Log in

View Full Version : madVR - high quality video renderer (GPU assisted)


Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 [584] 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329

cyberbeing
24th April 2015, 11:33
Out of curiosity which style of NNEDI3 is used with madVR currently (any of these, or something different)?
Going by MPDN terminology, madVR's implementation is most similar to 'Avoid Branches' (i.e. unrolled vector).

x7007
24th April 2015, 12:39
NNEDI3 still causing issues with potplayer , with subtitles or without. with subtitles it causing High Cpu Usage in pause mode randomly until I playback again and it lowers to 3% . and randomly if I pause too long or playback/pause fast the video freezes but the sound continues, if I try to pause and playback again Potplayer freezes and I need to close from task manager or hard close the player.


I7 3770K
970 GTX 350.12
Windows 7 x64
Potplayer 32bit Lav Filters 0.65 CUVID + Madvr 87.21

dansrfe
24th April 2015, 21:13
I've noticed that some of my files need to to have the black level adjusted to -16 and the white level adjusted to +20 in order to not appear washed out. However, madVR's OSD shows limited input range and the monitor I'm outputting to as well is set to PC Levels in the driver control panel as well as madVR settings. Is this a problem with the files themselves some subtle bug in madVR?

Nearly all other limited range files are rendered to PC Levels by madVR, as expected.

vivan
24th April 2015, 21:38
Hard to say, it sounds like a fullrange video that was not tagged (or tagged as limited range) since it works fine with other videos.
Can you post a sample of such file?

luk008
24th April 2015, 21:52
I have some doubts:

Should I use NNEDI3 (64 neurons) instead of Jinc3+AR for chroma upscale? Or NNEDI3 is only recommended for image doubling?

Is "change dither for every frame" recommended to be checked?

Should I use "reduce banding artifacts" even with 1080p videos?

Thanks.

Asmodian
24th April 2015, 22:34
I have some doubts:

Should I use NNEDI3 (64 neurons) instead of Jinc3+AR for chroma upscale? Or NNEDI3 is only recommended for image doubling?

This is easy, test it. Do you like the look more? Personally I don't use NNEDI3 for chroma.

Is "change dither for every frame" recommended to be checked?

Again, personal preference. I don't use it but it is the default. My theory is that LCDs (with their slow pixel response times) would rather not have pixel values change every frame if they don't have to but an OLED, plasma, or DLP, wouldn't care.

Should I use "reduce banding artifacts" even with 1080p videos?

Reduce banding artifacts on low does very little damage but does reduce banding. Even blurays have minor banding a lot of the time. Again this is personal preference but resolution is not what I would use to decide but rather how much banding is present, some 240p videos have no banding while some 4K videos have lots of banding.

Anime Viewer
24th April 2015, 23:08
I have some doubts:

Should I use NNEDI3 (64 neurons) instead of Jinc3+AR for chroma upscale? Or NNEDI3 is only recommended for image doubling?

Thanks.

Using NNEDI3 for chroma upscaling takes four times longer than Jinc3+AR. I recommend using Jinc instead of NNEDI3 from chroma if its only between those two. Having said that the effects of image doubling can be so effective in improving picture that you don't need to use something as powerful as Jinc. Many of the less taxing upscalers in combination with doubling will look just as good as Jinc plus doubling does. Experiment with some of the other upscalers in combination with doubling. The resources you save by not using Jinc and NNEDI3 for upscaling may be useful in bumping up other areas (like debanding or dithering).

dansrfe
24th April 2015, 23:15
Hard to say, it sounds like a fullrange video that was not tagged (or tagged as limited range) since it works fine with other videos.
Can you post a sample of such file?

Actually, I think it's the opposite problem. If you change the source levels via keyboard shortcut to TV double or triple expanded it looks better.

I think this mostly happens with YouTube videos streaming VP9 for HTML5 players.

mkv cut: https://www.dropbox.com/s/jexxo5u9m7zpp0o/cut.mkv?raw=1

I don't understand the root of the problem though. I thought files were only in TV or PC levels. Why are there even options for double/triple expanded TV -> PC levels?

vivan
25th April 2015, 00:10
Actually, I think it's the opposite problem. If you change the source levels via keyboard shortcut to TV double or triple expanded it looks better.

