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

Asmodian
27th November 2017, 01:12
I agree, both double again and NNEDI3 are not needed going forward.

SamuriHL
27th November 2017, 01:34
Yes, NNEDI3 and double again can go. I'm in the unfortunate position of waiting for my replacement 1060 to get here as my brand new card I got in August died on me, so, I'm using my built in intel 4000 graphics for now. Suffice it to say my 1060 can't get here fast enough LOL. This is painful. But evenso, I have NGU working on it even for upscaling DVD's. That's not a bad piece of work! :D

jkauff
27th November 2017, 01:36
No problem here with any of your three proposal questions.

Neo-XP
27th November 2017, 01:56
1) Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"? I'd really like to remove the "double again" option, because it's dramatically bad for performance and IMHO also looks worse. But here's your chance to object, before I remove it.

I am not happy with the "direct quadruple" option. It looks not natural at all ("plastic" look), not as sharp as the "double again" option and adds a lot of aliasing.
To everyone wanting the "double again" option removed, please post some comparisons.

2) Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.

No, it can go.

3) Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?

IMHO, NNEDI3 is not relevant anymore with NGU AA, and if it can help you to get rid of OpenCl, remove it.

:thanks:

Gopa
27th November 2017, 01:58
upscaling anime = blurriness + compression artifacts
NGU Sharp + RCA = sharpens & removes compression artifacts
RCA (High) = heavy GPU load
"Free RCA" = no more heavy GPU load

RCA rendertime (720p to 2160p upscale - GTX1070) = 9-10ms at strength 10 (high). What will I do with all that extra rendertime?!
This will be just like upgrading my GPU!

NGU Sharp + "Free RCA" = Anime Dream Combo (ADC)

I wonder what NGU/madVR, will look like, in the future

yukinok25
27th November 2017, 02:17
Let madVR decide will never switch to NNEDI3 if you haven't selected it. In your bug report there were 2 crashes. One was not clear. The other one pointed to a crash in the Nvidia OpenCL driver. madVR uses OpenCL only for NNEDI3, so my conclusion was that you probably used NNEDI3. Please check the OSD (Ctrl+J) to confirm that NNEDI3 is really not used, neither in chroma upscaling nor image upscaling. If you're really not using NNEDI3, then I'm at a loss why the Nvidia OpenCL driver crashed!

I have never used NNEDI3 as is terribly slow with my GPU, then I guess I'll have to try to use NGU-Anti Alias and see what's happening.

Thank you for your help, and this amazing software by the way! ;)

ABDO
27th November 2017, 02:21
I wonder what NGU/madVR, will look like, in the future

the best as he is now :D:)

huhn
27th November 2017, 02:52
2) Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.
and someone here reported a bug with nnedi chroma and RCA which is clear there just saying.
3) Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?
well deinterlancing is currently a painful topic with an nvidia card and nnedi3 is a deinterlancer so maybe you could use it for that until we get proper deinterlancing without mismatch chroma positions and stuff like this...
you know give it something useful todo.

If I could remove NNEDI3 completely, that would be a relieve, because I could completely get rid of OpenCL, with all the driver issues etc.
isn't there a fast and working shader version of it?

JarrettH
27th November 2017, 02:53
Aside from getting it for free, if you're asking why NGU is expected to benefit from RCA, it's probably because upscaling tends to emphasize compression artifacts because the algorithm doesn't realize they're noise. Removing them before/during upscaling can make the scaler behave much better.

Well, I usually watch high quality sources. I didn't know everything could benefit from it. :thanks:

ABDO
27th November 2017, 04:21
ok, since NGU direct quadruple Implementation, i did some comparison with double again, and i found some Reasons make me prefer to stay using double again In spite of the diffrent performance between them, for me double again look somehow a little bit best than direct quadruple, a little details look deeper, some lines here or here treated best, and the overall picture look little best for my eyes.

with double again i can make double high and quadruple meduim and set chroma to high, but with direct quadruple i forced to use medium at all, i think this would be a Problem to weak gpu owners like me. :D

ryrynz
27th November 2017, 06:37
You guys that like double again, try quadruple with some/more? Adaptive Sharpen.

madshi
27th November 2017, 12:02
My opinion is that NNEDI3 can go the way of the dodo. :)
I stopped using NNEDI3, so i vote yes, go ahead and remove it.
You can clearly throw this opencl based sh*t in the trash.
I never use NNED13 for anything. Prefer: NGU AA.
I dont use nnedi3 so imho it can go
I never ever use NNEDI3 since NGU AA come to madvr, so for me it is time to go.
I never ever use NNEDI3.
2) 3) NGU AA made NNEDI3 obslolete. No regrets.
Double again and NNEDI3 can go.
Yes, NNEDI3 and double again can go.
IMHO, NNEDI3 is not relevant anymore with NGU AA, and if it can help you to get rid of OpenCl, remove it.
I have never used NNEDI3 as is terribly slow with my GPU, then I guess I'll have to try to use NGU-Anti Alias and see what's happening.
Thank you guys, that's a pretty clear vote that we can let NNEDI3 go. Rest in peace, you've served us well.

