I also second this. This is could be stupid OP in certain situations. Also on a personal note, we just got out of luclin which was bards just playing mana and not doing much else lets not go back please.
I assume it's some kind of special sauce bard code. People in the thread seemed pretty convinced that you could only get hit once per tick. Trying to cast it twice would result in phantom mana that the server would correct when you went to cast. It's just an interesting discussion that seems to have some merit.
Is it worth the risk to breaking other things? That's like taking a hammer to this problem. If this was an AK/in era thing maybe just leave it in and maybe more unboxed Bards will be born from it.
Should just disable the song if its that big of a concern. On a server where all lowbies are KEIed up anyway, I doubt anyone even uses it.
I don't know why the client was desyncing for those people in the ancient forum thread we were discussing. It seemed like they didn't find conclusive evidence of anything and were misunderstanding what was going on, but I'm certain that our version of the game expects the mod to apply to it. I assume that this is how it worked on AK too but I can't be sure, I don't think most people would think to even use this thing, but now that I mentioned it maybe some people will want to? It does seem lame in the same sense as mages putting mod rods on the ground like I mentioned before, if people just pump mana with their bard, especially considering that it doesn't tie up a person because extra bot accounts are free here. I can disable it I guess, though I am excited to find stuff like this and it was interesting to research it. There are lots of things like this in the game, things that are stronger than other things, and they were ok 20 years ago when people didn't know about them, or they didn't realize how good it was or something. But where we have wikis and forums and youtube and people share these things and optimize for speed runs, it can really cut into the nostalgia and recreation of the original gameplay. Charming is really OP here, it was eventually fixed, but we made it true to AK. I guess it's a judgement call to decide what's ok and what's bad and not everyone will agree. Anyway, if this is going to cause a problem I can just disable, it's just 4 (mostly unused) songs.
Sometimes I think there is an insult to the intelligence of EQ players from 20 years ago. "Players were too dumb to think of X back then" works for obscure things; but when it's something obvious that players would try then it's evidence in itself that it was not possible.
An excellent post. Is leaving things "as is" an option? Things seem to be pretty balanced at the moment.
This was my understanding of how the discussed mana songs worked: -Ervaj's could be modded, but had a hard cap on it. This cap fixed the modded effect regardless of AAs/equipment/puretone once a certain modifier value was reached. This hard cap existed for several other songs on AK, but I'm blanking on what they were. I remember one or two being obvious based on future scaling implications (I think the +resists on PoV? Maybe faction song?). -Cassindra's was a 7 mana/pulse song. Spinjas and I never got around to testing multiple bard stacking. It was towards the time I was becoming burned out. But, Spinjas testing definitively demonstrated that Cassindra's stacked with Chorus/Wind of Marr. It was not affected by any mod whatsoever. Regardless of equipment and AAs, Cassindra's was always 7 mana. -I didn't use insipid ditty with other bards and never tested it. I'm pretty sure on the vast majority of the VT raids I used it on, I was the only bard. That was the only context I bothered memming the song. I always used an instrument with it, though. All evidence pointed to it being modded based on the research I did way back then. But, I remember being told that NPC regen outpaced the song. Didn't stop me from being Sisyphus.
Been digging into Chorus of Clarity and mods today. I found multiple casual (i.e. they didn't run tests) claims that mods worked on it. However I also found another 'phantom mana' claim about it on Usenet, and Break's link is highly specific about mana desyncs with that song. So that's two sources claiming that the server is granting less mana than the client is estimating. I have a hard time believing that you could just cast the song twice per tick and get double the mana that way let alone get 27x2. But I'm not very familiar with bards. Players were certainly using the level 32 song back then so it's not like they abandoned it in favor of the Luclin and PoP regens and didn't notice how much mana Chorus of Clarity granted because they weren't using it. If mods did work then I would say it was an oversight and unintentional anyway as Sony explicitly coded the icon regens to not mod the mana and it makes no sense to have your level 32 song grant much more mana than the new PoP AoE.
Thanks for sharing your experience Delorne. How did you determine that there was a cap on lost composition? I found the logic in the client code, so at least my findings line up with yours on that. I don't know how to explain the conflicting reports about the level 32 song. Did anybody personally experience phantom mana on AK when using a singing mod with these songs, or when pulsing them instead of letting them auto repeat? I pulsed this song in caster groups back in the classic days. If it didn't work, I had no idea, because the casters were happy when they saw the 'mind clears' message. I'm sure I can find some way to limit it to once per tick but from what I can tell these songs work the same way as the shaman canni spells, instant mana grant when it finishes casting. Anybody have more info on this? For insipid ditty, the stat debuffs are definitely modded but I don't think the mana effect is, because it's the same effect as what's in Chorus of Marr, just with a negative value. Just the fact that people are questioning this makes me think that there's more to it, and I found a way to disable the mod without desyncing the client, so I will do that for the level 20 and 32 songs. If we figure out more about it, it can always be revisited.
