I know everyone hates Adobe (I certainly share some of the grievances about their pricing structure) but I have been using Photoshop for something like 23 years, working across half that many creative disciplines, and cannot overstate how much I value the consistency, reliability, and neutron-star density of its feature set. You call it bloated, I call it a Swiss army knife. I don't want every tool to be that way, but I love that some are. (The problem with this analogy is that Photoshop actually /is/ the best tool for many of the things it can do, where the Swiss army knife is solidly in jack-of-all-trades territory.)<p>Given this, I am very concerned about Adobe's continued march into web apps. For most of my UI and web design work, I use Figma, and quite like its feature set, but that I cannot use it offline and that I cannot work in a file that requires over 2GB of RAM due to its browser RAM cap is infuriating to me when Photoshop (the supposedly worse program) enables me to manually allocate as much of my system's 64GB of RAM as I please. It's difficult to imagine any web app stably and efficiently growing to Swiss army knife-levels of capability. Not just in terms of performance and resource management, but also GUI, keyboard shortcuts, etc.<p>I've long assumed my CC subscription would die with me, unlike many (most?) of my peers, but if the web-based Photoshop somehow became the primary Photoshop, I may have to cancel my subscription, and then kill myself.
I discovered today that Adobe is introducing Project Shasta, a web-based AI-powered audio recording and editing tool. You can request access now [1].<p>It seems their intention is not only to provide AI-based audio editing and recording but also to cover the ground of what Riverside already offers.<p>Remote recording - Recording with others is as easy as sharing a link. Everyone’s audio is recorded in high quality locally, then Project Shasta syncs it back together in the cloud automatically.<p>"Soon to come features" list things we can already find in Descript, such as filler word removal or speech enhancement. And features already available in the beta include microphone checking for optimal quality and microphone distance, AI-powered audio, remote recording with guests, and project templates.<p>I was waiting to see what Adobe's take on AI-based audio editing and recording was, and it seems it's here. At the moment the project is referred to as Shasta; I imagine these capabilities either format a new product offering or are integrated with existing tools like Audition, Premiere, or After Effects.<p>What do you think?<p>[1] <a href="https://pages.adobe.com/shasta/request-invite" rel="nofollow">https://pages.adobe.com/shasta/request-invite</a>
Is it still called "Sherlock-ing" if Apple isn't the one doing it?<p>Adobe XD : Sketch :: Shasta : Descript (<a href="https://www.descript.com" rel="nofollow">https://www.descript.com</a>), right down to the advertising style
The denoising works very well. I found that mixing the enhanced audio with about -3 to -6 dB with the original voice audio works very well.
(short clip at a brewery)
<a href="https://youtu.be/M5pdHVoXQHE" rel="nofollow">https://youtu.be/M5pdHVoXQHE</a><p>I also tested with wind noise (outdoor speaker) and the audio "recovery" is amazing. It rescues bad field recordings that you normally can't use.
I'm impressed!
Hi! I’m Sam, lead designer of Shasta. You can check out the app at <a href="https://shasta.adobe.com" rel="nofollow">https://shasta.adobe.com</a> (OP's post links to one of our launch announcement pages). Our enhance speech and mic check features are already available for everyone to try.
This “Mic Check AI” - I actually just want to try this one feature.<p>I never know what gain I should set on my XLR mic + USB audio interface and I also don’t know how Teams/Webex will do to my signal. Is it compressing my signal? Noise gating it? No idea!
As a Figma employee, there's obviously been a lot of shakeup recently for us. That said, this is one of the areas that actually makes me really excited to partner with Adobe. They've always been at the forefront of AI-esque assisted editing (content aware fill in photoshop was so far ahead of its time). Definitely excited to try this out and to see other AI projects by Adobe to help creatives.
Really excited to see the transcription based editing. There is a ton of work that this can simplify. I really hope some of this gets integrated into Audition or AE etc as well.
Signed up for early access. We currently use Zencastr for recording our podcast, which has been great, but curious how Shasta will stack up and it'd be nice to not have to pay for two subscriptions.