I think this mostly happens with YouTube videos streaming VP9 for HTML5 players.

mkv cut: https://www.dropbox.com/s/jexxo5u9m7zpp0o/cut.mkv?raw=1Here're histograms for some frames:
http://i.imgur.com/lnPbCVO.jpg
http://i.imgur.com/hoOthFY.jpg
http://i.imgur.com/mSKHPUg.jpg
http://i.imgur.com/Fi9HZ7k.jpg
http://i.imgur.com/YMbb9A3.jpg
http://i.imgur.com/NilPlQA.jpg
http://i.imgur.com/RzzsHP6.jpg
There's something strange near 16 (UPD: actually those are borders, with them cropped there's nothing near 16/235) but other than that - it's clear that video was expanded... But it was expanded by different amount in different scenes.
I believe that it is intentional and was made for a "vintage" effect.

Does H.264 version looks different?

I don't understand the root of the problem though. I thought files were only in TV or PC levels. Why are there even options for double/triple expanded TV -> PC levels?I think that they're for fixing problems introduced either by bad mastering (TV range video was treated as PC and was converted to TV range) or drivers problem. I've never seen such problems, though.

huhn
25th April 2015, 00:20
I think that they're for fixing problems introduced either by bad mastering (TV range video was treated as PC and was converted to TV range) or drivers problem. I've never seen such problems, though.

i run in such issue sometimes when i watch livestreams on twitch.tv.
so this feature comes in pretty handy.

dansrfe
25th April 2015, 00:33
Here're histograms for some frames:
http://i.imgur.com/lnPbCVO.jpg
http://i.imgur.com/hoOthFY.jpg
http://i.imgur.com/mSKHPUg.jpg
http://i.imgur.com/Fi9HZ7k.jpg
http://i.imgur.com/YMbb9A3.jpg
http://i.imgur.com/NilPlQA.jpg
http://i.imgur.com/RzzsHP6.jpg
There's something strange near 16 (UPD: actually those are borders, with them cropped there's nothing near 16/235) but other than that - it's clear that video was expanded... But it was expanded by different amount in different scenes.
I believe that it is intentional and was made for a "vintage" effect.

Does H.264 version looks different?

I think that they're for fixing problems introduced either by bad mastering (TV range video was treated as PC and was converted to TV range) or drivers problem. I've never seen such problems, though.

Hmm...the H.264 version is also the same. I thought they would be different.

h.264: https://www.dropbox.com/s/t18sj0nmzlagm51/cut2.mkv?raw=1

There's definitely a vintage effect however, overall the video still looks slightly more washed out then it should. Maybe it's just a production choice to even out all the scenes and the overcast sky.

StinDaWg
25th April 2015, 04:07
OK, last post in a row:
- Is Catmull-Rom + Activate Anti-Ringing Filter + Scale in Linear Light still the best Image Downscaling method for the best quality? Lanczos shows better stats - less aliasing, but more sharpness. I am not going to argue with experts who say Catmull-Rom is better. I just wonder whether that has changed.
- Is Catmull-Rom also the best Image Downscaling method when Image Doubling/Quadrupling is used?
I prefer Lanczos3 AR downscaling on HD/4K sources. Slightly sharper and less aliasing like you said.

sexus
25th April 2015, 05:17
ok , im confused , could anyone here that has experience with madvr settings, since ive always thought mitchell netravali+AR+SL was the best for downscaling , please post the complete proper upscaling/downscaling/doubling settings for 480p-720p content and for 1080p and upwards content, mind you i have a unlocked ocd titan gtx as most people already know , thanks


btw whats this story about the latest nvidia drivers magically boosting NEEDI performance exponentially allowing for people to go from 16 to 32 doubling easily now on 1080p without noticeable dropped frames , this sounds more like partially broken opencl hence the so called boost in NEEDI performance

QBhd
25th April 2015, 05:33
Just do the compares yourself.... Every system is different... trust your own eyes.

For myself this is what I do for 720p==>1024x768:

chroma upscaling: NNEDI3-64
image doubling Luma: NNEDI3-128
image doubling Chroma: NNEDI-32
image downscaling: Catmull-Rom AR LL

And yes, all of the above is used and to my eyes it's the best PQ I have ever seen for 720p content on my TV (Plasma circa 2007)

QB

sexus
25th April 2015, 07:04
every system is different , not that different id say , we all have our displays perfectly calibrated , we all use displays with proper black levels and so on not sure what so called difference there should be , considering im talking to the high end users here, i as well use a plasma since OLEDs arent at that point yet

btw what do you use for 1080p=>


and why use 128 Needi what kind of gpu do you use that could support such beast of a setting? without a ton of dropped frames

id say itd be more realistic to go with >

chroma upscaling: NNEDI3-32
image doubling Luma: NNEDI3-64
image doubling Chroma: NNEDI-32
image upscaling:jinc 3+AR
image downscaling: Catmull-Rom AR LL

Siso
25th April 2015, 07:54
I've been testing some settings in madvr and I found that when I use the windowed fullscreen mode(new path), I get average rendering times around ~8.00 ms and present 0.14 ms, when I switch to fullscreen exclusive mode(old path) I get average rendering time ~16.00 ms and present 0.01 ms, it looks kinda strange to me, and in the FSE old path I can't increase the number of the backbuffer it always stays 3...Any opinions will be appreciated :)

x7007
25th April 2015, 08:22
Does anyone use subtitles with NNEDI3 ?