and someone here reported a bug with nnedi chroma and RCA which is clear there just saying.
Yes, I've had a couple of crash reports sent to me pointing to NNEDI3 being the cause of the crash. So clearly some users are still using NNEDI3. But are they using it because they compared it to NGU AA and found NNEDI3 to be better? I doubt that. They're probably using NNEDI3 either because they still have it active from before NGU AA existed, or because some tutorial told them to use NNEDI3, or because they are familiar with the NNEDI3 name, but don't know NGU AA.

In any case, I think the vote is very clear (see above), and it matches my own opinion. So NNEDI3 will be removed from the next build. I'll first remove just the options, but leave the code there, just in case definite proof comes up that it's still needed, after all.

well deinterlancing is currently a painful topic with an nvidia card and nnedi3 is a deinterlancer so maybe you could use it for that until we get proper deinterlancing without mismatch chroma positions and stuff like this...
you know give it something useful todo.

isn't there a fast and working shader version of it?
Yes, NNEDI3 was originally made for deinterlacing. However, if I wanted to, I could make a tweaked NGU variant for deinterlacing, too. Video mode deinterlacing is still on my list of things to look at. But priority for that keeps sinking because natively interlaced video content keeps becoming rarer every day. FWIW, it's not just a hour or two of making a deinterlacing solution available with NNEDI3, it would be a major effort, so not happening. I'm confident an NGU deinterlacing variant would beat NNEDI3 in terms of power vs quality. NNEDI3 was great for its time (it was actually very ahead of its time), but science/technology has caught up and overtaken it in the meanwhile.

I never use double again/twice.
No double/again upscaling also.
1) "Double again" is too sharp and slow, never used it since direct quadruple is appears.
Double again and NNEDI3 can go.
I agree, both double again and NNEDI3 are not needed going forward.
Yes, NNEDI3 and double again can go.
No problem here with any of your three proposal questions.
Ok, several votes for not needing "double again".

I am not happy with the "direct quadruple" option. It looks not natural at all ("plastic" look), not as sharp as the "double again" option and adds a lot of aliasing.
ok, since NGU direct quadruple Implementation, i did some comparison with double again, and i found some Reasons make me prefer to stay using double again In spite of the diffrent performance between them, for me double again look somehow a little bit best than direct quadruple, a little details look deeper, some lines here or here treated best, and the overall picture look little best for my eyes.
Ok, votes noted. Can you provide some screenshots (only the original small image suffices)? I've tested with my favorite test images and to my eyes direct quadruple usually looks better or at least not worse overall. In the "castle" image some lines are better and some worse with quadruple vs double again, so no clear winner there. But in the "clown" and "small" images, to my eyes direct quadruple clearly wins. See here:

http://www.madshi.net/clown.png
http://www.madshi.net/small.png

NGU Sharp + Free RCA!!!
Told you you'd love it! :)

Why is RCA so important to be built into NGU?
Because we're not photo editors who get RAW out of our cameras. We're getting highly compressed videos from the studios.

with double again i can make double high and quadruple meduim and set chroma to high, but with direct quadruple i forced to use medium at all, i think this would be a Problem to weak gpu owners like me. :D
Are you sure that double high + double medium is faster for you than direct quad high? That seems surprising to me.

foozoor
27th November 2017, 12:03
Hey madshi,
What config (mpv.conf) did you use to test FSRCNNX?
Did you use the D3D11, Vulkan or the old and slow Angle backend/api?

madshi
27th November 2017, 12:11
I used shinchiro's 2017-11-19 64bit build with "profile=gpu-hq". I didn't specify which backend to use. I trusted in mpv auto-picking the best option automatically. Bad idea? If D3D11 or Vulkan are better, why are they not the default backend/api?

foozoor
27th November 2017, 12:23
Vulkan doesn't work on old GPUs and the d3d11 renderer is not documented.
Igv released FSRCNNX_x2_r1_16-0-4-3.glsl that is quite faster too.

To get proper benchmark with vulkan, use this:
https://haasn.xyz/posts/2017-10-05-how-to-benchmark-mpvs-raw-throughput.html

Asmodian
27th November 2017, 12:50
Are you sure that double high + double medium is faster for you than direct quad high? That seems surprising to me.

