Open wide, motherfuckers.  This end of the 2 headed economy is going supersonic.

 The lion kingdom has more than the value of a house in the stonk market, so in terms of missing out on  the top head of dual headed inflation, lions probably have that covered.  Real estate is a volatile market just like stonks, manely investor funded rather than owner occupied nowadays, & not diversified.  Stonks & real estate are both in the top head, but stonk prices won't self limit because they don't count as inflation.

The 2 big risks are the pre rate cut media blitz & the pre election fear of Kamal.  It's been 30 years of nearly continuous democrat rule, but so far no sudden wealth tax.  Obama delivered a new tax on capital gains, a 4% bump for realized gains & only for AGI above $200k. Remember the good old days 10 years ago when $200k wasn't poverty?  Everyone is getting hit by the NIIT nowadays.

It's surprising Ego Monger's husband follows the cloth, since he pays himself in  dividends, probably makes over $250k, & is probably subject to the NIIT.  He moved to US in 2014 after the NIIT was enacted, so he probably thinks it always existed.  When lions see lots of short term  jobs like his, punctuated by long periods of self employment, it conveys being hard to work with.

Scott Manley is a salaried employee so he probably deducts his retirement account, mortgage interest, & doesn't collect any dividends.  His gootube income is probably kept inside an LLC without paying any dividends.  These guys file as married so their income threshold is $250k.

Rossguy said before that he would not vote for Trump so he would be firmly another blue state refugee who still votes democrat.  He probably keeps all his gootube income in an LLC without paying dividends.  He stopped making any vlogs before Bide's campaign went down the tubes though.  It's impossible to vote democrat without extensive tax shelters.

 

The NIIT didn't destroy the market like a 25% unrealized tax but it did extend the rotten years as animals sold more to pay their new pet tax.  Investors are going to fear a 25% unrealized tax & sell like no tomorrow before the election, but more likely we're going to get another capital gains tax increase & the dip will be short.  

Calif* has especially deep capital gains tax & amerikans see Calif* as the role model for the country.

 

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

Paw has had a really good week, since beginning metronome usage & attempting harder efforts.  Broke into the 9 minute regime briefly.  These harder efforts seem to be going better & promoting more improvement than endlessly taking it easy.  The metronome could be opening the door to harder efforts by changing the  running form.  Without it, those harder efforts just set everything back to the beginning.  Things still get set back.  It hurts on day 2 & there's a definite intensity limit, but it hasn't been the end.

Of all dietary supplements, whole chicken seems to be the most beneficial in terms of tendon recovery.  It's the staple food in Kenya.  It seems to have the optimum amino acid concentration.  Whey protein might benefit muscles but not tendons.

Lions would recommend dumping jello & buying more, cheaper whey.  Jello is more expensive per gram of protein.

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

HMSF has proven useful for still frames because you can see the frame positions while still seeing meaningful time in the movie, but the time bar leaves a lot to be desired.  Text stamps would ideally line up with seconds & minutes.  It originally divided the text stamps into 1 text stamp every 10 frames, 1 text stamp every 25 frames, & sensible numbers of frames between text stamps.  It looked good for frames but it made text stamps for HMSF line up all over the place instead of on seconds & minutes. 

It can be hacked to line up on seconds & minutes above 10 frames per text stamp.  The trouble spot is now 5 & 2 frames per text stamp.  It works for frame rates which are a multiple of 5 & 2 but not 24fps or 9fps.  Try to make it align on seconds for 5 & 2.

It gets erratic for 5 & 2, depending on the frame  rate.  It'll align every 3 frames per text stamp for 30 & 60, every 6 frames per text stamp for 60.  For 23.97 & 29.97, it falls apart at all zoom levels where each frame has a reticule.  For the irrational frame rates, it has to align on powers of 10 frames instead of discrete times.

For 5 & 2 frames per time stamp, the spacing between time stamps is not a constant time like it is for >= 10 frames per time stamp.  This is a unique condition of HMSF.  It should also apply to feet frames but that feature was never used & probably will get dropped.  The easiest solution is probably a ground up drawing routine.

For now, aligning HMSF on seconds & minutes is such a difficult problem, it's far easier to live with odd time stamps.  The bigger problem might be the way it doesn't align on 10 minutes like HMS does.  That's because it's aligning on powers of 10 frames.

The problem  is the seconds don't align on exact frame boundaries for irrational frame rates & certain zoom levels.  The tick marks follow the time stamps, so they end up shifted or uneven if the time stamps are shifted.  The solution is to quantize the positions of the time stamps so they're drawn on frame boundaries while still showing the round numbered times.  Then the tick marks should automatically line up with the frames.

