Sound and other improvements to Sega G-80 games. #7103

aaronsgiles · 2020-08-19T20:35:53Z

Sound and other improvements to Sega G-80 games: [Aaron Giles, couriersud]

Added netlist-based sound to Eliminator, Zektor, Space Fury, and Astro Blaster.
Split the Sega Universal Sound Board and Speech Boards into their own separate files.
Improved Universal Sound Board implementation for better accuracy in Star Trek and Tac/Scan.
Wrote netlist-based backend for Universal Sound Board; currently disabled due to limitations in the system.
Wrote netlist-based backend for Speech Board; currently disabled pending future sound system changes.
Implemented wait states and the vector DRAW flag to help improve timing.

SP0250 Improvements: [Aaron Giles]

Matched clock divider to real chip measurements.
Fixed behavior when not fed enough data; addresses "gapping" in speech in Sega games.
Implemented accurate LFR noise generator according to real chip measurements.
Added pulse-width modulation DAC output mode for future consumption by netlist.

Netlist additions: [Aaron Giles]

Added compile-time option to record nltool-compatible CSV files.
Improved CD4020 implementation.
Fixed CD4053 behavior.
Added 74139 device.
Added TL082 device.

8253 PIT changes: [Aaron Giles]

Added explicit synchronization to all writes.
Cleaned up some timing calculations to avoid double<->attotime conversions.

…n-nl2

…nnections. All sheet 8 sounds working.

… quickly.

…acfury.

…ently stalls on first noise source due to AFUNC interference.

* Removing devices is not a really good idea. Disabled for now. Perhaps should print some INFO.

…e is being output" This reverts commit 140a5be.

…s through speech board. Fix on/off controls.

* Moved netlist version behind an #ifdef for now, until some tricky issues can be resolved. * HLE code now runs at full fidelity, fixing Tac/Scan sounds.

* HLE shot sounds * add option to simplify SONAR sounds * updated documentation on current issues

* This is a short term solution. A NET_MODEL_OVERWRITE directive will be added to netlist later.

cuavas · 2020-08-25T08:23:51Z

src/devices/sound/sp0250.cpp

+		save_item(NAME(m_filter[index].F), index);
+		save_item(NAME(m_filter[index].B), index);
+		save_item(NAME(m_filter[index].z1), index);
+		save_item(NAME(m_filter[index].z2), index);


Why not use STRUCT_MEMBER to help save these and avoid clutter in the debugger?

Ok, didn't know about it and there's fairly few examples of it in the code.

cuavas · 2020-08-25T09:59:36Z

src/lib/netlist/nl_setup.cpp

+		{
+			// FIXME: Add an directive MODEL_OVERWRITE to netlist language
+			//throw nl_exception(MF_MODEL_ALREADY_EXISTS_1(model_in));
+			log().info(MI_MODEL_OVERWRITE_1(model, model_in));
+			m_abstract.m_models[model] = def;
+		}


Is this really a good idea? Being able to overwrite a model has potential to wreak havoc, and building things that depend on it sounds like a workaround for some other issue. I think @couriersud had it right previously in preventing models from being overwritten.

This is a commit from @couriersud directly as a workaround for limitations in the current model situation. I believe he intends to support a cleaner mechanism in the future.

cuavas · 2020-08-25T10:09:54Z

src/mame/audio/nl_elim.cpp

@@ -0,0 +1,1225 @@
+// license:BSD-3-Clause


Weren’t we trying to standardise on CC0 licensing for netlists? An actual description of the circuit doesn’t qualify for copyright protection. The only things that are potentially protected by copyright are the organisation/formatting of the file, and comments that contain sufficient creative work. @couriersud made an effort to get the netlists contributed by MAME developers made available under CC0 some time ago, but recently several have been added with BSD-3-Clause.

If you were, that's news to me, but I don't care as long as it's not GPLx.

cuavas · 2020-08-25T10:15:12Z

src/mame/audio/nl_segausb.cpp

+	// unconnected inputs to analog switch -- what's the right approach?
+	NET_C(GND, U8.2, U8.5, U7.2)
+	NET_C(GND, U15.2, U15.5, U16.2)
+	NET_C(GND, U30.2, U30.5, U31.2)


If they’re not connected to anything, they’ll look like a high-impedance path to the substrate, not a low-impedance ground connection.

cuavas · 2020-08-25T10:27:30Z

src/mame/audio/segag80.h

+	u8 m_lo_mask;
+	u8 m_hi_mask;
+	bool m_has_psg;


Why aren’t these const? They should only be set on construction, and bad things will happen if they get modified while running.