On my Titan X (Pascal) at 1569 MHz scaling 528x288 to 2048x1152 NGU Sharp high 2x -> medium 2x is ~1.8 ms faster than high 4x. :(

However, a medium-high 4x option seems better if all we want are more steps to maximize the power available in a GPU. :p

madshi
27th November 2017, 13:25
On my Titan X (Pascal) at 1569 MHz scaling 528x288 to 2048x1152 NGU Sharp high 2x -> medium 2x is ~1.8 ms faster than high 4x. :(
With my Nvidia 1070, when upscaling 720p by 400%, I'm getting 13.3ms with 4x High, and about 14.0ms with 2x High + 2x Medium. In any case, it seems rather close, so is this really a reason to keep "double again"?

FWIW, once I switch to D3D11 rendering, NGU High + Very High will probably get a small speed boost. NGU Medium + Low probably not.

huhn
27th November 2017, 13:36
Yes, NNEDI3 was originally made for deinterlacing. However, if I wanted to, I could make a tweaked NGU variant for deinterlacing, too. Video mode deinterlacing is still on my list of things to look at. But priority for that keeps sinking because natively interlaced video content keeps becoming rarer every day. FWIW, it's not just a hour or two of making a deinterlacing solution available with NNEDI3, it would be a major effort, so not happening. I'm confident an NGU deinterlacing variant would beat NNEDI3 in terms of power vs quality. NNEDI3 was great for its time (it was actually very ahead of its time), but science/technology has caught up and overtaken it in the meanwhile.

if you can use NGU for deinterlancing sometime in the future well rip nnedi.

Sunset1982
27th November 2017, 13:51
Thank you guys, that's a pretty clear vote that we can let NNEDI3 go. Rest in peace, you've served us well.


But are they using it because they compared it to NGU AA and found NNEDI3 to be better? I doubt that. They're probably using NNEDI3 either because they still have it active from before NGU AA existed, or because some tutorial told them to use NNEDI3, or because they are familiar with the NNEDI3 name, but don't know NGU AA.

I would also vote for deleting NNEDI3.

And I doubt it too. I think there are a lot of outdated How to's out there where NNEDI3 was the best solution, before the era of NGU started. I think it will be fine to let NNEDI3 rest in peace.

madshi
27th November 2017, 14:08
Vulkan doesn't work on old GPUs and the d3d11 renderer is not documented.
Igv released FSRCNNX_x2_r1_16-0-4-3.glsl that is quite faster too.

To get proper benchmark with vulkan, use this:
https://haasn.xyz/posts/2017-10-05-how-to-benchmark-mpvs-raw-throughput.html
Ok, I've done some more tests. It seems for my Nvidia 1070, there's not much of a speed difference between Angle and d3d11, for some reason. Didn't test Vulkan. Doesn't make much sense if I can't get comparable measurement results.

The new FSRCNNX_x2_r1_16-0-4-3.glsl is faster than FSRCNNX 32-0-2-2, but slower than FSRCNN 32-0-2. But at least it runs (although not smoothly). FSRCNNX 32-0-2-2 just freezes on my PC, for some reason, when trying to play video.

So speed on my PC from fastest to slowest is:
FSRCNN 32-0-2 > FSRCNNX 16-0-4-3 > FSRCNNX 32-0-2-2

Comparing quality, FSRCNN vs FSRCNNX has no clear winner, IMHO. Either one beats the other in some image areas. Overall I might prefer FSRCNN by a tiny amount. However, this is just my personal opinion, and only based on the one "Spirited Away" test image. So take it with a pinch of salt.

madshi
27th November 2017, 14:09
if you can use NGU for deinterlancing sometime in the future well rip nnedi.
I would also vote for deleting NNEDI3.

And I doubt it too. I think there are a lot of outdated How to's out there where NNEDI3 was the best solution, before the era of NGU started. I think it will be fine to let NNEDI3 rest in peace.
Ok, thanks!

leeperry
27th November 2017, 17:07
Double again and NNEDI3 can go
+1, I'm not willing to waste CPU cycles on "Double again" even if it looked a bit better.

but non-RCA NGU Sharp will still be possible and medium/high will still be an option please?

Neo-XP
27th November 2017, 18:53
Ok, votes noted. Can you provide some screenshots (only the original small image suffices)? I've tested with my favorite test images and to my eyes direct quadruple usually looks better or at least not worse overall. In the "castle" image some lines are better and some worse with quadruple vs double again, so no clear winner there. But in the "clown" and "small" images, to my eyes direct quadruple clearly wins. See here:

http://www.madshi.net/clown.png
http://www.madshi.net/small.png

Those tiny images do not represent the real content I use madVR for.