Because the subtitles are causing tons of issues, randomly the subtitles are covered with coloured square or just cut on the first and the last words, like the words just cut them self in some places of the latter.

No one has ever countered issue with NNEDI3 with Nvidia ?
Jinc is working perfectly, but I really like NNEDI3, it has better quality.

michkrol
25th April 2015, 11:10
every system is different , not that different id say , we all have our displays perfectly calibrated , we all use displays with proper black levels and so on not sure what so called difference there should be , considering im talking to the high end users here, i as well use a plasma since OLEDs arent at that point yet

btw what do you use for 1080p=>

I know you were replying to QBhd, but I have to disagree with you that everyone here uses perfectly calibrated display and won't see a difference.

There are major differences between for example using a TV, a computer monitor and a projector for video playback.
Some of us are using older or cheaper hardware for various reasons and the little differences in setups add to different experience.
I have myself an old display that uses some "tricks" to get the 8bit color precision, hence I can't tell a difference between some of the madVR's settings which make a day and night difference for others.

As for the 1080p scaling settings, it would depend greatly on your target resolution. In my case, playing 1080p videos on 1080p screen means only chroma gets upscaled and that's it, so not much I could help you with.

As always it is best to chose the settings that you like. I'd say everything other than nearest neighbor scaling is a good setup, depending on the circumstances ;)
Does anyone use subtitles with NNEDI3 ?

Because the subtitles are causing tons of issues, randomly the subtitles are covered with coloured square or just cut on the first and the last words, like the words just cut them self in some places of the latter.

No one has ever countered issue with NNEDI3 with Nvidia ?
Jinc is working perfectly, but I really like NNEDI3, it has better quality.
Using MPC-HC's built-in subtitle renderer I have no issues on my Geforce 750Ti. Never had any.

That said, there were problems with Kepler GPUs.
Even if you're not on Kepler, I'd suggest you update your drivers to 350.12 and delete HKEY_CURRENT_USER\Software\madshi\madVR\OpenCL registry key
as mentioned by cyberbeing here http://forum.doom9.org/showthread.php?p=1718353#post1718353

If the problem persists some more information might help: what OS, player and subtitle rendered do you use? what subtitles format do you use (SRT/SSA/DVD's or BluRay's subs/...)?

QBhd
25th April 2015, 12:03
every system is different , not that different id say , we all have our displays perfectly calibrated , we all use displays with proper black levels and so on not sure what so called difference there should be , considering im talking to the high end users here, i as well use a plasma since OLEDs arent at that point yet

btw what do you use for 1080p=>


and why use 128 Needi what kind of gpu do you use that could support such beast of a setting? without a ton of dropped frames

id say itd be more realistic to go with >

chroma upscaling: NNEDI3-32
image doubling Luma: NNEDI3-64
image doubling Chroma: NNEDI-32
image upscaling:jinc 3+AR
image downscaling: Catmull-Rom AR LL

Well my system specs are listed below in my signature, and I push my GPU to the max with every profile for every different resolution I watch. And no there is not a single dropped frame during playback. For 1080p24 playback I can push this through my GPU (only 63% usage for this particular res.):

chroma upscaling: NNEDI3-64
image downscaling: Catmull-Rom AR LL

But as I said, why not just test this yourself... it's how I found my settings.

QB

x7007
25th April 2015, 14:48
I know you were replying to QBhd, but I have to disagree with you that everyone here uses perfectly calibrated display and won't see a difference.

There are major differences between for example using a TV, a computer monitor and a projector for video playback.
Some of us are using older or cheaper hardware for various reasons and the little differences in setups add to different experience.
I have myself an old display that uses some "tricks" to get the 8bit color precision, hence I can't tell a difference between some of the madVR's settings which make a day and night difference for others.

As for the 1080p scaling settings, it would depend greatly on your target resolution. In my case, playing 1080p videos on 1080p screen means only chroma gets upscaled and that's it, so not much I could help you with.

