ENB Comparison with Benchmarks

[Gunthral]

 

Hello together,

 

I've finally decided to make an information thread here about my ongoing and extending comparison for ENBs.

 

What's included so far in short:

 

51 ENBs with

77 variants/presets

1668 screenshots taken

120 benchmarks

 

Site on the Nexus: http://www.nexusmods.com/skyrim/mods/60685/?

 

Because I've wrote a really long description that should answer a lot of questions you're probably have, here's the whole text from the nexus in the spoiler:

 

 

 

Update: 30-shots-series updated for 21 presets

So far...
51 ENBs with
77 presets/variants
1668 screenshots taken
120 benchmarks

English description
[/size]

• Introduction
• Components of this comparison
• List of compared ENBs
• Data of the selection-spreadsheet
• At what this comparison is(n't) aiming
• How the scenes are selected and why
• How benchmarks are done here
• Interpret the benchmarks the right way
• What a ENB should have to be covered here
• Credits and thanks

1. Introduction

Hello together,

I’m a German user and working on my new Skyrim-installation for quite some time. Concomitant to that I’m also writing a German guide to introduce gamers how to build a proper working mod installation – not to give a strictly „mod instruction“, but rather to supply them with the necessary tools. One point of this is the search procedure for an appropriate ENB preset. And because I’m assuming that at least this part can also be helpfully for other users I want to share it here.

To putting my benchmarks in relation to different hardware specs you’re needing of course my specs – so here they are:

CPU: i7-5820 @Stock - 6x 3,3GHz
GPU: sapphire vapor-x radeon r9 290 tri-x oc 8GB VRAM (@Stock)
RAM: 32GB Gskill DDR4-2400 CL15
HDD: OS on Crucial-512GB-SSD, Skyrim on 1TB-Samsung 840 basic
Display: EIZO FS2434 (FHD, 1920x1080, 60Hz)

One word on this: Yeah, I'd surely better suited with an "normal" 1150-system and an 4790k - but I'm also using this system for some CAD work, rendering and so on... in fact for most of the time, when currently not modding
Anyway, I don't even need to OC my CPU when you're looking into the benchmark results.


2. Components of this comparison series/quick reference

So, what can you expect when downloading the archive, whats’ integrated so far? For a much more detailed description of each component (except the pdf file) refer to the coressponding points further down.

• Of course the screenshots (muaha).
  Each ENB has its own folder and sometimes also subfolders which then represent different presets (e.g. Aeon ENB or The Grim and Somber ENBs). Every screenshot series consists of 18 or 30 screenshots. The single settings for the sceneries 01 to 18 were settled mostly by following the excellent guide at skyrim-comparison (couldn’t find the name of the author). The sceneries 19 to 30 were defined and tested by myself. If you’re want to do a “real” comparison for an single situation, I suggest to search the archive for the corresponding number (all screenshots having a double-digit number, beginning with 01 counting to 18). Additional informations about this can be found in part 5.
• A chart with fundamental informations about the benched ENBs
  This sheet hosts informations about the most generally characteristics of each examined preset. This is the reference if someone would like to know what version of the preset is covered, who's the author and what are the (in)compatibility states with the most common ENB-related mods. Maybe additional information will be added over time. For a more detailed description of this part refer to chapter 6 and 7.
• The benchmark results
  This should be self-explaining. Each examined ENB has one (or more) result row(s). It contains information about the FPS, GPU (and VRAM), CPU and the allocated RAM. Bear in mind that different variants of a ENB aren't benched individual every time - only when they're differing in terms of performance. Different performance versions are normally benched seperate (but not for every ENB with every possible configuration). Read chapter 7 for more informations.
• A .docx document showing the used configuration
  This describes in a little more detailed way the additional used mods and optional files at the benchmark and screenshot series of the corresponding ENB. This includes weather and additional lighting mods, options of the ENBs itself (performance, coloring, etc.) and other important settings. Thereby it should be easier to reconstruct the screenshot settings.

3. List of compared ENBs so far

Note: Don't confuse "preset" with "performance options" . "Preset" means you'll get at least a different look and feel (and maybe performance). So screenshots are taken for these (obviously), benchmarks then only if necessary (e.g. the performance options from the Aeon Fantasy and Natural presets don't differing in my benchmarks). "Benchmarks" finally declares mostly performance options. So for different performance options you'll see different benchmark results (that doesn't mean for ALL options...).