With a 1280x694 source upscaled to UHD ( Source / NGU very High + Direct quadruple very High / NGU very High + Double again very High ) :

https://thumb.ibb.co/dbCrxm/Source.png (https://image.ibb.co/eXwJcm/Source.png) https://thumb.ibb.co/k9eX3R/NGUvery_High_Direct_Quadruple_Very_High.png (https://image.ibb.co/iqmC3R/NGUvery_High_Direct_Quadruple_Very_High.png) https://thumb.ibb.co/gg4ziR/NGUvery_High_Double_Againvery_High.png (https://image.ibb.co/kfGqq6/NGUvery_High_Double_Againvery_High.png)

Do you want to know what I find ugly ? Look at the external edges of his right ear :

http://screenshotcomparison.com/comparison/124784

It looks like plastic, very fake, and less sharp.
Even NGU + Jinc looks better than Direct quadruple IMHO.
NGU + Double again looks a little better than NGU + Jinc.

I do not use it now (not enough power), but I will with my next GPU because it is the best option.

Asmodian
27th November 2017, 19:05
I have to say I don't see it in your screen shot, double again looks a little sharper but no plastic look. I find direct quadruple results in a more natural look with my content.

Neo-XP
27th November 2017, 19:09
The edges are blurred, no texture, ugly.

It is as if the worst aspects of NGU were increased, but not the good ones.

Last comparison to try to save the Double again algo ( Source / Direct quadruple / Double again ) :

https://thumb.ibb.co/ib8s3R/Source.png (https://image.ibb.co/kRt3A6/Source.png) https://thumb.ibb.co/e2YZiR/Direct_Quadruple.png (https://image.ibb.co/gra3cm/Direct_Quadruple.png) https://thumb.ibb.co/mtyZiR/Double_Again.png (https://image.ibb.co/iBwLOR/Double_Again.png)

Look at this :

http://screenshotcomparison.com/comparison/124788

I let you judge.

aufkrawall
27th November 2017, 19:23
I used shinchiro's 2017-11-19 64bit build with "profile=gpu-hq". I didn't specify which backend to use. I trusted in mpv auto-picking the best option automatically. Bad idea? If D3D11 or Vulkan are better, why are they not the default backend/api?
It likely was already with D3D11 since it's higher in the autoprobe order list. OGL -> ANGLE -> DX11 presentation is not the default anymore with recent git builds.
profile=gpu-hq btw. also activates deband (and dithering).
Vulkan doesn't show lower GPU usage than D3D11 on Nvidia.

JarrettH
27th November 2017, 19:52
Because we're not photo editors who get RAW out of our cameras. We're getting highly compressed videos from the studios.

How apt, because I am a photographer :D

cork_OS
27th November 2017, 20:02
Look at this : http://screenshotcomparison.com/comparison/124788
I let you judge.
Shirt on the shoulder is definitely better with double again. "Double edges" is the ugliest NGU effect.

Braum
27th November 2017, 20:02
The edges are blurred, no texture, ugly.

It is as if the worst aspects of NGU were increased, but not the good ones.

Last comparison to try to save the Double again algo ( Source / Direct quadruple / Double again ) :

https://thumb.ibb.co/ib8s3R/Source.png (https://image.ibb.co/kRt3A6/Source.png) https://thumb.ibb.co/e2YZiR/Direct_Quadruple.png (https://image.ibb.co/gra3cm/Direct_Quadruple.png) https://thumb.ibb.co/mtyZiR/Double_Again.png (https://image.ibb.co/iBwLOR/Double_Again.png)

Look at this :

http://screenshotcomparison.com/comparison/124788

I let you judge.

Double again looks better indeed, there's more details and the edges are sharper, less blurred.

Asmodian
27th November 2017, 22:39
Shirt on the shoulder is definitely better with double again. "Double edges" is the ugliest NGU effect.

I agree, on both counts. In this scene double again helps reduce the "double edges" artifact which is very beneficial.

Some of my content has this issue with NGU Sharp and some doesn't, for anything that shows double edges with NGU Sharp I use NGU AA. :o

Siso
27th November 2017, 23:51
I have an issue with HPET enabled in OS and HPET disabled in OS...madvr shows 2 different refresh rates, with HPET in OS, (unfortunally can't disable in bios)...madvr shows 72.003 HZ refresh rate, when it is disabled in OS, it shows 72.004 HZ. I can't tell which one is the accurate one...
Also when it is stopped in OS, reclock adjusts itself after every PC boot, when it is ON in OS, reclock is adjusting only once after reboot, and then it is always green, no time how many restarts I make.

steakhutzeee
28th November 2017, 01:08
I know this is nooby but i've two questions, i use mpc be, lav and madvr.

Now, i think this is some sort of zoom settings in the madvr settings. I'm watching a move right now, i'm on 16:9 monitor and movie is 1:85,1. So i've a little black bars on top and bottom, just a few pixel.