Good point. Also the netlist and scale factor.

cuavas · 2020-08-25T10:33:01Z

src/mame/audio/segaspeech.h

+
+#pragma once
+
+#define ENABLE_NETLIST_FILTERING (0)


This name is far too generic for a macro that ends up in a header that could get included before other headers.

Got it. It doesn't need to be in the header anyway so I'll just move it into the cpp file.

cuavas · 2020-08-25T10:34:08Z

src/mame/audio/segag80.h

+class segag80_audio_device : public device_t, public device_mixer_interface
+{
+public:
+	segag80_audio_device(const machine_config &mconfig, device_type type, const char *tag, device_t *owner, u32 clock, u8 lomask, u8 himask, bool haspsg, void (*netlist)(netlist::nlparse_t &), double output_scale);


This constructor should only be called by derived classes, it should not be public (and the class should possibly be called segag80_audio_device_base to indicate that it’s an abstract base, not a concrete device implementation).

That's fair.

cuavas · 2020-08-25T10:36:12Z

src/mame/audio/segausb.cpp

+#include "includes/segag80r.h"
+#include "includes/segag80v.h"


If you start needing the driver header in a device used in the driver, you’ve created circular dependencies. Is there a something you can do to avoid this?

They aren't needed; probably vestigial from when I split segasnd.

cuavas · 2020-08-25T10:39:50Z

src/mame/audio/segausb.cpp

+	for (int tgroup = 0; tgroup < 3; tgroup++)
+	{
+		timer8253 &group = m_timer_group[tgroup];
+		for (int tchan = 0; tchan < 3; tchan++)
+		{
+			timer8253::channel &channel = group.chan[tchan];
+			save_item(NAME(channel.holding), tgroup * 3 + tchan);
+			save_item(NAME(channel.latchmode), tgroup * 3 + tchan);
+			save_item(NAME(channel.latchtoggle), tgroup * 3 + tchan);
+			save_item(NAME(channel.clockmode), tgroup * 3 + tchan);
+			save_item(NAME(channel.bcdmode), tgroup * 3 + tchan);
+			save_item(NAME(channel.output), tgroup * 3 + tchan);
+			save_item(NAME(channel.lastgate), tgroup * 3 + tchan);
+			save_item(NAME(channel.gate), tgroup * 3 + tchan);
+			save_item(NAME(channel.subcount), tgroup * 3 + tchan);
+			save_item(NAME(channel.count), tgroup * 3 + tchan);
+			save_item(NAME(channel.remain), tgroup * 3 + tchan);
+		}
+		save_item(NAME(group.env), tgroup);
+		save_item(NAME(group.chan_filter[0].capval), tgroup);
+		save_item(NAME(group.chan_filter[1].capval), tgroup);
+		save_item(NAME(group.gate1.capval), tgroup);
+		save_item(NAME(group.gate2.capval), tgroup);
+		save_item(NAME(group.config), tgroup);
+	}


Why not use STRUCT_MEMBER for saving members of structures in arrays?

cuavas · 2020-08-25T10:40:24Z

src/mame/audio/segausb.cpp

+{
+	LOG("%s:usb_data_r = %02X\n", machine().describe_context(), (m_out_latch & 0x81) | (m_in_latch & 0x7e));
+
+	m_maincpu->adjust_icount(-200);


This should guard against disabled side effects.

aaronsgiles and others added 30 commits July 22, 2020 14:54

Added 74139 and TL082 devices to netlist.

5545c47

Add CD4011 and an attempt at CD4053 devices.

3156b88

Whoops, forgot to add 4053 to MAME.

0818fe1

Add CD4024 and 7407 devices.

c807415

Fixes to new devices.

ae61a86

Some fixes/improvements to the templated device.

1f199cb

Very preliminary Eliminator/Zektor netlist. Much debugging left.

b6c782e

Merge branch 'master' into aaron-nl2

b1f17bf

Remove a couple of hacks for now. Fix logging.

5edd59f

Merge branch 'master' into aaron-nl2

7308f91

netlist: Fix 4011 implementation.