As always it is best to chose the settings that you like. I'd say everything other than nearest neighbor scaling is a good setup, depending on the circumstances ;)

Using MPC-HC's built-in subtitle renderer I have no issues on my Geforce 750Ti. Never had any.

That said, there were problems with Kepler GPUs.
Even if you're not on Kepler, I'd suggest you update your drivers to 350.12 and delete HKEY_CURRENT_USER\Software\madshi\madVR\OpenCL registry key
as mentioned by cyberbeing here http://forum.doom9.org/showthread.php?p=1718353#post1718353

If the problem persists some more information might help: what OS, player and subtitle rendered do you use? what subtitles format do you use (SRT/SSA/DVD's or BluRay's subs/...)?

I am on Maxwell 970.

Windows 7 x64
PotPlayer x32
Subtitle Output : Display on Overlay/VMR/EVR surface
Font Renderer : Vector text renderer - fast , modest quality recommended . ----- Should I use bitmap ? What is the difference between them I never understand.

Display bitmap subs outside vid.image X
Display text subs inside video image X

I load SRT subtitles
Do I need to use image doubling NNEDI3 to double Luma resolution Always - if upscaling is needed for 1080p or 720p ? do I need it at all or Chorma Upscaling is good enough compare to jinc 4 taps ?

Anime Viewer
25th April 2015, 15:06
Does anyone use subtitles with NNEDI3 ?

Because the subtitles are causing tons of issues, randomly the subtitles are covered with coloured square or just cut on the first and the last words, like the words just cut them self in some places of the latter.

No one has ever countered issue with NNEDI3 with Nvidia ?
Jinc is working perfectly, but I really like NNEDI3, it has better quality.

I haven't noticed any problems (like the ones you are reporting) with NNEDI3, XYSubFilters, and my Nvidia 680m GTX combination.

Are you using LAV hardware acceleration? (If so you could disable hardware acceleration in LAV video settings and see if that makes a difference).

What Nvidia driver are you using, and what OpenCL version does your system list it is using with your Nvidia GPU? (I'm using 350.12 with OpenCL 1.2. Some of the older drivers have 1.1 tied into them. If you're using an older driver which is using 1.1 try updating to 350.12 and OpenCL 1.2 and see if that makes a difference).


I load SRT subtitles
Do I need to use image doubling NNEDI3 to double Luma resolution Always - if upscaling is needed for 1080p or 720p ? do I need it at all or Chorma Upscaling is good enough compare to jinc 4 taps ?

Like usual on this forum opinions vary. In some cases it depends on your screen resolution, and how much of an upscale from the source material is being done. I feel doubling is a waste to use when upscaling 720p to 1080p, and doesn't provide any significant benefit. Likewise if your screen resolution isn't at least double your source file resolution you're probably not going to see a significant (if any) benefit.

x7007
25th April 2015, 15:14
I haven't noticed any problems (like the ones you are reporting) with NNEDI3, XYSubFilters, and my Nvidia 680m GTX combination.

Are you using LAV hardware acceleration? (If so you could disable hardware acceleration in LAV video settings and see if that makes a difference). - Doesn't make a difference

What Nvidia driver are you using, and what OpenCL version does your system list it is using with your Nvidia GPU? (I'm using 350.12 with OpenCL 1.2. Some of the older drivers have 1.1 tied into them. If you're using an older driver which is using 1.1 try updating to 350.12 and OpenCL 1.2 and see if that makes a difference).

Using 350.12 and had the same issues with all the drivers since 970

Like usual on this forum opinions vary. In some cases it depends on your screen resolution, and how much of an upscale from the source material is being done. I feel doubling is a waste to use when upscaling 720p to 1080p, and doesn't provide any significant benefit. Likewise if your screen resolution isn't at least double your source file resolution you're probably not going to see a significant (if any) benefit.

The answers

James Freeman
25th April 2015, 18:40
Any idea what Avisynth 64-bit version actually works with MPC-HC & MadVR 64bit?
There are a lot of them.

Why?
I need some basic Red, Green, Blue sliders for hue correction.
I can't find a shader for this nor does MPC-HC or MadVR has them.......hint.hint..

luk008
25th April 2015, 20:20
A simple question: I've read that the rendering time needs to be below 30ms for smooth playback. But if the time is above 30ms and there's no repeated or dropped frames I shouldn't worry about this, right?

Asmodian
25th April 2015, 20:35
Right, you need 30 ms for 30 FPS content but for 24 FPS content you only need below 40 ms. For 60 FPS content you need below 15 ms.

