I open a new thread on av1_vaapi because the other one is only about
Intel drivers.

Thank you, Andrea,  for starting a new thread and a reminder to myself and others that if your reply is not related to the Subject line, it is best to start a new thread.  Otherwise, I am having a hard time finding something when I need to get back to it !

I put the results of my system ffmpeg and CinGG's internal ffmpeg.

$ ffmpeg -encoders | grep av1
 V....D libaom-av1           libaom AV1 (codec av1)
 V....D librav1e             librav1e AV1 (codec av1)
 V..... libsvtav1            SVT-AV1(Scalable Video Technology for
AV1) encoder (codec av1)
 V....D av1_nvenc            NVIDIA NVENC av1 encoder (codec av1)
 V..... av1_qsv              AV1 (Intel Quick Sync Video acceleration)
(codec av1)
 V....D av1_amf              AMD AMF AV1 encoder (codec av1)
 V....D av1_vaapi            AV1 (VAAPI) (codec av1)
 A....D wmav1                Windows Media Audio 1

$ /home/paz/cinelerra5/cinelerra-5.1/thirdparty/ffmpeg-6.1/ffmpeg
-encoders | grep av1
 V....D libaom-av1           libaom AV1 (codec av1)
 V....D av1_vaapi            AV1 (VAAPI) (codec av1)
 A....D wmav1                Windows Media Audio 1
...
  I see that Phyllis'
system has only two libraries for av1 and all the others are missing.
...
Now it makes sense after Andrea's research in explaining this in more detail.  Yes, my libva is out of date because my O/S is old, but on a duplicate laptop with a partition loaded with current Fedora 38, I do see av1_vaapi as an encoder.

I am attaching a new version of the preset for av1_vaapi, but there
are many conditions for it to work: you have to install CinGG
(appimage is no good, unless you know how to open and recreate it);
new generation GPU (Nvidia 4000; AMD Radeon 7000; Intel Xe 2 or Intel
Arc); system with updated libva (possibly Rolling...). If anyone meets
the requirements, I would like to know if the preset works or what
errors it gives. I used average options between quality and speed but
could not test them because I lack the right hardware.

I did verify that the attached render formats have correct values, even though I lack the right hardware too, by just adding garbage onto each line which then generates an error message (also, changed encode.opts in the above directory to "loglevel=verbose").  This verifies that it does process the parameters for correctness before it fails on the hardware.  So I will be checking these into GIT -- the user can easily change values to suit their needs.

Too many of the newest speed-ups are so hardware dependent which is so varied for everyone.  Not too many people want to update their graphics board all of the time at $500 or so a pop and then struggle with updating the software too.