Helas, even that still results in rounding errors in certain conditions with high frame rates or low zoom levels.  The favored solution is increasingly a scratch built routine just for HMSF.  Younger lion showed quite an ability in getting it to work as well as it did, 30 years ago.

Another go at this involved keeping the text unchanged but aligning the drawing of the bits on frames.  This still left some rounding errors & overlapping text, depending on the frame rate.

Nothing was working for every combination of zoom level & frame rate, so there was an idea to advance 1 frame at a time for all zoom levels where the frames were over 1 pixel wide.  Then determine if we're on a whole multiple of the current timestamp spacing.

For zoom levels where the frames are under 1 pixel wide, use the existing algorithm for continuous time formats.

That gets the tick marks always to line up on frames, but what time stamps to draw remanes difficult.  The part about determining if the current frame is on a multiple of the current timestamp spacing is the problem.  That suffers from rounding errors.  It alternates between passing on even & odd frames.

The concept of a constant timestamp spacing might be unworkable.  It might need a list of timestamp priorities & to draw lower priority timestamps based on available space.  It'll do 1 pass for each priority in the heirarchy table, drawing all the timestamps in the current pass which fit in the remaneing room.

That algorithm seemed to get closer than the rest, creating the necessarily uneven text marks & drawing the whole second time stamps.  It still can't consistently detect frames on the text intervals.  There is a problem with the text formatting rather than the algorithm.  It sometimes skips frames with the irrational frame rates.  This has been fighting the problems with time stamp alignment.

Units::totext has been the scene of much carnage over the last 30 years.  TIME_HMSF appears to be the 1 trouble spot.  There seems to be not enough accuracy to get below .5 frames of accuracy in the text representation.

The problem was narrowed down to the frames & ticks hitting 1 alignment but the cursor hitting frames at a slightly different alignment.  The tick for 30:00:00 draws at 1799.998 seconds or frame 29:59.832.  The cursor hits this frame slightly to the right at 1800 seconds & time 30:00:00.  To draw the GUI, it added a .5 frame fudge factor to the text formatting so the cursor hit the frame slightly left of the tick & the tick was bumped up 1 frame.

Calculation for the cursor at 30:00:00:

MainClock::update 56 position=1800.000000 107892.107892 107892.107892 0.000000 0:30:00:00

Calculation for the tick at 29:59.832:
MTimeBar::draw_time 283 tick_position=1799.998200 107892.000000 107832.167832 59.832168 0:29:59:59

This could be a limitation of physics.  A whole number of frames just can't line up to an exact second while the cursor can be drawn at an exact second.  The fudge factor doesn't seem to matter if align to frames is enabled, but it is making the cursor hit .5 frames ahead of where the actual 30:00:00 is.

The right solution is the .5 frame fudge factor for tick marks but no fudge factor for the cursor.   The problem is the 2 times have to agree for selections & for aligning to frames.  To make it user friendly, the fudge factor has to be applied to both uses.

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

NOTE 1: the fudge factor is being applied to irrational frame rates where it's needed & whole frame rates where it technically shouldn't be needed.  In whole frame rates, it always makes the cursor advance 1/2 a frame early while in irrational frame rates, the lead ahead varies.  Testing showed eliminating the fudge factor requires more than setting it to 0 for whole frame rates.

NOTE 2: Another problem is going to be extremely high frame rates above 1000fps.  It's using 1/1000 for equivalence operations.  In lion defense, the footage might be  1000fps but the timeline is never going to be over 120fps.

NOTE 3: The bigger problem is Cinelerra representing time using floating point.  Integer nanoseconds wouldn't do the job either.  Video time needs an integer numerator & denominator, but it would still not be possible to convert between exact audio samples, frame numbers, & seconds with integers.

Audio sample rates are locked into multiples of 44100, 48000 by hardware oscillators.  Young lion knew about the sample rate restrictions by 1995 but was naive about the reach of standards & thought arbitrary rates would always be used for video.  Helas, converting between frame rates proved impossible early on.  Young lion's 1st challenge in 1999 was discovering video couldn't be captured at 24fps.

When it didn't show round numbered time stamps, it showed the true time when each frame appeared but it was hard to navigate.  The true time they appear is not on round numbers of seconds.  It's now making a compromise in rounding the display time to legible time stamps for the sake of usability.

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

HMSF gradually replaced the 30 year long preference for HMS.  Simple things like the time stamps lining up on round numbers make a difference.