ChanServ changed the topic of ##yamahasynths to: Channel dedicated to questions and discussion of Yamaha FM Synthesizer internals and corresponding REing. Discussion of synthesis methods similar to the Yamaha line of chips, Sound Blasters + clones, PCM chips like RF5C68, and CD theory of operation are also on-topic. Channel logs: https://freenode.irclog.whitequark.org/~h~yamahasynths
<kode54> I'd appreciate other people's perspectives on things too
<kode54> I wouldn't tell them that "LGR did it" but instead "I'd appreciate your perspective on this subject, and probably a lot of other people would, too"
<Lord_Nightmare> its kind of like how some people get fed up with certain youtube personalities' lack of technical research sometimes
<Lord_Nightmare> theres definitely different grades of quality
<Lord_Nightmare> and different grades of presentation
<cr1901_modern> https://twitter.com/cr1901/status/1115037726494068743 I think this ultimately sums it up. I'm sad I missed out, and yes I do like sharing information about topics that I don't think most ppl know about.
<cr1901_modern> Call it a character flaw lol
<cr1901_modern> And there's a line between "liking to share knowledge you don't think most other ppl know" and "wielding knowledge like a cudgel for the sake of asserting superiority in a niche community", and I'm not exactly sure where that line begins and ends
<cr1901_modern> >its kind of like how some people get fed up with certain youtube personalities' lack of technical research sometimes <-- this is why I can't watch 8-bit guy
<cr1901_modern> I objectively know that almost nobody watching his vids will give a shit about pedantry, and if I were to go comment on his vids, I would- rightfully so- be called out on said pedantry as wanting to show off.
<cr1901_modern> mmm okay, "almost nobody" is too strong. The rest still stands tho
<Lord_Nightmare> some of his videos are pretty decent, but his video about speech synthesis was really poorly researched in my opinion
<ej5> yeah i think he's aiming for more mainstream techy audiences
<Lord_Nightmare> he doesn't seem to understand the difference between speech synthesis as a way of producing pre-analyzed speech vs unlimited text to speech where the speech is fully synthesized from arbitrary user supplied text with prosody etc
<cr1901_modern> I'm not a fan of the "how sprites work" video
<Wohali> i don't watch any of that crap, but i think all it takes is a statement by whomever "i'm no expert on this, if you want to learn more, go see these other vids"
<Wohali> but humility is rare in 2019
<Wohali> just look at most politicians.
<ej5> heh
<Lord_Nightmare> how sprites work would be really interesting if done right, starting with pong and its shift counters
<Lord_Nightmare> which are the earliest 'sprites', unless you count the analog ball in tennis for two and the digital ships in spacewar
<Lord_Nightmare> but spacewar used a framebuffer
<Lord_Nightmare> well, a vector buffer
<Lord_Nightmare> so it isn't exactly sprites
<ej5> there's a neat venn diagram though. two circles: people who understand sprites and the history thereof, and people who are good at making well-produced youtube videos. these circles don't overlap.
<Lord_Nightmare> the ball in tennis for two is an analog sprite of sorts
<Lord_Nightmare> the paddles and ball in pong i'd consider the earliest 'really modern' sprites, even if they're squares
<Wohali> ej5: again, humility. personalities need to interview the others and do joint videos. i've seen some on this (in way other fields, like math or history)
<Lord_Nightmare> the first arbitrary sprite would be the ufo and spaceship in computer space
<ej5> isn't it just a circle generator (tennis for two)
<Lord_Nightmare> done with diode roms
<Lord_Nightmare> ej5: it might be, in which case, its not really a sprite, but a precursor to a sprite
<ej5> what about spacewar on the pdp1?
<ej5> technically vector, but still
<Lord_Nightmare> spacewar was vector
<Lord_Nightmare> i guess those are sprites, but they're not raster sprites
<ej5> do we even know why they're called sprites?
<Lord_Nightmare> pong and computer space i'd say was the first raster sprites?
<Lord_Nightmare> I do not
<Lord_Nightmare> I have no idea why they're called that
<cr1901_modern> Wohali: "but humility is rare" niche community drama in a nutshell?