Even if 27 mana per pulse is AKurate, there's an AKuracy vs. balance question here. 54 stackable mana per bard per tick is way more gamechanging than the Mental Corruption bug, for example. Turning a full group of wizards into "necros with SoS" is... yeah, good. Interesting archaeology here regardless.
I agree that it's strong and would be worth using, especially if players know how it's strong already rather than having to stumble on it, but what's bugging me is that for something that was a commonly used song in the classic days, there isn't any definitive answer, we don't know if it was modded or not, there are people saying it both ways. It is certainly modded in our AK client, so whether an oversight or intentional, it does work as modded. I can't say the same thing about haste, for example, no matter how I mod that, the client is coded to cap it. I can't make Chorus of Marr's mana portion become modded even if I tell the client that it's puretone; it's only the level 20 and 32 songs that do this. So then the argument is that people don't remember, didn't know about it, or it just didn't work for some reason even if they did try, but I haven't seen anything definitive. I don't imagine anybody has a packet capture from playing this song with mods on AK or another pre-2004 server lol. So just like charm being strong and later adjusted, this was eventually fixed by converting these instant songs to regular buff songs which we know don't get mods, so if this really was moddable it was relatively short lived and it would have been over 20 years ago now (except AK). Torven found a patch note that the buff conversion happened in Nov 2003 around the time of LDoN, so shortly after our time line. For the past 20 years of EQ, mana songs have not been moddable AFAIK. Bards weren't even that popular back then because they mostly sucked during Kunark/Velious. I played bard on an off through the years and I always 'knew' that mana songs couldn't be modded, to stay in line with other similar effects from other classes. It never occurred to me that the old crappy song that doesn't even put up a buff could be better than the awesome combo regen buff songs, and even if it were it was nice not to be chained to the level 32 song anymore once there were new ones. There were no mods for singing in original EQ - in fact one of the ways they tried to fix the haste stacking problems early on is by making the level 10 haste song singing instead of stringed, because that meant it couldn't be modded [1]. Then the bard epic was created, and the idea was that it gave bards 80% more power to their songs, all of their songs including the ones that couldn't be boosted with anything else before. This obviously isn't how it worked because a lot of songs don't scale properly with mods but the bard epic was still good even if not as good as intended. So then singing stuff could be modded, but the previous dev assumption was that it couldn't. I think this is when mana song became moddable. The song would have gone from 7 to 12 which isn't crazy good; enchanters got Clarity II which is 11, doesn't require spamming the spell and can be applied to more than 5 players at once. Once we got into Velious the level 34 song appeared, which I think was the first combo regen song that had a buff icon. This didn't completely obsolete the level 32 song, but could be used alongside it which matched the new mana spells that enchanters got in Velious. I think that during this time, it wasn't so common to have an epic. There was not an easily available singing mod for the masses until Luclin, the Amplification song, and scouring forums you can find that people are claiming to get extra mana when amping their level 32 song. Maybe it was quietly nerfed with some crude hack even before the LDoN overhaul, during the introduction of the Luclin singing mods? So there is still the open question from Torven, does it really work to pulse the song and get the benefit more than once a tick because of how it's an instant song rather than a buff? The 3 second twisting is kind of core to the bard class, but other than this song, or applying an offensive debuff that was resisted, there isn't much reason to ever apply the same song twice in a row or more often than every 6 seconds like auto repeat would do. Another thought I had was that perhaps there was a bug - when casting the song the mod gets applied, but when it's auto repeated it does not, or vice versa. We had similar bugs here in the past, so that made me think Sony could have had a similar situation, and that could explain some of the conflicting and phantom mana reports maybe. Edit: I think the level 55 song predates the level 34 song and was added in Kunark not Velious, but works the same way and the level 34 one was a backfill for lower levels. [1] http://www.afterlifeguild.org/Thott/haste_stacking.php
A lot of posts I read that are lost to time and my bad memory. I think it was also a part of the excel sheet I was a big fan of back during my playing days. I never did any definitive tests on mana drains during my AK days. I don't know any of the historical stuff. All really cool to read, solar. I didn't play a bard until I started on AK in 2009 or 2010, so it's great getting this insight from someone who played one during the heyday of the game. There are a couple niche use-cases of re-applying songs rather than pulsing. Off the top of my head, Shield of Songs agro generation and Saryrn's Scream of Pain for non-DoT damage (back when I had more free time than sense, I did the math on optimal song twisting using Amp+Saryrn's+Bombastic). There are probably more, but my barding days are behind me. Those are definitely exceptions to the rule rather than class defining, but I have always felt bards were an exception to everything in EQ.
