeds, especially for an initial release. But even if you write scripts with another program, you can import them into Story to take advantage of its real selling point: its role in a metadata-enhanced workflow.
Starting with Story, metadata is passed from one program to the next, automating tasks at each step. When imported into OnLocation, the Story script generates a shot list consisting of placeholders, a process comparable to a traditional script breakdown. Once you execute the shot list, OnLocation can link the media to the corresponding placeholders—uniting their metadata, as well.
By utilizing the original script as a guide, Adobe Media Encoder can transcribe the speech in each clip far more accurately than it could in CS4. In Premiere Pro, the ability to search clips according to their metadata transcripts expedites the editing process. The metadata is even retained in a Web DVD exported from Encore, so that the viewer can search the online video according to keywords.
The workflow these features promote favors script-driven projects that use tapeless media. And the preproduction tasks they automate aren’t replicated in every detail. For example, whereas a traditional script breakdown would generate a shot list that includes all of a scene’s coverage—cover shot, close-ups, reaction shots—OnLocation only lists a shot for each scene heading in the script. So an assistant cameraperson will still have to modify the shot list manually. And because speech-to-text relies on a reference script for accuracy, ad-libbing is discouraged. Transcripts are accurate enough to be truly useful, but are still as difficult to edit as they were in CS4.
That said, implementing this workflow could save time and present new possibilities. Other workflows could benefit, as well. An interview-driven documentary doesn’t have a script, but a manually generated transcription would enable the speech-to-text feature to produce metadata accurate enough to help an editor sift and edit hours of footage.