<Lord_Nightmare> and i'm not afraid to admit it
<ej5> huh, wikipedia claims it was the ti 9918 vdp
<Wohali> cr1901_modern: indeed. niche communities are deeply guilty of a) personality worship, and b) loving being a big fish in a small pond
<Lord_Nightmare> the noughts and crosses symbols on the manchester SSEM though i do not consider sprites
<Lord_Nightmare> they were dots on a williams storage tube, effectively memory and framebuffer and display all at the same time
<cr1901_modern> Wohali: I'm not innocent either... I mean I'm kinda sad I missed out on a piece of the vintage computing pie and now basically 2 people are seen as the end-all of all vintage knowledge
<cr1901_modern> and shit I could use w/ some free donations of vintage crap :D... it would go to a good home
<ej5> by dumb people lol
<Wohali> cr1901_modern: why you crave the fame? do you want to "hold court"? it's not like there's money in it
<ej5> chuckg on vcf is way more knowledgeable than those two on pre 1990s stuff
<cr1901_modern> ej5: Chuck Guzis is one of the smartest people I know on vintage stuff
<Lord_Nightmare> this reminds me: when is vcfed gonna fix https on their forums, which is still broken
<Lord_Nightmare> its 2019 ffs
<cr1901_modern> Wohali: It's human nature to want attention on a visceral level. This is why I think it's a character flaw
<ej5> don't they keep having issues with coin miners?
<Wohali> i always see that as vcf'ed
<cr1901_modern> The truth it I wouldn't handle the attention well
<Wohali> like it's a swear word
<ej5> ALSO i really enjoy being able to browse it on my amiga.
<cr1901_modern> Someone will come up w/ an SSL accelerator sooner or later to bypass the https problem
<cr1901_modern> the truth is*
<ej5> there's a port of openssl on the amiga, but none of the amiga browsers support it yet lol
<cr1901_modern> oh, problem "solved" lmao
<ej5> the best part: there's a browser that's slated to support it in the next few months. but it's a commercial product.
<ej5> an amiga web browser. commercial product. let that sink in.
<Wohali> lolol
<Wohali> well that company is making brand new A4000 mobos
<cr1901_modern> in the DOS world, I've basically concluded the most painless way to solve the https problem is to create a battery powered dongle running a C-friendly core (riscv? arm?) w/ a bearssl port and a handshake protocol over the LPT port
<ej5> why not proxy it
<Wohali> ^^^
<cr1901_modern> b/c I find it viscerally displeasing
<ej5> ? you can use mtcp and then proxy over your lan
<cr1901_modern> I... don't find it enjoyable to use a computer (read raspberry pi) several times more powerful to retrofit an old machine.
<cr1901_modern> I mean bad enough these machines have to ultimately talk to servers thousands of times more powerful :P
<cr1901_modern> Plenty of old 8-bit computers had coprocessors that were as powerful or perhaps even more so than the computer itself. Commodore disk drive is an example in the thread.
<ej5> don't forget the apple laserwriter.
<cr1901_modern> Oh, what's in the laserwriter?
<ej5> 68K running at a higher clock than the mac
<cr1901_modern> Hahaha awesome!
<cr1901_modern> ej5: Anyways, a CPU that's slightly more powerful doing work ill-suited to the 8-bit computer as a coprocessor is an aesthetic (tm) I like.
<Wohali> i mean, unless you're browsing to another machine on your local network, you're using ghz processors all along the network to wherever else
<Wohali> and even then your switch is probably running at ghz
<Wohali> this is inches away from period capacitors :P
<cr1901_modern> That's a terrible idea
<Wohali> imo, scsi2sd is a necessary evil. having lived with oodles of dying hdds, floppies, etc. relying on them 100% of the time is asking for data loss, possibly permanent
<ej5> yeah...
* ej5 still uses floppy disks though
<ej5> mostly just for data transfer
<cr1901_modern> I mean sure, eventually all my hard disks will die. And since nobody is probably ever going to make an open source MFM drive, XTIDE and SD cards will be a concession I have to make
<ej5> hmm, an open source MFM drive...
