Br 12 vs Br Beta -- performance results

Before beginning, I should probably post up my pedigree.  I was invited into the Bridge prerelease program ca. 2005-06 by Chris Cox.  I have tested every prerelease, release, and beta version of Bridge since then.  That’s likely in the vicinity of 100+ versions and many thousands of volunteer hours invested in assisting Adobe in the development of Bridge. 

I have used Br as the primary asset management tool in our studio’s production environment for all that time.  It is unlikely that anyone in this conversation has significantly more experience in testing and using Bridge in a high-volume production studio than I do.  My testing parameters have been refined over the years to provide the most accurate, predictable, and repeatable results. It works well in all our studio folder hierarchies. Your conditions are not ours, so your milage may vary.

Here are the parameters of my most recent performance comparison.  The test bed is a brand-new Windows 11 laptop with an AMD processer comparable to a higher end i7 laptop. 

The attached videos will demonstrate the relative efficiency of Br 12 vs the latest beta build of Br 14 when browsing a hierarchy of folders that contains the following:

• Total file size of the folder hierarchy 14,659 MB
• Number of folders: 25
• Number of files: 4076
• Roughly half of the files are a mix of Psd, Png, and Tif files.  The vast majority of those are less than 4 megabytes each.  The other half are proprietary txt files (.flame / .chaos) that serve the same purpose to their parent software as .Xmp files serve to their companion RAW files.  Those are only a few kilobytes each.

I have given both Br 12 and Br beta their own independent cache location on the primary drive. All of the files have been cached in advance by both builds to eliminate caching behavior as a performance variable.

The process of the test should be obvious from the linked videos below, but in short, Bridge is launched and a folder hierarchy is collapsed to reveal all the assets in that hierarchy. Immediately an attempt is made to operate on those assets, in this case using various filters to narrow the selection down.  This is a common daily activity in our studio.

It has been well known for many years that there is typically a slight delay when accessing the subfolders of a hierarchy while Br accesses the cache and generates thumbs. Accessing thumbs from the cache takes some time. The performance in the filter panel in Br 12 and earlier is not affected by this.  Br 13 trashed that performance, and the results still remain in Br 14 as you will see in the following videos.

Also notice the fact that the valuable arrow keys on the panel scroll bars were eliminated with the first Br 13 release and never returned to the interface, while giving no equally valuable alternative in return. The feature was simply removed, or perhaps more accurately never incorporated into the new architecture. Why?

As far as the conversation about scrolling speed of thumbnails is concerned, I would submit that hobbyists and other casual users may not notice the inconvenience.  On the other hand there are many, many highly skilled artisans who have had their studio’s productivity severely compromised by this defect.  It took years of effort, testing and feedback from prerelease testers to help Adobe establish the performance level of Br 12.  A great deal of that progress has been severely damaged or destroyed by the new architecture.  Of that much I am certain.

I’ve seen a number of inconsequential features added to Br over the years, but at least those changes have always been easy to simply ignore. The destruction of years old professional workflows that has been wrought by the subjugation of Br 12 by Br 13-14 is impossible to ignore. The pity is that Adobe was repeatedly warned about this eventuality beginning almost 20 months ago, before Br 13 ever became a public release.

Here are links to the head-to-head testing results of Br 12 vs the latest Br 14 beta. I’ll leave them on Dropbox for a bit, but I’d recommend downloading the two files to your local drive for the most accurate results.