You need the rendering time to be a little below 1000/[frame rate].

toniash
25th April 2015, 20:48
Any idea what Avisynth 64-bit version actually works with MPC-HC & MadVR 64bit?
There are a lot of them.

Why?
I need some basic Red, Green, Blue sliders for hue correction.
I can't find a shader for this nor does MPC-HC or MadVR has them.......hint.hint..
perhaps this?// (C) 2011 Jan-Willem Krans (janwillem32 <at> hotmail.com)
// This file is part of Video pixel shader pack.
// This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 2.
// This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
// You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

// brightness, contrast and gamma controls for RGB
// This shader can be run as a screen space pixel shader.
// This shader requires compiling with ps_2_0, but higher is better, see http://en.wikipedia.org/wiki/Pixel_shader to look up what PS version your video card supports.
// If possible, avoid compiling with the software emulation modes (ps_?_sw). Pixel shaders require a lot of processing power to run in real-time software mode.
// This shader is meant to work with linear RGB input and output. Regular R'G'B' with a video gamma encoding will have to be converted with the linear gamma shaders to work properly.

// fractions, either decimal or not, are allowed
// RedBrightness, GreenBrightness and BlueBrightness, interval [-10, 10], default 0
#define RedBrightness 0
#define GreenBrightness 0
#define BlueBrightness 0
// RedContrast, GreenContrast and BlueContrast, interval [0, 10], default 1
#define RedContrast 1
#define GreenContrast 1
#define BlueContrast 1
// RedGamma, GreenGamma and BlueGamma, interval (0, 10], default 1
#define RedGamma 1
#define GreenGamma 1
#define BlueGamma 1

sampler s0;

float4 main(float2 tex : TEXCOORD0) : COLOR
{
float3 s1 = tex2D(s0, tex).rgb;// original pixel
s1 = s1*float3(RedContrast, GreenContrast, BlueContrast)+float3(RedBrightness, GreenBrightness, BlueBrightness);// process contrast and brightness on the original pixel
// preserve the sign bits of RGB values
float3 sb = sign(s1);
s1 *= sb;
return (sb*pow(s1, float3(RedGamma, GreenGamma, BlueGamma))).rgbb;// process gamma correction and output
}

jmartinr
25th April 2015, 22:12
Any idea what Avisynth 64-bit version actually works with MPC-HC & MadVR 64bit?
There are a lot of them.
This is avisynth+ r1576, installed this week. I use it to do some simple cropping through ffdshow. Works OK till now.

http://www.roelofs-coaching.nl//blobs/pictures/capture2.png

pokazene_maslo
27th April 2015, 21:20
i have a bug to report: switching 1080p video to fullscreen freezes the video, sound keeps playing but picture is frozen. Seeking 1080p video running at fullscreen also freezes the picture.
switching 720p video to fullscreen works without any problem!

my setup: win8.1 x64, MPCHC x86 1.7.8, latest LAV filters, ATI HD6950, 2560x1600 desktop resolution, LAV splitter and LAV DXVA decoder, jinc chroma upsampling 3taps+anti-ringing, image upscaling jinc 3taps+anti-ringing
the remaining settings are default

baii
27th April 2015, 23:46
i have a bug to report: switching 1080p video to fullscreen freezes the video, sound keeps playing but picture is frozen. Seeking 1080p video running at fullscreen also freezes the picture.
switching 720p video to fullscreen works without any problem!

my setup: win8.1 x64, MPCHC x86 1.7.8, latest LAV filters, ATI HD6950, 2560x1600 desktop resolution, LAV splitter and LAV DXVA decoder, jinc chroma upsampling 3taps+anti-ringing, image upscaling jinc 3taps+anti-ringing
the remaining settings are default
Try disable exclusive mode or lower frame to present for exclusive mode.

Sent from my 306SH

x7007
28th April 2015, 15:18
i have a bug to report: switching 1080p video to fullscreen freezes the video, sound keeps playing but picture is frozen. Seeking 1080p video running at fullscreen also freezes the picture.
switching 720p video to fullscreen works without any problem!

my setup: win8.1 x64, MPCHC x86 1.7.8, latest LAV filters, ATI HD6950, 2560x1600 desktop resolution, LAV splitter and LAV DXVA decoder, jinc chroma upsampling 3taps+anti-ringing, image upscaling jinc 3taps+anti-ringing
the remaining settings are default


This happens to me only when using NNEDI3, Jinc Chroma Upscaling works perfectly

pokazene_maslo
28th April 2015, 17:55
Try disable exclusive mode or lower frame to present for exclusive mode.

