New, simplified LoZ3 MSU-1 code

I'll post that volume thing at bszelda forum. I know that Ikari adjusted the timing in 1.7 but the hardware is limited. You cannot blame him to have too silent music output.

The results of my initial testing are looking really good. I still need to go through and re-normalize all of my PCM tracks to figure out where the sweet spot is, but I'm pretty much convinced that the SD2SNES MSU-1 volume boost (v1.7+ Press X>Configuration>Chip Options>MSU volume boost>+12dBFS) is sufficient to bring the SD2SNES's MSU-1 volume up to the same, or nearly the same volume level as higan, so you should NOT be using separate emulator/hardware patches anymore, and always use $FF as max volume. If the tracks are too loud at $FF volume, the track files themselves should be adjusted.

I definitiely won't adjust all pcm:s. If it gets confirmed, I will just set the volume of my patch to 60 without 2 separate patches , higan adjustments

I guess that works too... either way, it still achieves the same result of both higan and the SD2SNES using the same code and audio tracks.  Consistency is the real goal, regardless of how you get there. After playing around a bit more, I think I'm settling in on audio tracks normalized to -9dBFS, with SD2SNES MSU-1 boost set to +12dBFS (the +12 makes the SD2SNES play at the same level as higan, the -9 makes the PCM tracks the same volume as the original SPC).

Ok, so here are my current tests.  The opening title and overworld themes are MSU-1, the others are SPC fallback.  As you can tell, the volume levels are really close (though MSU-1 is a bit too loud in both videos, so I probably need to tweak that down just a bit).  Don't compare the videos to each other, what matters is each video relative to itself, compare the MSU-1 tracks vs the SPC tracks, and the MSU-1 track vs the SFX like the sword swinging.  This is with the same exact ROM and the same exact PCM files in both.

Also, forgive my lousy playing in the SD2SNES video, there are several seconds of display lag on the HD PVR 2, so I had no idea where I was the whole time...

higan:


SD2SNES:

I wait for the playtesters how it turns out with the bszelda pcm:s:
http://z9.invisionfree.com/bszelda/index.php?showtopic=1572&st=15&#last

if it will be tested successfully, there's a bit of work done (my alttp, smw patches, all must be updated, also re-upload at romhacking all that f* no one is eager to do ).
I hope that was the last change of policies. I didn't update the alttp and smw with higan 0.97 instructions either

The message I wrote was before I saw your videos. Yes, looks like the volumes are pretty much the same  
Though I still want to wait for the bszelda results before changing anything.

I think I've settled on -10dBFS for the track normalization.  So, if you're not going to re-normalize your audio, then that actually works out really well, because -10dBFS = half volume, so that should work perfectly if you use $7F as your full volume and $3F as half volume (or $80/$40 if that works out easier for your fade routine).

Edit: I may have done the math wrong on that, it's been awhile since I've done anything in log scale. >_<

Would you be interested to shortly test bszelda? Somehow my call at the bszelda forums didn't attract playtesters

It's relatively simple to test:
1. Download the rom:
http://bszelda.zeldalegends.net/zips/pbszelda.zip

2. Download the patch folder:
http://bszelda.zeldalegends.net/zips/m1_msu.zip

3. Download the pcm files (2 GB):
https://onedrive.live.com/redir?resid=90BB3EA2AF91197A!566&authkey=!AGZXTrLsfFjjQgM&ithint=folder%2czip

4. Apply the patch "m1_msu_video.ips" to the rom bszelda_map1.smc

5. Apply the patch "m1_bsnes_higan_compatibility.ips" from the subfolder bsnes_higan_compatibility to the rom bszelda_map1.smc

This step 5 sets the $2006 to volume level 60. Then it is just a test like you did with the ALTTP videos above whether the sfx are covering the msu1 on sd2snes (too silent). You must wait until minute 7 before you can move and check the difference between sfx (sword throwing etc...)

I can probably do a short test, with recordings, but I doubt I'll be able to do a full playthrough.  I've been having trouble sleeping lately, and it's been absolutely destroying my ability to get anything useful done, plus I'm getting kind of burned out on the whole Parallel Worlds v1.2 project.  Right now it's coming in at just under 10,000 lines of asm (though, to be fair 60-75% of that is the FastROM patch).  I'm almost completely done with it, but now comes the far less glamorous and more tedious parts, like script typos and checking all 300 rooms for glitched doors like the one in the guardhouse.

*lol, of course no in-depth test is needed, just a short test like you did on ALTTP, how the bsnes setting of volume 60 turns out on sd2snes firmware1.7. That's all, 3 minutes of slaying swords while there is talking and or msu music in the background (backbone music volume is lowered while anybody is speaking, so the voice must be clearly understandable with sword sfx and such).

And we're in no rush, get some sleep first, man!

Great job on PW btw. Thought you were already finished, as you do this ALTTP simplified code now, but I guess you were just burned out from PW and did this as distraction

I wrote this because I had a small feature I wanted to add to PW, but I couldn't figure out how to make it work with your code. I probably could have gotten it working, but instead I just got curious and decided to see if I could write my own.