1- In some scenes there are the records of an old camera, just a vertical image showed and black bars left and right. During these scenes the video is fitting the screen, like zooming, to perfectly 16:9. Just to turn back when the scene ends. Why? What options take me to achieve this?

2- In zoom control in madvr can you explain to me what the option "disable scaling if image size changes by only:". I'm newbie :/

Thanks! :D
Anyone can calrify this to me? Thanks!

ryrynz
28th November 2017, 01:28
I think most voted to remove double again simply because we don't use it and probably taking your word for double again being inferior on all accounts. This may not be so cut and dry especially with HD images rather than small photos.

Might be better for it to stay.
Have u guys tested with sharpening on quadrupling yet for comparison?

SamuriHL
28th November 2017, 02:36
I think most voted to remove double again simply because we don't use it and probably taking your word for double again being inferior on all accounts. This may not be so cut and dry especially with HD images rather than small photos.

Might be better for it to stay.
Have u guys tested with sharpening on quadrupling yet for comparison?

Having read the posts for the past day or so on this subject, I have to agree that this doesn't seem as cut and dry as some of us may have thought. I would opt to leave it until this can be sorted out, as well.

Asmodian
28th November 2017, 04:22
Might be better for it to stay.
Have u guys tested with sharpening on quadrupling yet for comparison?

I have tested sharpening, it works for the difference in sharpness but not the double edge look. However, even double again has double edge artifacts with content that does that, it is less obvious but I would still want to use NGU AA instead.

I think the double edge is in the source, a UHD image that when downscaled 50% looked like the source would have weird double edges. The edge has a faint line in the center, probably an artifact from sharpening (ringing from both sides of the edge), which is emphasised by NGU Sharp. I don't think it is an artifact from NGU itself because too many sources do not show it at all.

ABDO
28th November 2017, 04:24
Can you provide some screenshots (only the original small image suffices)?
sorry for late replay, a have a hard day.
Ok, here some screenshots
http://screenshotcomparison.com/comparison/124805
http://screenshotcomparison.com/comparison/124814
http://screenshotcomparison.com/comparison/124815
http://screenshotcomparison.com/comparison/124816

screenshots sours
http://www.mediafire.com/file/dws1bg7lmu7g349
/screenshots++sours.rar

it is little bit diffrent as i said.

Are you sure that double high + double medium is faster for you than direct quad high? That seems surprising to me.
ops, this is my big mistake (direct quad high) faster than (double high + double medium) for 1 or .5 ms, seem like ffdshow make unfunny stuff some times.

then i think that we can remove "double again" especially the diffrent visual between them hardly noticed.

edit
well, i had serious tests today, madshi i think double again is Power point in madvr, and we ( and madvr) will lose it if it removed,but definitely you are the boss.

ryrynz
28th November 2017, 06:42
http://madVR.com/doom9/stuff/asRnnRca.png


Someone went a bit crazy with the sharpening on that.
Hey madshi, super serious request, shoehorn awarpsharp into MadVR ;)


https://thumb.ibb.co/dbCrxm/Source.png (https://image.ibb.co/eXwJcm/Source.png) https://thumb.ibb.co/k9eX3R/NGUvery_High_Direct_Quadruple_Very_High.png (https://image.ibb.co/iqmC3R/NGUvery_High_Direct_Quadruple_Very_High.png) https://thumb.ibb.co/gg4ziR/NGUvery_High_Double_Againvery_High.png (https://image.ibb.co/kfGqq6/NGUvery_High_Double_Againvery_High.png)


Double again looks quite good to me here, really brings out the finer details and looks more alive so to speak, most noticeably the jacket and facial features.

Werewolfy
28th November 2017, 20:12
To wet your appetite, here's what I'm working on: I'm trying to fuse "NGU Sharp" + "Reduce Compression Artifacts" into one. Well, not really, but partially. Uhm. Ok. So if you don't use NGU Sharp, everything will stay as it is now. But if you activate both NGU Sharp + Reduce Compression Artifacts at the same time, basically RCA will come for free. So you get a nice speed reward for combining NGU Sharp + RCA. However, you can still use both separately, of course.

First results seem promising. I'll need your tests to check if there's any quality loss when using the fused algorithm instead of doing both separately. I hope there'll be no quality loss, but I'll let you test it, once it's ready to go. Maybe ready next weekend, maybe the weekend after that, but probably no later then 14 days from now.

For a start, it will only be available for NGU Sharp, maybe I'll also offer it for NGU Standard/Soft/AA, don't know yet. It's all a lot of work, so...

It's a great idea but why did you choose NGU Sharp? It's not a very good algo for bad quality videos, NGU AA is a better choice in this case and I think that RCA is mainly aimed to this type of videos. Do you think that RCA combined with NGU Sharp woul provide a good alternative for these videos? Not criticizing at all, just asking.