Sent from my 306SH

That solved the problem! Thank you very much!

o00o
28th April 2015, 20:24
I cannot possibly get NNEDI3 to work properly on my r290x in any driver besides 13.12, which offers flawless playback, for some very odd reason. My flicks are extremely choppy even if the Neurons threshold is set at its lowest setting. I've tried using Jinc @ 8 taps as a temporary solution, a task which my GTX580 can handle just fine with a mild overclock, but unfortunately, a noticeable amount of frames are still being dropped in the great majority of drivers I've tested. It goes without saying that a GPU packing twice the amount of transistors shouldn't pull a sweat handling this kind of workload. I would like to add that I am using MadVR in conjunction with SVP.

It would be much appreciated if someone can shed some light on the matter so that I can finally get to the bottom of this.

Asmodian
28th April 2015, 21:07
8 taps is too many for Jinc, 3 or 4 would be better. More taps does not mean better quality with Jinc, Lanczos, and Spline.

Unlike Nvidia's, AMD's drivers do a copy to system memory and back when using NNEDI3 (OpenCL). Do you have slow PCIe or is SVP using up all the bus'es bandwidth?

XMonarchY
29th April 2015, 00:04
That solved the problem! Thank you very much!

Looks like the new exclusive mode is bugged or needs updating...

leeperry
29th April 2015, 02:10
I cannot possibly get NNEDI3 to work properly on my r290x in any driver besides 13.12, which offers flawless playback, for some very odd reason. My flicks are extremely choppy even if the Neurons threshold is set at its lowest setting.
Exactly my experience, it's choppy-land >13.12 with a 7850 on W7SP1 x86 over PCI-E 3.0 even with 16 neurons.

You may wanna let them AMD know about it. (http://forum.doom9.org/showpost.php?p=1702655&postcount=27842)

PS: Oops, hopefully madshi could fix the broken link but apparently it's been broken on purpose, making this bug report essentially useless.

x7007
29th April 2015, 04:51
I've found the issue I had with NNEDI3 , Disabling the artifact removal - reduce banding artifacts fixes all the issues I had. video freeze, subtitle corruption and random continually frame drop

Shiandow
29th April 2015, 08:24
I've found the issue I had with NNEDI3 , Disabling the artifact removal - reduce banding artifacts fixes all the issues I had. video freeze, subtitle corruption and random continually frame drop

Have you tried enabling the trade quality for performance option "don't rerender frames when fade in / out is detected"?

LocutusEstBorg
29th April 2015, 09:27
It needs to output the frames right at the time it calculated, and not when the GPU and screen think the next one needs to go out.

I don't understand the problem. This is only required if it's interpolating frames, right? Why can't the renderer just output the original untouched frames at the frame rate of the movie? Why can't the renderer just assume that the refresh rate is exactly the same as the movie frame rate?

The whole point of G-SYNC is that the frame can be presented at any arbitrary point between the regular Vsync intervals. If the frame rate is low, the G-SYNC module internally internally clones the frames until the next frame is received from the PC. That's an abstraction that the renderer shouldn't care about.

nevcairiel
29th April 2015, 10:03
I don't understand the problem. This is only required if it's interpolating frames, right? Why can't the renderer just output the original untouched frames at the frame rate of the movie? Why can't the renderer just assume that the refresh rate is exactly the same as the movie frame rate?

But thats simply not how madVR is designed. Its designed to work against a fixed VSYNC, and use this VSYNC for perfect presentation. If it doesn't have a VSYNC to work with, it needs to implement a whole new logic.
And then the question is, whats there to even gain over a fixed VSYNC? The only single case would be true-VFR video, which is extremely uncommon as it is.

That also doesn't solve the problem that this method would not allow for presenting any frames in advance to accommodate for glitches. It would be entirely reliant on a accurate clock and perfect timing in the PC, which is something that sounds like it might glitch out all the time.

Its a much better idea to just use a fixed VSYNC for video. A 144Hz GSYNC screen can render at a multitude of refresh rates which work for all common video frame rates, so I don't even see the point of bothering with GSYNC anyway, especially considering its a gaming feature available over DP only, so TVs or Home Cinema projectors couldn't even benefit at all.

Ultimately, madshi decides of course if he thinks its worth trying to bother with it in madVR, but personally, I would rather have him work on so many other features. :)

x7007
29th April 2015, 14:27
Have you tried enabling the trade quality for performance option "don't rerender frames when fade in / out is detected"?


