Yet another bug got fixed, whereby it was clearing an edit handle before pasting between 2 edits with the same source, resulting in the after edit being lost & replaced by the rest of the movie in the before edit.  It was clearing in all cases where the selected area had zero length.  The EQUIV operator is more difficult to use than == & won't work if the frame rate is in the thousands.  Lions are now assuming a start & end coming from user input are always going to work with the == operator.

 

 

During this process, 1/16 sample zoom exposed a few anomalies. 

Noted if the source samplerate doesn't match the project samplerate, it draws the samples offset from the sample ticks.  

Also noted the 48khz samples line up with the 24 frame rate but the 44.1khz samples do not.  It's possible for the insertion point to line up on a frame but not a sample in certain cases: double click an edit, ctrl seek to label, alt seek to edit.  Align cursor to frames is ignored in these operations.

 Align cursor to frames is only honored when single clicking on positions & edit points.  Single clicking an edit point aligns on the sample if  frame alignment is disabled & aligns on the frame otherwise.

Lions recall intending cursor positions to always be quantized to the sample but never finishing it.  That later gave way to quantization based on align cursor to frames which too was never finished.  It makes seeking really clunky.  

It would be quite difficult to get ctrl/alt seeking to follow align cursor to frames.  It would have to escape the nearest frame or sample before beginning the search.  It's justifiable to ignore align cursor to frames when double clicking, but not consistent with the single click behavior.

-------------------------------------------------------------------------------

 

The firehose of low value which previously didn't exist between us continues with a claim all of adobe's software has now been replicated from just the documentation.

 

https://github.com/storytold 

 The names were obfuscated.  The READMEs were all generated in that chicken scratch claude style with snazzy headlines you've come to love.

 

 

We're utterly dependent on the language model to condense its own massive amount of generated text into an estimate of reality.  If even that 50% level of functionality existed 30 years ago, it would have been way beyond what lions had in those days & lions wouldn't have bothered down the path of Cinelerra.  

 Just missing are any kind of precompiled binaries.  It's almost impossible to build.  Instead of any demo videos, we only have AI generated sales pitches imagining what it would look like if it worked.

Exercising due diligence in an age of overwhelming generated dead ends, cloned it. 

 It seems to have largely transcribed prior open source code to rust.  "Codec::name() matches FFmpeg's codec_name for these."  The big database is defined in crates/project/src/lib.rs

 //! All time is an integer number of [`Tick`]s at [`TICKS_PER_SECOND`] = 254 016 000 000/s. That
//! rate divides evenly into every broadcast frame duration (23.976, 24, 25, 29.97, 30, 48, 50,
//! 59.94, 60, 120 …) and every common audio sample duration (8 k … 192 kHz, including the 44.1 k
//! family), so edits, frame math and audio alignment never drift.

 Not likely picking a certain unit for all time codes eliminated all drift.  Younger lion concluded the == operators would still be subject to near misses but integers would always give reproducible values.  The constant is based on only standardized formats but would also work with powers of 10, powers of 4.  That's a 5 byte fraction.  It leaves enough bits for 10,000 hours.

The language model claims to have achieved instant compatibility with every standard file format. 

The program is manely in crates/.  The main windows are defined in crates/ui-egui/src/panels.

It's hard to imagine why anyone just starting out in editing media would use anything but an AI generated replica of an industry standard.  The big question is what new capabilities AI is going to invent instead of transcribe from prior work.

 The only reason to continue using some odd ball editor or heaven forbid, developing a new one with its own odd ball file formats, might be a preference for a certain user experience or just a fascination with writing software.  

It makes lions wonder what project they did in the last 30 years would still have any relevance in a resume or what else they could have done for the last 30 years.  Writing a video editor is now as germane as breathing.  The window of relevant work is shrinking to just the last nanosecond.  There's a role for window dressing a finished product, guiding a project, creating training data, but the software industry is now like being a musician.

Lions would never have started manually penning the X11 calls for an audio editor 30 years ago on this day, but would have simply prompted a language model for an audio editor that compiled on a Mac.

Probably would never have migrated to UNIX or bothered learning all the things lions did in the last 30 years.  That body of knowledge is like so many scientific constants now, that no-one bothers remembering.  The last 30 years of source code still had to get created & uploaded for the language models to learn from it, though.  It would have been a long wait.  Doing nothing would have been like waiting until all music became free on gootube before hearing it.

 There's unlikely to be a need for any more major changes like the time format in Cinelerra. It would have to be some esoteric feature never previously created.

After 30 years, lions are once again in a lost phase, wondering what's next after a path they were on fizzled.  1st it was music.  Then it was biology.  Then it was software.  A minimum wage job would not pay the bills but it would take a lot of heat off.  It wouldn't be the most optimum use of time, though.  Perhaps retirement is the new engineering & lions are once again over extending the status quo.  Perhaps it was the memory of watching everyone else making money.  Eventually, not being the ones making money starts feeling natural.