I don't really think there is any kind of once-per-tic cap on reapplying songs, but I was just trying to make up explanations for the phantom mana thing. I think this mana desync stuff was just some temporary problem that was eventually fixed and was not a permanent thing. Did anybody get mana desyncs like that on AK? I scoured Thott's site and didn't find anything about modding the level 32 song, but he mentions doing a Cantata, Cassindra, Cassindra twist and I don't think he would have done that if it wasn't effective.
by “mana desync” do you mean such as the blue mana bar not correctly displaying your actual mana value? Because this happened all the time on AK. Most often it was noticeable when you summoned your horse after being rezzed with old models, your mana bar would not move at all as you regen’d, only when you cast a spell would it instantly jolt and update to its true value (say, from 15m to 80m, etc.) Regarding cassindra’s, we had a bard or two (maybe Hyacinth?) who would play that in group with freshly rezzed players so i believe they knew it stacked and gave an extra boost, but can’t speak as to whether it was modded by items or AA.
Yeah, the horse mana desync came to mind for me too. I remember it happening when using a horse in old models. The client thought I wasn't mounted, so my mana should regen slowly; the server thought I was mounted and should regen quickly. Casting a spell or clicking an insta-clicky forced the client and server to resync, which visibly increased my mana bar. TAKP has, or anyway had (I haven't tested in a while), a similar mana desync with the SK Mental Corruption bug. The client thinks an MC proc should refill my whole mana bar, so it zings up to 100% very briefly then drops down after a quick resync with the server. I don't remember other mana desyncs on AK, but I barded rarely and badly there, and of course it's been a million years.
Now that you guys mention horses, this is something I worked on here with Luclin. We work around that issue by telling the client how much mana it has frequently, rather than letting it free wheel. There are several other bugs related to old/new model mixing, like the helmet appearances that I'm trying to find a workaround for. Mental Corruption is interesting, I think the spell is not correctly filled out by designers, or is meant for a later era client (this one: https://lucy.alkabor.com/spell_3613.html ). It has 128 for the formula, but they only go up to 123. The Macintosh client hits a default case and thinks it gives (128 * level + 50) = 8370 but it should cap at 100 and it isn't doing that, and I spent way too much time to figure out why. The windows client's game logic doesn't differ in many places but this is one of them; it knows the formula and only ends up with 390 before the 100 cap. There was no good way to do this and make it work nicely with both clients so this is a TAKP specific issue. AFAIK this bug with ignoring the cap was present on the AK server too and this was popular to use for infinite mana. The bard song has a cap too, but the mods are added after that. For other nerds like me, disassembly dump from the Macintosh client: Code: [ Only registered users can see the bbcode. Click Here To Register... ] There is a sign error here, [esi+edi+0x7a] contains the formula value, 128. The cmp instruction performs a subtraction and sets the sign bit which is tested by the js instruction following. Basically 128 is treated as a negative value and that is the error in the Sony code that made this return 8370 instead of 100 like they intended. Had the formula been 0-127 or had they used a word wide version of the cmp subtraction it would have worked. This is likely a programmer error and the compiler generated these instructions based on type casting in the C++ code.
That is interesting to me. Where is the difference being applied? In the custom eqw code that's incorporated? The base client before any dll mods was the downloaded one from the AK patcher. I remember building mine this way when the PC hack was available.
The windows client is not the 'real' client for AK/TAKP. It's really close in the game logic but not identical, it's a slightly different version, and I try to work around things or patch them with a hack on the windows client, never the Mac client since that's the one I consider the standard. It's likely that Hobart or whoever applied various patches to the Mac code, or that some things were fixed in their main code base after the fork, that didn't make it there. We know that he piecemealed the spell data because there are obvious copy/paste errors in it, and it includes many out of era spells that don't match the spell code, like this MentalCorruptionRecourse spell. So I wish I knew the full history but I am just guessing here. Edit: do you have the for-sure original files from the patcher? I can compare if you send me those.
Incidentally this was in the Dec 4 2002 patch notes: That's why I think AK may have forked late November 2002. That patch also fixed perma pac and Beastlord fizzles.