No mate, I don't like to trade quality for performance cause I don't have performance issue, just some kind of bug issue with something that is causing a issue.

What this should fix ?

Won't it cause other issues ? and when installing madvr there are default selected options in there, which one are those and should I keep using them ? cause I have nothing selected in the trade quality for performance.

Ha I read it now


Quote:
Originally Posted by generalmx View Post
Hmm, interesting. But I do think I've found a significant source of dropped frames that I think is just related to madVR: blending of high contrast / overexposed and normal contrast scenes. With my settings, every now and then, there will be a big string of dropped frames from rapid transitions between the different contrast levels, which I'm guessing is a side-effect of manipulating luma and chroma.

Try those scenes with disabled debanding. Then try with both "strength" and "fade in/out" set to the same level.

With "don't rerender frames ..." trade quality for performance option disabled, when madvr detects a fade in/out it drops frames and rerenders them using other debanding setting. If this works as intended you don't even notice it, but if you have maxed out settings and one fade in/out after another in the video you rerender frames almost constantly with different settings. That's why some quality trade-offs are enabled by default

In you previous post you have mentioned judder with some sources, are you sure it's not in the source? If your encode or deinterlacing goes wrong you will get duplicate frames, which means judder during pan scenes. Easiest to check by frame-stepping through the panning scene.


EDIT : It seems that it fixed that, with the option enabled and debanding enabled.

Should I even use the banding artifact option ?

EDIT : But the movie is still freezes with audio still happens

I have this error log before the freeze happens

The description for Event ID 13 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video7
NVRM: Graphics TEX Exception on (GPC 3, TPC 3): TEX NACK / Page Fault


The description for Event ID 13 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video7
Variable String to Large



EDIT : STILL Crashing,, I don't know what is happening, everyone says it's the driver, Overclocking or no overclocking Prime95 tested, no memory errors, just this error happens in movies with Madvr and NNEDI3. I don't know why.

kasper93
29th April 2015, 21:41
PS: Oops, hopefully madshi could fix the broken link but apparently it's been broken on purpose, making this bug report essentially useless.

Probably there is no hope anymore that someone will look at this test case... and fix it.

SecurityBunny
30th April 2015, 05:03
Two somewhat problematic bugs I've experienced.

This issue exists in all windows 10 nvidia dx12 display driver versions. (349.65, 349.90, & 352.63)

I have narrowed down what is causing playback problems with nvidia dx12 drivers on windows 10 when playing back content in MPC-HC with madVR.

Test environment.
Windows 10 Pro Insider Preview - Build 10074 (clean install)
Nvidia Display Driver - GeForce 352.63
GTX 780 Classified (Kepler)
Tested playback in native video windowed resolution and fullscreen mode upscaling to 2560x1440 (monitor resolution).
Tested numerous sources with matching resolutions multiple times to confirm whether or not playback is functioning as intended with that resolution type.

madVR v0.87.21 x64 (http://madshi.net/madVR.zip)
MPC-HC 1.7.8.162 x64 (https://nightly.mpc-hc.org/)
Exported MPC-HC & LAV Filters registry settings. (https://www.dropbox.com/s/mw8uyeesd4cvmu3/mpc-hc-settings.reg?dl=0)
Exported madVR registry settings. (https://www.dropbox.com/s/boienw2pbho6cf8/madVR.reg?dl=0)
Generated 3DLUT calibration file used, saved in main madVR directory. (https://www.dropbox.com/s/xw3im6qa3dl12aj/3DLUT.3dlut?dl=0)

Chroma Upscaling set to NNEDI3 64 neurons
640x480 (NV12, 8 bit, 4:2:0), playback is fine.
704x400 (NV12, 8 bit, 4:2:0), playback is fine.
704x480 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
704x528 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
720x480 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
720x480 (MPEG2, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
848x480 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
848x480 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
854x480 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".

Chroma Upscaling set to NNEDI3 32 neurons
1280x690 (h264, 10 bit, 4:4:4 -> Y410, 10 bit, 4:4:4), playback is fine.
1280x692 (h264, 10 bit, 4:4:4 -> Y410, 10 bit, 4:4:4), playback is fine.
1280x720 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), playback is fine.
1280x720 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1280x720 (h264, 10 bit, 4:4:4 -> Y410, 10 bit, 4:4:4), playback is fine.
1440x1080 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
1440x1080 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
1448x1080 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1520x1080 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), playback is fine.
1916x1036 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1916x1072 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1920x816 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
1920x1038 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1920x1080 (h264, 8 bit, 4:2:0 -> NV12, 8 bit, 4:2:0), black screen with it outputting in the corner "madVR reports: - resetting Direct3D device failed (8876086c)".
1920x1080 (h264, 10 bit, 4:2:0 -> P010, 10 bit, 4:2:0), playback is fine.
1920x1080 (HEVC P010, 10 bit, 4:2:0), playback is fine.

Chroma Upscaling set to Jinc 3 taps
All test samples used play back fine and as intended.

Using NNEDI3 on 'image doubling' to double Luma resolution and Chroma resolution, playback is fine in all resolution tests. The issue seems to be isolated around NNEDI3 being used in Chroma Upscaling.

This issue exists in all nvidia display drivers newer than version 347.88. (350.05 & 350.12, including all dx12 driver versions.)

Under rendering > general settings in madVR, with 'enable automatic fullscreen exclusive mode' unchecked to use the newer windowed fullscreen mode, the display driver crashes when the video player enters windowed fullscreen and tries to playback content. Enabling exclusive fullscreen mode, playback occurs as intended without any driver crashing. So it seems the windowed presentation path does not want to function with the latest nvidia drivers. Last functional driver version that worked with windowed fullscreen playback is driver 347.88.

ryrynz
30th April 2015, 05:15
You shouldn't really be posting anything about the state of Windows 10 and or it's drivers in the forum at this stage, not until it's released.

SecurityBunny
30th April 2015, 06:43
You shouldn't really be posting anything about the state of Windows 10 and or it's drivers in the forum at this stage, not until it's released.

Oh, alright. Just thought I would pass along my findings. The second issue is unrelated to windows 10 and dx12 drivers though. Seems windowed fullscreen mode broke in the latest windows 8.1 nvidia drivers.

Asmodian
30th April 2015, 07:03
This issue exists in all nvidia display drivers newer than version 347.88. (350.05 & 350.12, including all dx12 driver versions.)

Under rendering > general settings in madVR, with 'enable automatic fullscreen exclusive mode' unchecked to use the newer windowed fullscreen mode, the display driver crashes when the video player enters windowed fullscreen and tries to playback content. Enabling exclusive fullscreen mode, playback occurs as intended without any driver crashing. So it seems the windowed presentation path does not want to function with the latest nvidia drivers. Last functional driver version that worked with windowed fullscreen playback is driver 347.88.

I am running Nvidia drivers 350.12 with a GK110 and Windows 8.1 x64 and I always use "fullscreen windowed mode [new path]" without any issues.

yuriks
30th April 2015, 09:03
I'm sorry if this has already been reported, I looked back a few pages and didn't spot anything relevant. EDIT: Oops, looks like I should've gone one page further back.

I'm getting artifacting on video playback when using NNEDI3. It manifests itself in the form of rows/columns of black dots, in different positions in each frame. The frequency of the problem seems to scale with neuron count: at 16 neurons there's barely any (one every few frames), with 128 neurons the image is practically full of them. You can check a screenshot here: [Large PNG] (http://i.imgur.com/WOPbqEf.png)

Disabling NNEDI3 makes the problem go away entirely. The screenshot is from a 720p source, but the same happens on 480p videos.

This didn't happen in the past, and is not related to a madVR version upgrade (I was using .13 before, I updated to check if the issue was fixed.) I also used FurMark to check for memory corruption and there didn't seem to be any issues. All signs indicate that the culprit is me updating the nvidia drivers to 350.12 a few days ago.

It is very likely that this is a driver regression, and so there's not much madshi can do about it, but I thought I'd post anyway in case you want to look into it.

Windows 8.1 x64
Playback setup from CCCP + madVR + XySubFilter
madVR v0.87.21
Nvidia Geforce GTX 770 2GB
Nvidia drivers 350.12

(Ask away if I forgot any relevant info!)

ryrynz
30th April 2015, 09:13
I'm sorry if this has already been reported

It has been, revert to 347.09 or earlier or select a different Error Diffusion mode other than Option 1.

yuriks
30th April 2015, 09:18
@ryrynz: This happens no matter what dithering method I choose. Even None. Reverting would be inconvenient for me, so I'll have to do without NNEDI3 for the time being I suppose. :(

sneaker_ger
30th April 2015, 09:27
Does this help?
http://forum.doom9.org/showthread.php?p=1718353#post1718353

yuriks
30th April 2015, 09:33
@sneaker_ger: It does! Thank you for linking that post. At least I have a workaround for the time being.

Sorry for disturbing the topic with a redundant question, I should've searched a bit more.