<cr1901_modern> I would be _very_ surprised if someone had access to equipment to make one of these
<Wohali> home floppies
<ej5> hardest part would be making the platters and heads i think.
<cr1901_modern> r/w head would be hell to make
<cr1901_modern> platters prob less so
<ej5> platters would be reasonably doable, but the head would be quite tricky
<Wohali> i mean, home vinyl is possible. maybe we need vinyl based storage.
<cr1901_modern> home floppies I think someone will eventually do lol
<Wohali> you could replace the head with a magnetoscope instead
<Wohali> if the density is low enough you can just view it directly
<ej5> hmm, or some sort of thin film design maybe
<cr1901_modern> but isn't vinyl kinda... write-once?
<ej5> sintered ferrites would be a challenge.
<Wohali> i know LCM is using that for media retrieval
<Wohali> cr1901_modern: ;) low-tech CDROM!
<cr1901_modern> oh right, open CD-ROM manufacture would be interesting
<cr1901_modern> and not the burning method while fools CD-ROM drives, but actually _making_ the pits
<ej5> i should dig through some of my old hard disk heads
<ej5> maybe some of the 60's stuff could be duplicated
<cr1901_modern> time to dissect it?
<ej5> i'm not cutting up any of those lol
<cr1901_modern> ej5: You have access to the RAMAC at all? I think some of the designers are still alive
<ej5> i'm related to one of the designers :P
<ej5> and i have a few RAMAC heads :P
<cr1901_modern> Ahhh no kidding!
<ej5> sadly he passed away a few years ago, but his boss is still around IIRC
<cr1901_modern> That makes things easier. Just ask/figure out what the heads were made out of and perhaps if any documentation on how they were constructed survives LOL :P
<cr1901_modern> "Just"
<ej5> i bet the CHM has a bunch of the docs
<ej5> but those early heads were pretty primitive and required compressed air
<ej5> i'll have to tweet some photos at some point
<cr1901_modern> looking forward to it
<futarisIRCcloud> 5MB of data in a one ton drive.
<cr1901_modern> Now all we need is an open source "my first fission reactor" kit and "my first IC fabricator" kit :3
<futarisIRCcloud> http://sam.zeloof.xyz/
<cr1901_modern> yup I know about him, but that's not a kit :D
<ej5> hmm
<cr1901_modern> (not yet anyway...)
<ej5> you know, this may not be as hard as i thought
<ej5> a preexisting ferrite core, with some modifications, could work
<ej5> basically you'd take a toroid core and cut a slit in it
<cr1901_modern> think this is how floppy heads work?
<cr1901_modern> at least Shugart's internal manuals document it like a "toroid with a slit"
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ej5> realistically it'd be easier to refurbish MFM drives
<cr1901_modern> right... that would be an interesting hobby
<ej5> basically get a pile of broken drives, maybe have substitute electronics designed for them, ready to go
andlabs has joined ##yamahasynths
<ej5> but find the ones that crashed and repair them
<ej5> platters could be refinished and new oxide spun on
<ej5> heads, if in reasonable shape, could be polished again
<cr1901_modern> primitive enough to be repairable for future generations to struggle w/ g=c000:0005 :D
<cr1901_modern> (fun fact, doesn't work on the original IBM controller)
<ej5> the xebec controller is weird
<cr1901_modern> indeed :D... I wonder why the z80 is there
<cr1901_modern> it's clearly not doing much
andlabs has quit [Ping timeout: 245 seconds]
<ej5> hmm, a challenge would be finding the right oxide formula and thickness
<futarisIRCcloud> We need to recreate https://www.amazon.com/Inside-Personal-Computer-Illustrated-Introduction/dp/0896595048 and the usborne books.