05069fc

Merge branch 'aaron-nl2' of https://github.com/mamedev/mame into aaro…

d08dba2

…n-nl2

Merge branch 'master' into aaron-nl2

3decddf

Split Zektor from Eliminator. Document differences. Verify sheet 8 co…

7323040

…nnections. All sheet 8 sounds working.

All sounds mostly working. Hooked up PSG into netlist for Zektor.

2ec47e5

sega80v: Improved interrupt handling. Misc cleanups.

3ce8621

Add clipping diodes to torpedo sounds to allow them to retrigger more…

a2f40ea

… quickly.

Merge branch 'master' into aaron-nl2

1f61f90

Initial pass at Space Fury. Most sounds do something, but very slow.

4ed165f

Add clipping diodes to all 4011-based triggers. Complete review of sp…

b52d7c8

…acfury.

Minor optimization to remove sqrt from inner loop.

adc2973

Added generic netlist CSV creation. Define NETLIST_CREATE_CSV to enable.

d2454e1

Add message to static_assert

b7d3ed7

Better implementation of CD4053

6f9b54a

Fix incorrect sense in warning.

51a63fc

Split Sega USB from segasnd. Implement the backend as a netlist. Curr…

bd210ca

…ently stalls on first noise source due to AFUNC interference.

Didn't mean to add this.

8b38d2d

Fix up a few USB items.

1aed093

netlist: Disable dangerous optimization.

b7d6a6f

* Removing devices is not a really good idea. Disabled for now. Perhaps should print some INFO.

Add remaining Sega G80 sources to nl.lua

69f3a1d

aaronsgiles and others added 27 commits August 11, 2020 09:40

sp0250: Correct LFSR to hardware. General cleanup.

a884e52

segasnd.cpp: Boost volume out of netlist

c886a9e

segasnd -> segaspeech

aaec09b

netlist: append tag to name of csv files

771b547

netlist: Convert error to warning

cecf993

Revert "Netlist: add 'Q' parameter to VARCLOCK to indicate which phas…

a798391

…e is being output" This reverts commit 140a5be.

netlist: Add THRESH parameter to AFUNC

8f3f47b

segaspeech: Clear up schematics discrepancies. Route non-speech sound…

8792e10

…s through speech board. Fix on/off controls.

segausb/astrob: Latest WIP work

a948195

Merge branch 'master' into aaron-nl2

41d034e

Fix merge errors

7a8c30f

segausb: Remove fake 8253 attempt; it wasn't any better performing.

c15fd2e

segaspeech: Disable netlist filtering until new sound changes.

0d5fa42

segausb: Restore original HLE implementation.

2783d57

* Moved netlist version behind an #ifdef for now, until some tricky issues can be resolved. * HLE code now runs at full fidelity, fixing Tac/Scan sounds.

segag80v: Reduce wait states back to 2, which seems more plausible.

dc9e8d8

segag80v: Update flags

cae2b22

segausb: Reduce amplitude a bit

a96ef53

segag80v/segag80r: Cleaner implementation of waitstates

912038d

astrob: Sound improvements:

12c0c5f

* HLE shot sounds * add option to simplify SONAR sounds * updated documentation on current issues

Fix compile error

979d1a3

astrob: Some more optimizations.

7d05f24

astrob: Rebalance sound

e0363f6

netlist: Allow models to be overwritten.

59b2c7f

* This is a short term solution. A NET_MODEL_OVERWRITE directive will be added to netlist later.

astrob: use 12V minimum specs for CMOS circuits.

3f182fc

segag80v: Add model override for these netlists as well

af5ff17

astrob: Hook up MUTE. Couple of small optimizations.

03829e2

Fixed merge failure.

f8cc967

aaronsgiles merged commit f7b263d into master Aug 20, 2020

aaronsgiles deleted the aaron-nl2 branch August 22, 2020 22:05

cuavas reviewed Aug 25, 2020

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Sound and other improvements to Sega G-80 games. #7103

Sound and other improvements to Sega G-80 games. #7103

aaronsgiles commented Aug 19, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

aaronsgiles Aug 25, 2020

cuavas Aug 25, 2020

		#include "includes/segag80r.h"
		#include "includes/segag80v.h"

Sound and other improvements to Sega G-80 games. #7103

Sound and other improvements to Sega G-80 games. #7103

Conversation

aaronsgiles commented Aug 19, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment