From: Martin G. <mar...@gm...> - 2025-04-25 00:25:12
|
On 25/04/25 00:12, Nicolas Graner wrote: > Thanks very much for your detailed answer. My pleasure. > I can wait for the next sox_ng release in May, no hurry. :) Thanks. I'm working on other stuff too, libdolbyb to put Dolby B encoding and, more importantly, decoding as many people have old archives but don't want to have to buy a more expensive Dolby B licensed deck and spend the money on better heads and electronics. It's working in sox_ng trunk now, giving bit-for-bit exact same output as the original Pascal. I was sniffing ReZound and may get around to it. It's original developer is still with us but is doing other things. He seems more interesting in me implementing or finishing things he was working on or had in his TODO list but he is responsive to technical questions so I consider myself to be blessed in this respect. It's a fab graphical audio editor whose only lack is a log frequency axis spectrogram display as an alternative to the waveform. Give me time :) My latest is maintaining Professor David Turner's interpreter for the seminal pure lazy functional programming language Miranda miranda.org.uk I was Prof. Turner's pupil, then colleague, then friend and translated the interpreter of his previous lazy flanguage, KRC, the modestly-named Kent Recursive Calculator, from BCPL-for-EMAS to C-for-Unix. in the 2000s he even regaled me damson, the Sparc SunStation on which he had developed Miranda in the 90's. I remember him bounding up the stairs to deposit damson in her new home and then being so terrified of the steep and narrow staircase to descend that I almost had to carry him down in my arms, bottle-bottom shortsighted that he was. codeberg.org/DATurner/miranda It now works on all off the GCC Compile Farm except for, paradoxically, all Sun Sparc/Solaris machines on which it dumps core as soon as you ask for 1 + 2. > comments to store application-specific data into MP3 files OK, I'll come clean. I broke ID3-through-LAME writing in 14.5 by including UTF16-writing code that didn't check the return values from id3_set_fieldvalue() or somesuch and, in fact, it was failing. I've reverted it and am working on a patch release for 14.5.1 but had to stop today when I found myself screaming at the screen. The pushers, putas and punkabestia under my window must have wondered who was doing what to me, but it was just me. It's OK in trunk, so grab that. > tell users to run sox_ng instead of sox Absolutely! Not only are all CVEs fixed and a slew of grave, noisy and talkative bugs,but it's now sprouting more stuff. The new ones are: * softvol, a volume control that can slowly increase the volume and, when it would have clipped, adjust the software volume control so it exactly peaks. The crunched I expected are totally inaudible, presumably because they coincide with the peaks of the sound wave and have latency 0 * a new combining method for sox synth that uses the synthesized wave to modulate a delay of any length. Well, up to a year, 15GB. * dolbyb for Dolby B noise reduction application and removal. It emulstes in software a real Dolby B unit's electronics, complete with the calibration procedure in which you have to earth certain points and adjust some pots until the voltage is right at some other point on the circuit. I'ts bang on, hats off, based on dolbybcsoftwaredecode by translating its Free Pascal into a C library. It doesn't work on bigendian systems yet but I think it's just a question of how it assembled and disassembles the samples when they arrive as bytes from a file. For your need case, I'd provide both a sox interface and an id3 interface that does the same thing. That way as working ID3 handling in SoX and id3 comes and goes your customers will have a working system. > Keep up the good work on sox_ng! You know where the cookie jar is. Offer me a beer :D M This message was composed with Thunderbird so, if the lines are backwards or one on top of the other, please apply the appropriate filters. |