Here you go.  Again, my Hauppague has something like 3+ seconds of lag for composite inputs, and I don't currently have a splitter setup, so there was no way I was going to be able to actually play through it. Same exact files in both videos.

higan v097


SD2SNES (MSU-1 volume boost +12dBFS)

Sounds pretty much the same, what's your opinion on this?

Do you have to manually switch on the msu1 boost or is it automatically switched on? It says on Ikari's page:

◦MSU1 volume boost: In case MSU1 audio is too quiet you can now increase volume in several steps. Available via Configuration->Chip Options.

So if you manually switch that on I must mention it in the readme

Guess I will have to change all my patches (renounce on the compatibility patch or leave it for firmware 1.6 and prior), for the bszelda as well as smw and alttp.

Thanks for the test

In my opinion, they are close enough to not require any separate patch. It's not 100% exact, but that's me being picky with a direct side-by-side comparison. Most people will only play one or the other, and those people will never know the difference. It's definitely not like before where the SD2SNES was less than half as loud, which was blatantly and immediately obvious. At this point I would consider the issue completely fixed.

Personally, I wouldn't even bother with compatibility patches, just tell people to upgrade their firmware, but that's just me. My next aLttP DX update is going to break the volume patch on RHDN anyway because I'll be using this MSU code instead, so it won't line up to the right addresses anymore. I'm not going to go to the trouble of fixing it, it wasn't even mine in the first place. I just spent an hour re-leveling my audio tracks, so everything works. I'm not going to go out of my way to cater to old broken firmware when we have a real solution now.

You do have to manually enable it in the SD2SNES, but it a persistent and global change, so you do it once and it's done. I've asked ikari if he would consider making it the default, but I don't know how often he checks comments on his blog, and don't know a more direct way of contacting him. Be sure to set it to the max +12dBFS setting, or maybe +9dBFS, but anything less and you might start to see a difference (personally, I think +10 might be the perfect level, and he should just use that and completely remove the config option, there's no need to keep the old wrong level around, but now I'm starting to sound like byuu...)

Alright, again many thanks for everything. Will update in the near future.

I pm:ed you the mail address of Ikari

I'm almost finished with the update. Just one further small question as I do not own a sd2snes and thus do not know the menu. If you klick Configuration->Chip Options... can you here adjust the volume to 10db?

Maybe you can send a screenshot or anything...

Edit: ok, updated all my msu1 patches (alttp, f-zero, smw, bszelda map1/2, ast). Default is now the low volume which can be increased with an optional patch to sd2snes 1.6 settings (FF). I hope this was the last time of changing policies, though I suggested the following:
http://board.byuu.org/phpbb3/viewtopic.php?f=8&t=1156

The boost option is increments of 3. For the sake of your readme, I'd say just turn it all the way to the max (+12). When I said 10 it was just me being really picky, but also it was kind of a complete guess, I would have to actually analyse the audio to determine it that finely. My videos were recorded at +12, might as well just keep it simple.

I think I leave it at 10 for now. Except you, the participation on my testing request at bszelda was rather frustrating. I do not have a sd2snes, and if people do not even like to test, I see no point satisfying the compatibility by guess. If they complain at 10, I will give 12 in the tutorial, but for now, I leave it at this value as recommend

10 isn't an available option, it increments by 3, so the choices are 3, 6, 9, 12. I only threw 10 out there hypothetically. That's why I'm saying 12 is simpler (just say turn it all the way up to max). Or maybe it's by 1.5, I don't remember... basically, either max or maybe one step less than max.

I see... it would really help if you'd send me a screenshot of the menu to avoid such misunderstandings.

I now wrote
Firmware v0.1.7 ff: choose Configuration-> Chip Options and set the msu1 boost to +12dBFS

e.g., I do not know what the FS shall be, but you e.g. used that in above post...

dBFS = decibels relative to full scale.  Commonly abbreviated to just dB unless you actually need to specify the difference, but it's labeled dBFS in the menu.

Press X>Configuration>Chip Options>MSU-1 volume boost

Awesome, thank you, this really helps. Will do the next update within the next days. I'll rewrite the bml being v97 compatible... byuus answer to my suggestions sounded somewhat he got pissed by all the suggestions. But there are only 5 coders maybe who can implement msu1. Though he finally gave tipps how to play the sd2snes files without renaming them in higan97, Changing policies all the time DOES NOT MAKE THINGS BETTER FOR US!
(small rant here from my side)

As far I can tell the difference between v97 and v94 bml is that he thought a child would fit better there than there. This makes the v94 incompatible with v97 bml.

So. I will delete v94 from my tutorials and rewrite the bml to be v97 compatible. If he again changes things in v98, I will say play bsnes v70 and won't support higan anymore
*rant over.

hex_user actually posted a solution to the bml issue. It's ugly and kludgy, but apparently you can write a single bml that is compatible with both versions of higan. That, combined with the track name specifier should pretty well cover things.

I will try to make that with the instruction by hex_usr. Do you know how I get the board node? Looks like some weird hash code.