05803: nemesis, nemesisuk, gradius, twinbee, konamigt, rf2, gwarrior: Nemesis/Bubble System games too dark

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
05803	Color/Palette	Minor	Always	Dec 18, 2014, 17:11	Jan 15, 2015, 23:37

Tester	NRS	View Status	Public	Platform	MAME (Official Binary)
Assigned To		Resolution	No change required	OS	Windows Vista/7/8 (64-bit)
Status [?]	Closed			Driver	konami/nemesis.cpp
Version	0.156	Fixed in Version		Build	Normal
		Fixed in Git Commit		Github Pull Request #

Summary	05803: nemesis, nemesisuk, gradius, twinbee, konamigt, rf2, gwarrior: Nemesis/Bubble System games too dark
Description	The Konami Bubble System games as well as the ROM version of Nemesis had always been way too bright in MAME, until in early 2014 when the analogue amplification circuit was added to video/nemesis.c. Unfortunately, now the games are somewhat too dark when comparing to video captures from the original hardware (e.g. ). Trying to understand the Nemesis schematic (http://tamdb.net/v2/index.php?action=downloads;sa=view;down=3996), it seems that there are two relevant resistors: a 1K Ohm pull-up resistor, and a 470 Ohm resistor (to ground). In video/nemesis.c in nemesis_state::create_palette_lookups(), there are these lines: static const res_net_info nemesis_net_info = { RES_NET_VCC_5V \| RES_NET_VBIAS_5V \| RES_NET_VIN_OPEN_COL, { { RES_NET_AMP_EMITTER, 1000, 0, 5, { 4700, 2400, 1200, 620, 300 } }, { RES_NET_AMP_EMITTER, 1000, 0, 5, { 4700, 2400, 1200, 620, 300 } }, { RES_NET_AMP_EMITTER, 1000, 0, 5, { 4700, 2400, 1200, 620, 300 } } } }; video/resnet.h indicates that the second resistor field (set to 0 in video/nemesis.c) is for a "pull-down" resistor. I'm not sure whether the 470 ohm resistor in the schematic really qualifies as a "pull-down" resistor, but changing the RES_NET_AMP_EMITTER lines to the following: { RES_NET_AMP_EMITTER, 1000, 470, 5, { 4700, 2400, 1200, 620, 300 } }, does produce brightnesses in the Gradius logo that match almost exactly said YouTube video. I have attached a zip file with screenshots that compare the output from stock MAME 0.156 and my suggested patch to the YouTube capture.
Steps To Reproduce
Additional Information
Github Commit

Flags	Verified with Original
Regression Version
Affected Sets / Systems	nemesis, nemesisuk, gradius, twinbee, konamigt, rf2, gwarrior

Attached Files	gradius pictures.zip (213,837 bytes) Dec 18, 2014, 17:11 Uploaded by NRS

No.11314 Tafoid Administrator Dec 20, 2014, 12:20 edited on: Dec 20, 2014, 12:22	From what data I have collected any Devs involved - this suggested change is wrong and only improves the screen completely by accident. The 470 ohm resistor you are looking at on the schematic definitely is not the "pulldown resistor" used in the resnet.c calculations. Overall brightness level is like overall volume level... trying to make it consistent across drivers is a bit beyond the scope of the project. That said, the same can be accomplished by increasing your Contrast Slider in MAME to 120%. Therefore, no changes will be made. Closing. If you need to discuss this further, find a forum that is frequented by Devs and work it out there.
No.11373 couriersud Developer Jan 15, 2015, 23:37	For the records: The 470 resistor is not a pull-down resistor. It acts as an emitter resistor in a emitter-follower configuration. For the non electronics-savvy: This translates into a "more-less-than-more" deduction of 0.7 V (maxed at 0) of the signal level. This is reflected by using RES_NET_AMP_EMITTER. In the end, the developers couldn't care less about the video level. Every arcade monitor had potentiometers to adjust the amplification on a per channel basis. If anybody thinks the brightness is not sufficient: Don't bother trying to modify the driver. Do it like the ol' operators. Use the monitor adjustments.

No.11314

Tafoid

Administrator

Dec 20, 2014, 12:20

edited on: Dec 20, 2014, 12:22

From what data I have collected any Devs involved - this suggested change is wrong and only improves the screen completely by accident. The 470 ohm resistor you are looking at on the schematic definitely is *not* the "pulldown resistor" used in the resnet.c calculations. Overall brightness level is like overall volume level... trying to make it consistent across drivers is a bit beyond the scope of the project. That said, the same can be accomplished by increasing your Contrast Slider in MAME to 120%.

Therefore, no changes will be made. Closing.

If you need to discuss this further, find a forum that is frequented by Devs and work it out there.

No.11373

couriersud

Developer

Jan 15, 2015, 23:37

For the records: The 470 resistor is not a pull-down resistor. It acts as an emitter resistor in a emitter-follower configuration.
For the non electronics-savvy: This translates into a "more-less-than-more" deduction of 0.7 V (maxed at 0) of the signal level.
This is reflected by using RES_NET_AMP_EMITTER.

In the end, the developers couldn't care less about the video level. Every arcade monitor had potentiometers to adjust the amplification on a per channel basis. If anybody thinks the brightness is not sufficient: Don't bother trying to modify the driver. Do it like the ol' operators. Use the monitor adjustments.