1) Do we need "double again/twice" for NGU Sharp/Standard/Soft? Or is everybody happy with "direct quadruple"? I'd really like to remove the "double again" option, because it's dramatically bad for performance and IMHO also looks worse. But here's your chance to object, before I remove it.

Very happy with "direct quadruple"! Like you said, "double again/twice" is really bad for performance. By the way, I hope to see "direct quadruple" for NGU AA if it's possible one day ;)


2) Anybody still using NNEDI3 for chroma upscaling? If so, is there a good reason why you prefer it over NGU AA? I'd really like to remove NNEDI3 chroma upscaling, because it's simply worse than NGU AA, IMHO, and slower, so removing the option should be all good, and make the settings dialog simpler.

No, remove it. NGU AA is a better choice.


3) Anybody still using NNEDI3 for image upscaling? If so, is there a good reason why you prefer it over NGU AA? I've recently seen an image comparison where NGU AA Medium was significantly better than NNEDI3-128 (sharper, more detailed *and* more natural), so I really wonder if there's any need for NNEDI3, anymore?

No, remove it too ;)

Blackwalker
29th November 2017, 18:06
My opinion is that NNEDI3 can go the way of the dodo. :)

same here :)

Skwelcha
29th November 2017, 18:42
Wanted to report a minor bug in madhccontrol v0.92.9, when I set my A/V receiver to "receiver, processor, switch" and the display connected as well to receiver, I get "An error occured in madvr" with the attached log when I try to open the display modes port on a different device. I know it's a weird thing to set in the first place but there's no display connected to my receiver that's why I set it that way... Obviously I just set the connected device back to digital display, just wanted to report my "finding".

I'm also voting for removing nndedi and doubling again/twice.

kalston
29th November 2017, 19:20
I haven't used NNEDI3 in a long time either. Certainly wouldn't miss it now that we have several very good (and much faster!) NGU options. Not sure about doubling twice, been using 4x all this time and didn't think there would really be any difference besides performance. The posted screenshots are interesting though, I guess I should give it a closer look but it seems very minor and I can't decide which I like better with those specific examples.

Been really happy with NGU sharp lately (used standard or AA for a long while). It's suitable for pretty much all I watch. I've also found that supersampling (with NGU sharp of course) actually does a lot of good on the 1080p 27" monitor I have now (for 720p sources and DVDs). On my (now defective) 1440p/27" display supersampling 1080p files (or DVDs) was a complete waste of resources on the other hand. Could just be because the pixels are bigger so I can see the difference better (my sitting distance has not changed).

chros
29th November 2017, 20:04
Yes, there is, using F5 for different HDR->SDR conversion:
- pixel shader conversion is OK
- 3dlut conversion is Not (image is much darker)

PrintScreen in Windowed mode works fine for both of them.

Can you give me a few more details. How is 3dlut conversion not ok? Is the 3dlut not applied for the screenshot? Or is it applied incorrectly? What happens if you disable the 3dlut? Are F5 and PrintScreen identical then for SDR movies?
Here are the screenshot made on a 120 nits, 2.2 gamma, SDR 3dlut calibrated monitor:
- 3dlut : HDR->SDR 3dlut conversion
- pixel: pixel shader conversion
- madvr-decide: using the "let madVR decide" option

Using F5 in FSE mode (with MPC-BE):
hdr-sdr-3dlut-fse-f5 | hdr-sdr-pixel-fse-f5 | hdr-sdr-madvr-decide-fse-f5
https://thumbs2.imgbox.com/1f/c3/tMNnliJJ_t.png (http://imgbox.com/tMNnliJJ) https://thumbs2.imgbox.com/22/74/MTWFs3FR_t.png (http://imgbox.com/MTWFs3FR) https://thumbs2.imgbox.com/f6/ac/BBuDIU75_t.png (http://imgbox.com/BBuDIU75)

Using PrintScreen in Windowed mode:
hdr-sdr-3dlut-window-PS | hdr-sdr-pixel-window-PS | hdr-sdr-madvr-decide-window-PS
https://thumbs2.imgbox.com/d4/69/Mjp1UTjt_t.png (http://imgbox.com/Mjp1UTjt) https://thumbs2.imgbox.com/18/3e/KxcheCDX_t.png (http://imgbox.com/KxcheCDX) https://thumbs2.imgbox.com/52/81/BEk9Kl3I_t.png (http://imgbox.com/BEk9Kl3I)

Nyago123
29th November 2017, 21:19
Interesting. But why not stick to 385.xx until the issue if fixed?


Yeah, I found it interesting that on my system HDR switching behavior varied between players, though as I'm sure most everyone knows that component installation configuration (splitter, renderer, etc) can vary as well even within one system.