<cr1901_modern> REing formulas seems to be a whole different animal compared to code
<cr1901_modern> We can't even recreate damascus steel* or the coke formula
<cr1901_modern> * (Sarayan has a good counterpoint to this)
<ej5> take an old platter, section and polish, take a look under a SEM, figure out the oxide thickness
<ej5> not sure about the formula. i'm no good at chemistry
<cr1901_modern> I remember that a mole is a very large number, and something about sigma and pi bonds
andlabs has joined ##yamahasynths
l_oliveira has quit [Quit: ChatZilla 0.9.92-rdmsoft [XULRunner 35.0.1/20150122214805]]
ej5 has quit [Read error: Connection reset by peer]
_whitelogger has joined ##yamahasynths
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
nukeykt has joined ##yamahasynths
<cr1901_modern> ValleyBell: How does libvgm-emu (sic) ultimately end up using "rwfuncs" function pointers? https://github.com/ValleyBell/libvgm/blob/master/emu/EmuStructs.h#L72 In particular, why is there a need to store them as an array of structs w/ function pointers?
<cr1901_modern> As opposed to having a single "read" and "write" function per core that the rest of libvgm-emu calls
<cr1901_modern> I see RWF_READ, RWF_WRITE, RWF_QUICKREAD, etc. And not all cores implement more than RWF_{READ, WRITE}
<cr1901_modern> does that mean the abstraction over all the cores chokes if it needs, e.g. a "RWF_QUICKREAD" function then and there, but the core doesn't support it?
<cr1901_modern> Okay I think I see... the player is expected to request the r/w functions it needs: https://github.com/ValleyBell/libvgm/blob/ec936e12befc8ab8b6974db83e02a8912a235893/emu/SoundEmu.c#L387
<cr1901_modern> ValleyBell: Better question- is there an absolute minimum number of r/w functions that a libvgm core _must_ support? from the list of RWF_* given in EmuStructs.h?
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
andlabs has joined ##yamahasynths
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
andlabs has joined ##yamahasynths
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nukeykt has quit [Quit: Page closed]
andlabs has joined ##yamahasynths
l_oliveira has joined ##yamahasynths
<Wohali> cr1901_modern: PCBs arrive, more to come
<Wohali> they look good!
<cr1901_modern> Wohali: Beautiful, thanks!
cr1901_modern has quit [Ping timeout: 245 seconds]
cr1901_modern has joined ##yamahasynths
<Wohali> cr1901_modern: i should be able to get shipping estimates to the 4-5 countries people were interested in and get the shopping cart up this week.
<cr1901_modern> you are amazing :D!
ej5 has joined ##yamahasynths
<whitequark> ej5: cross section requests: quartz crystal and ceramic resonator
<whitequark> esp the latter
<whitequark> i have no idea what's inside
<ej5> hmm, ceramic resonator could be tasty
<Ultrasauce> oh youre tubetime. didnt realize you were here
<Wohali> ...is that a fertilisation joke?
<Ultrasauce> get your mind out of the gutter
<ej5> probably, who knows
<Wohali> :/
<Wohali> we're talking about you, not me.
<Ultrasauce> (yes)
<TD-Linux> I tore down a quartz crystal previously, it's not great for cross sectioning, but great to look at
<ej5> neat, that looks like an oscillator, not just a crystal
<TD-Linux> yes, that's the sot23 chip
<ej5> the little sot23 is the oscillator/buffer device
<TD-Linux> (it came off due to the heat of cutting the case open)
<ej5> an hc49 crystal looks similar. it's why i hesitate to cross section it, there's a good chance if i do it crosswise that the crystal will shatter
<ej5> maybe if i do a light top-down cross section that doesn't touch the crystal itself...
<TD-Linux> if you go through one of the mounting springs it will also fall off
<TD-Linux> just cutting the can in half and viewing it from the side would almost count as a cross section
<whitequark> oh that's cool
<whitequark> the sot23 chip is just an inverter right
<Ultrasauce> were there capacitors on those pads or is it just using the parasitics for that particular xtal?
<TD-Linux> I think it has additional buffering
<TD-Linux> but yeah
<TD-Linux> Ultrasauce, I don't think there were caps there due to the lack of solder. there was ether a cap or pullup on the left
<TD-Linux> possibly for some cuts of crystal they would need them populated
<Ultrasauce> yeah
<TD-Linux> also some of the high frequency ones have internal plls, not cut one of those open
<whitequark> oh nice
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]