• Aeon ENB – by LucidAPs - (3 presets, 3 benchmarks for perf. options)
• Air ENB - by hrk1025
• Atomic ENB - by ixm
• BLEAK ENB - UNBLEAK ENB - by akiro (2 presets, for each 2 benchmarks)
• Bleak Realism ENB - by ZenonSeth (2 benchmarks)
• CR ENB - by Bigeyescarl (3 benchmarks)
• Clarity ENB - by bretton
• Elder Blood ENB - by CommanderBrightside (3 benchmarks)
• Falkonian ENB - by XxDCExX
• Far Off ENB - Claymay - Performance – by far327
• Far Off ENB - Nirn - CGI (COT Pure Weather ELFX ready) – by far327
• Far Off ENB - Nirn ( For COT and ELFX ) – by far327
• Fire and Storm ENB - CoT compatible – by FireStorm974 - (2 presets, each benched)
• Gameplay Performance ENB - by Scrabbulor
• Horizon ENB - by ARYAYADA (2 benchmarks)
• HRK ENB TWO - by hrk1025
• Insomnia ENB – by Insomnia - (2 presets, 4 benchmarks)
• K ENB Pure Light – by Kyokushinoyama - (3 benchmarks)
• Kountervibe ENB Northern Light – by Kyokushinoyama - (3 benchmarks)
• Lumen ENB – by Jafin16 - (3 benchmarks)
• M ENB - by Sloth
• Megatherium ENB- by Sloth
• Minimal ENB Final Edition - by ewi65 (2 presets, each benched)
• Noir ENB -- Fogbound - by Rusey
• Organic ENB optimized for ELFX and COT compatible – by skysan4298
• Paradise ENB - by ProtologolusX
• Phinix Natural ENB – by phinix
• Poupouri ENBs - Extended – by MACHIIMACHII - (2 presets, each benched)
• Project ENB - by Bronze316 (2 benchmarks)
• PureVision ENB – by SkyrimTuner
• RealVision ENB – by SkyrimTuner - (2 benchmarks)
• Rudy ENB - by Rudy102
• Ruvaak Dahmaan ENB - by Dragens
• Saraan Suum ENB – by MACHIIMACHII - (2 benchmarks)
• Seasons of Skyrim ENB True HDR – by bronze316 - (2 presets, for each 2 benchmarks)
• Shift ENB - Performance – by bretton
• Skylight ENB - by Aiyen
• Somber Dreams ENBs – by Maeldun0 - 2 presets, for each 2 benchmarks)
• Straylight ENB - by fadingsignal (2 benchmarks)
• T.A.Z Visual Overhaul - by TAZ
• The Grim and Somber ENBS – by wolfgrimdark (6 presets, totally 11 benchmarks)
• The Mistveil ENB – by Modd3r - (4 presets, each benched)
• TK ENB - by tamu75
• Toric law of nature ENB - by magica2 (3 presets, for each 2 benchmarks)
• True Vision ENB – by bronze316 - (2 benchmarks)
• Tundra ENB – by wiplashwang - (3 Presets, 2 benchmarks for perf. options)
• unkoENB - by unko9622 (2 benchmarks)
• Vividian ENB - Weather and Lighting – by Mangaclub (2 Presets, for each 3 benchmarks)

4. Data of the selection-spreadsheet
In this chapter I want to (shortly) describe the columns in the selection-spreadsheet and give you a hint what informations can be get out of it.