It's not really all that much of an inconvenience when I factor that 99% of content is still SDR, and the Win 10 system I'm using is doubling as a game machine so I prefer having the latest drivers. Taking a few seconds to play with the Windows HDR toggle switch when things go wrong is annoying but it's a First World Problem... lol.

If I were that concerned, I'd probably take your advice and go with Win 8.1 in a dual boot, using the 8.1 when I want to do HTPC and Win 10 otherwise but I'm just too lazy. :p

Randi
30th November 2017, 00:07
after googling, looking through madshis bugtracker and searching here didn't find anything: anybody has an idea why when I try to test a custom mode (nothing fancy, just the currently active one anyway) I always get a BSOD with "unexpected kernel mode trap"?
trying to get rid of reclock :-)
nvidia 1030 directly connected to a 4k hdr tv, driver version 385.29 (as recommended)

Gopa
30th November 2017, 03:23
It's a great idea but why did you choose NGU Sharp? It's not a very good algo for bad quality videos, NGU AA is a better choice in this case and I think that RCA is mainly aimed to this type of videos. Do you think that RCA combined with NGU Sharp woul provide a good alternative for these videos? Not criticizing at all, just asking.
)

For most of my anime, NGU Very Sharp (image upscale) + RCA High, provide, images, that look sharper & also, free of artifacts, vs NGU AA (need more sharpening refinements). Mostly, average to good quality anime, though. High-quality images: NGU Very sharp + RCA High. I will test with lower quality anime (NGU AA + RCA High), to see if this is preferable (to me). Both, NGU Sharp & NGU AA, combined with RCA, would probably be nice, I guess, especially, for "other than anime" videos (?). Looking forward, to combined, NGU Very Sharp + RCA High, first, for most of my anime, at least. I do, usually, use NGU AA, for chroma upscale. Am I missing something?

EmuAGR
30th November 2017, 10:29
I've been trying to enable HDR but it didn't work, my TV doesn't change to HDR mode and colors look really poor. My current setup is Windows 7 x64 + AMD RX 580 + MPC-HC + LAV 0.70.2 + madVR 0.92.9 .
I've checked passthrough and metadata (for the TV), exclusive mode and D3D11 rendering options.

I've tried everything I saw online with no luck.

nevcairiel
30th November 2017, 10:38
I've been trying to enable HDR but it didn't work, my TV doesn't change to HDR mode and colors look really poor. My current setup is Windows 7 x64 + AMD RX 580 + MPC-HC+ LAV 0.70.2 + madVR 0.92.9 .
I've checked passthrough and metadata (for the TV), exclusive mode and D3D11 rendering options.

I've tried everything I saw online with no luck.

With an AMD card you'll probably need Windows 10 for HDR, because madVR only supports HDR passthrough using the NVIDIA API (on any OS) or the Windows 10 API. Unless I missed that this changed along the way..

Edit: Apparently it did change, read the bottom of the next post.

madshi
30th November 2017, 10:59
I know this is nooby but i've two questions, i use mpc be, lav and madvr.

Now, i think this is some sort of zoom settings in the madvr settings. I'm watching a move right now, i'm on 16:9 monitor and movie is 1:85,1. So i've a little black bars on top and bottom, just a few pixel.

1- In some scenes there are the records of an old camera, just a vertical image showed and black bars left and right. During these scenes the video is fitting the screen, like zooming, to perfectly 16:9. Just to turn back when the scene ends. Why? What options take me to achieve this?
I'm confused about "1-". How can the video fit the screen to perfectly 16:9, if you have black bars left and right?

2- In zoom control in madvr can you explain to me what the option "disable scaling if image size changes by only:".
If you have a movie with a resolution like 1919x1079 and play it on a 1920x1080 TV, then most media players by default ask madVR to scale 1919x1079 to 1920x1080, which is not really good for image quality, and wastes GPU power. So with this option activated, madVR will "override" the media player wish to have this video scaled up by 1 pixel and instead draw it unscaled, with a 1 pixel sized black bar at the right side and the bottom of the video.

It's a great idea but why did you choose NGU Sharp? It's not a very good algo for bad quality videos, NGU AA is a better choice in this case and I think that RCA is mainly aimed to this type of videos. Do you think that RCA combined with NGU Sharp woul provide a good alternative for these videos?
One of the key reasons why NGU Sharp is not a good algo for bad quality videos is because of compression artifacts. So the combination of RCA + NGU Sharp seems a natural fit. There were also a number of users who reported that they switched from NGU AA to RCA + NGU Sharp.

But as I said, I may also try to create a fused version of RCA + NGU AA, but that's not decided yet.

Wanted to report a minor bug in madhccontrol v0.92.9, when I set my A/V receiver to "receiver, processor, switch" and the display connected as well to receiver, I get "An error occured in madvr" with the attached log when I try to open the display modes port on a different device. I know it's a weird thing to set in the first place but there's no display connected to my receiver that's why I set it that way... Obviously I just set the connected device back to digital display, just wanted to report my "finding".

