This LP on Luaka Bop could have gone very wrong instead it went very, very right. Floating Points on the controls, Pharoah Sanders on his horn - still going strong after all these years & The London Symphony Orchestra on everything else. A peace & love filled audio antidote to the angst of the last 12 months.
This chapter should provide some relief if the first one stressed you out ;). Thanks for reading!
Looks like there's some kind of listening party scheduled for this coming Sunday! It also looks like it's just the first of a three part series of some kind 🤔
Light details and RSVP here: https://www.listentopromises.com/
Listening to the broadcast of Promises by Floating Points and Pharoah Sanders. I feel a big influence of Eno (especially the serie "Ambient") so I was wondering what you guys are thinking about it
I'm trying to understand how numbers are stored whne talking about float, decimal, double types.
In int, for example, we have the following: 32 digits, lsb is (0 or 1) * 2^0 bit. From right to left, msb is (0 or 1) 2^30. 31 digit tells is if it's positive (0) or negative (1).
And what about float? msb is the same. Exponent is integral part of number and mantissa is fraction? How free digits are distributed between exponent and fraction? And how they're ordered? Numeration from right to left?
Double floats only support up to 53 bits of significand precision, exponents down to -1023, and there is significant trouble with floating point errors. But as of 2020 we have calculated 50 trillion digits, and there are many infinite Mandelbrot set zoom videos. How is this done with limited floating point arithmetic?
Hey - missed out on the pre orders for Floating Points, Pharoah Sanders & The London Symphony Orchestra – Promises and am super bummed. Please reach out if you have a color variant you would be willing to part with.
I am a simple Judy/V stan. Been enjoying other's works for a while now and wanted to share something of my own. Check it out on AO3: https://archiveofourown.org/works/30104145/chapters/74148798
Reminder: this is title syntax, it'll work anywhere in Foobar that accepts title formatting.
I use the Discord Rich Presence component and I'm the kind of audio geek/snob that likes to show off not just what he's listening to, but what it's sample rate is. But Foobar's lack of floating point math makes this fun if you want to display 44.1khz or 88.2khz rather than just 44khz or 88khz. You also need to account for rates that aren't divisible by 1000 perfectly and the length of the string above 100khz and...yeah...here:
$ifgreater($mod(%samplerate%,1000),0,$insert($div(%samplerate%,100),'.',$sub($len($div(%samplerate%,100)),1)) kHz,$div(%samplerate%,1000) kHz)
Basically if the mod of the samplerate divided by 1000 is anything but 0, then divide the sample rate by 100 but drop the "." at one less than the stringlength of samplerate/1000. Otherwise just divide that sucker by 1000 and spit it out.
If you're the kind of person that has a habit of listening to stuff with MHz sample rates; then perhaps you'll like my entire "Bottom Line" for the discord component:
$if($stricmp($right(%filename_ext%,3),iso),Super Audio CD - 2.8MHz,%codec% - $ifgreater(%samplerate%,1000000,$insert($div(%samplerate%,100000),'.',$sub($len($div(%samplerate%,100000)),1)) MHz,$ifgreater($mod(%samplerate%,1000),0,$insert($div(%samplerate%,100),'.',$sub($len($div(%samplerate%,100)),1)) kHz,$div(%samplerate%,1000) kHz)) - %bitrate% kbps)
I'm looking for a little clarity on the bones of floating point files.
If I have a mix I like in my DAW but the main output is clipping and I don't want to adjust my levels so that it's not (for whatever reason), if I export a 32-bit floating point wav, is the clipping no longer an issue?
I know the converters would have been clipping so I would have heard that while mixing but if the bounce doesn't go through the converters does that fix/allow a work around for the problem?
If this is the case should tracks be delivered for mastering as 32-bit FP in general?
Any clear explanations floating around the web?
There’s a hunk of tree/rock floating in space and it has mechanical stuff built into it. At one point, you ski somewhere? And there’s hookah? It was all vibes. I found it on a random website back in the late 00’s that also had other flash games, videos, text posts, etc.
I have an hardware in which all the computations are done in fixed point arithmetic. Yet, the input is in floating point. Therefore, I am trying to convert floating point inputs to their corresponding signed fixed point representation. My first attempt to simply copy to_sfixed function from float_pkg provided by Vivado to my package. (I have also copied Class function to my package.) The function is provided below.
... keep reading on reddit ➡
-- purpose: Converts a float to sfixed function to_sfixed ( arg : float; -- fp input constant left_index : INTEGER; -- integer part constant right_index : INTEGER; -- fraction part constant round_style : BOOLEAN := fixed_round_style; -- rounding constant overflow_style : BOOLEAN := fixed_overflow_style; -- saturate constant check_error : BOOLEAN := float_check_error; -- check for errors constant denormalize : BOOLEAN := float_denormalize) return sfixed is constant fraction_width : INTEGER := -minx(arg'low, arg'low); -- length of FP output fraction constant exponent_width : INTEGER := arg'high; -- length of FP output exponent constant size : INTEGER := left_index - right_index + 4; -- unsigned size variable expon_base : INTEGER; -- exponent offset variable validfp : valid_fpstate; -- Valid FP state variable exp : INTEGER; -- Exponent variable sign : BOOLEAN; -- true if negative variable expon : UNSIGNED (exponent_width-1 downto 0); -- Vectorized exponent -- Base to divide fraction by variable frac : UNSIGNED (size-2 downto 0) := (others => '0'); -- Fraction variable frac_shift : UNSIGNED (size-2 downto 0); -- Fraction shifted variable shift : INTEGER; variable rsigned : SIGNED (size-1 downto 0); -- signed version of result variable result_big : sfixed (left_index downto right_index-3); variable result : sfixed (left_index downto right_index) := (others => '0'); -- result begin -- function to_ufixed validfp := class (arg, check_error); classcase : case validfp is when isx | nan | quiet_nan => result := (others => 'X');
I first heard of Floating Points on this sub (I believe JV said he was listening to the album Crush during or just before his AMA) and I had to plug the latest release here...an incredibly moving album.
Thoughts? I know there isn’t much in relation to boc but I figured lots of you have been listening to it and this is the only place on this site that I can trust to have good discussion.
Also curious if anyone knows some pressing info, I’m really kicking myself for not pre-ordering and it appears to be sold out now.
Thinking about jumping for a copy on discogs but I’m in the US and prices are pretty steep right now.
If you don't think jazz & classical can be ambient then you need to check this LP.
Floating Points on the controls, Pharoah Sanders on his horn - still going strong after all these years & The London Symphony Orchestra on everything else. A peace & love filled audio antidote to the angst of the last 12 months.