I'm also voting for removing nndedi and doubling again/twice.
Attached stuff to this forum doesn't work, unfortunately, so please upload it somewhere else and then link to it here.

Here are the screenshot made on a 120 nits, 2.2 gamma, SDR 3dlut calibrated monitor:
- 3dlut : HDR->SDR 3dlut conversion
- pixel: pixel shader conversion
- madvr-decide: using the "let madVR decide" option

Using F5 in FSE mode (with MPC-BE):
hdr-sdr-3dlut-fse-f5 | hdr-sdr-pixel-fse-f5 | hdr-sdr-madvr-decide-fse-f5
Hmmmm... The 3dlut stuff is surprising. I'll check if I can reproduce it here. Which pixel shader conversion "display peak nits" value did you choose? I think with "let madVR decide" madVR uses 400nits. You've probably chosen a much lower value? That would explain the difference between "let madVR decide" and the pixel shader math option. FWIW, I'm planning to lower the default "display peak nits" value in a future version.

If I were that concerned, I'd probably take your advice and go with Win 8.1 in a dual boot, using the 8.1 when I want to do HTPC and Win 10 otherwise but I'm just too lazy. :p
I never advised to use dual boot. If you have a pure HTPC, there's no reason to use Win10 at all. If you need Win10 for some reason, e.g. for DX12 games, then by all means, use Win10 instead of Win8.1. You may suffer a bit, but that's how it is. You can still fix the HDR problems by using an older driver, as has been suggested about a million times already.

after googling, looking through madshis bugtracker and searching here didn't find anything: anybody has an idea why when I try to test a custom mode (nothing fancy, just the currently active one anyway) I always get a BSOD with "unexpected kernel mode trap"?
trying to get rid of reclock :-)
nvidia 1030 directly connected to a 4k hdr tv, driver version 385.29 (as recommended)
A BSOD is almost always the fault of a driver. In this case most probably the GPU driver. I'd suggest that you try installing a different driver version. I'm sorry, but there's not much else I can suggest here. As you wrote yourself: Nobody else has reported this issue, so it seems specific to your PC somehow. Most probably some sort of installation/configuration issue with your OS. If all else fails, an OS reinstall could possibly fix it (but is a lot of work). With a bit of luck just reinstalling the GPU drivers, or installing a different GPU driver version may fix it. If you have extra bad luck, it could be a GPU BIOS or hardware issue.

I've been trying to enable HDR but it didn't work, my TV doesn't change to HDR mode and colors look really poor. My current setup is Windows 7 x64 + AMD RX 580 + MPC-HC+ LAV 0.70.2 + madVR 0.92.9 .
I've checked passthrough and metadata (for the TV), exclusive mode and D3D11 rendering options.
Please upload a screenshot of the OSD (Ctrl+J), when playing HDR content, to some image sharing host and link to the image here (don't attach here, doesn't work).

I'm not completely sure if the AMD HDR API supports Windows 7, but I think it should. Do you have the Windows 7 Platform Update installed?

With an AMD card you'll probably need Windows 10 for HDR, because madVR only supports HDR passthrough using the NVIDIA API (on any OS) or the Windows 10 API. Unless I missed that this changed along the way..
Actually, I've added support for a private AMD HDR API a while ago, so it should work with AMD in Windows 8, maybe also in Windows 7. However, the AMD API has some extra annoying requirements: It only works in fullscreen mode (doesn't have to be exclusive, though), and D3D11 presentation must be used with 10bit backbuffer.

mclingo
30th November 2017, 12:41
anyone else getting a black screen on latest MADVR, dropping back to previous version fixes, I get this playing all movies.

note, i've just updated to windows 10 fall creators update, i'd removed it due to instability with BSOD and gfx drivers, thought i'd trying it again with latest releases.

Latest LAV seems ok but i'm having to drop back to a previous MADRV, sorry not with my PC at the moment so cant tell you which version is working for me.

jkauff
30th November 2017, 12:56
anyone else getting a black screen on latest MADVR, dropping back to previous version fixes, I get this playing all movies.

note, i've just updated to windows 10 fall creators update, i'd removed it due to instability with BSOD and gfx drivers, thought i'd trying it again with latest releases.

Latest LAV seems ok but i'm having to drop back to a previous MADRV, sorry not with my PC at the moment so cant tell you which version is working for me.
This has been reported before. The Fall Creators Update and Full Screen Exclusive mode do not work well together. In fact, they don't work at all.

Before madshi jumps in and tells you to go back to Win 8.1 :), let me suggest you use full screen windowed mode instead. You probably don't absolutely need FSE.