• Name and author: should be self-explainable...
• Version and date: Stating the version and the last updated date of the ENB that's included in the comparisons. For ENBs with multiple presets I only stating the latest version.
• ENB-dll: The version of the ENB binary that was used in the comparisons and benchmarks. In the uttermost cases I'm sticking to the description page of the ENB for this, only changing it when a newer version is reported to be compatible by the author.
• Style: A fully subjectively categorizing for each of the ENBs in a few coarse groups of their visual appearance, going from quite "normal" ones (real) to ENBs that are so far away from reality that's really obvious (surreal).
• Various compatibility columns: Here you can find some information about the compatibility with the most popular weather and lighting mods. In particular these are: CoT (Climates of Tamriel), NLA (Natural Lighting and Atmospherics for ENB), PW (Pure Weather), RCRN (RCRN AE -- HDR Lighting and Weather Enhancement), RLO (Realistic Lighting Overhaul), RLS (Relighting Skyrim), ELFX (Enhanced Lights and FX), ELE (Enhanced Lighting for ENB).
  Comment: Again, I'm here sticking to the description of the authors. It may be the case for some presets that even a mod that has stated a "no" for compatibility may look quite good ingame but it's not recommended or supported by the author - that's the difference.
• Night-Eye-Compatibility: An additional column I've inserted per user request. It's important to say that I don't installing such mods especially because I'm not focusing on using it. Therefore I'm searching the description and the comments of each ENB to probably find an indication about compatibility with these. The status is separated into 4 levels. "No" means either it's simply incompatible or I couldn't find any information about that. "Yes" should be self-explainable, the author itself states here that Night-Vision should be compatible. Yes or No with additional brackets then means there are more adjustments needed (Yes) or you've to change quite an bunch of parameters to get it working (No).
• VSync: Just a information on the status of the VSync-parameter in the (most times) included enblocal.ini. And just to declare that here, I'm forcily disabling VSync in any case, but just for the benchmark purpose!
• Included since/Last update: Should be pretty self-explainable. Just an information since which version the ENB was included in the comparisons and in which it was updated the last time.

5. At what I'm aiming with this comparison - and at what not

So I've given a little insight about what I want to do with this guide in the introduction... now get a bit more detailed.

The top priority for me in this comparison are the authenticity of the screenshot series. This includes that they should most exactly look as intended by the individual authors. But that could of course be somewhat difficult especially in terms of the comparison aspect. Installing ALL the recommended mods for each of the ENB presets would need a real astonishing additional amount of installation work. In fact it should multiply the needed time for setting up an specifc ENB installation (calculate some hours installation for an half-hour of taking the screens and benchmarking. Also this could lead to an much greater likeliness for failures or inconsistences in the screenshots and benchmarks. Another disadvantage of installing to much additional mods would be the greatly increased difficulties of comparing one preset with another. Where do you end when installing additional mods for ENBs? That counts especially for texture and similar image-enhancing mods.

Therefore I've set up a simple rule for installing additional mods when arrange an specific ENB installation: Install only the really required mods. This includes mods like:

• Climate-changing mods, e.g.: Climates of Tamriel, Pure Weather, Removed Exterior Fog
• Lighting-Mods, e.g.: Realistic Lighting Overhaul, Enhanced Lights and FX, Relighting Skyrim
• ENB-related patches, e.g.: Subsurface scattering patch, Particle patch
• additional needed dlls, e.g.: enbhelper.dll
• various other things, e.g.: mostly specific data, dedicated Data-Archives, special sunsprite, cloud or similar textures specially for this ENB

These points should give some hints for what shouldn't be expected to see her: pure high quality screenshots of the presets. That's mostly because of the absolute focus on the comparison target. But also because I'm thinking that the "real good" screens are hosted way better at the authors mod sides than here.
Another thing that should be considered: I won't give explicitly support for all the covered ENBs. Primarily because I would pretty surely give some misinformation or can't help you anyway. That doesn't mean you can't ask me for advice when experiencing problems. But don't expect much helpful input aside from the most generic themes.

Comment: The above rule counts especially for the sceneries 01 to 20. That's because these are enviromental screens and things such as bland textures aren't becoming too obvious. Anyway I'm making an exception from these for the scenes 21 to 28 and 30. Otherwise the character textures of vanilla would becoming to apparent here. So the additional mods installed here are:
- CBBE - Caliente's Beautiful Bodies Edition: Slim, Vanilla, Ultra Textures
- SG Female Face Textures Renewal
- XCE - Warpaint and Dirt
- CMO - Complete Male Overhaul (Quality version)
- High Res Face Maps for Men
- Frost Atronach HD Retex

So I'm coming to the target of the second big part of this comparison: the benchmarks. You'll see alot of it and with certainty a greater quantity of them then the screenshot series. They should give you an relative (considering the non-professional claim) accurately estimate of the performance impact or the needed system ressources - if read in the right way. I will that explain in chapter 7. Bear in mind that the benchmarks aren't focusing a real sophisticated target in terms of statistics. I'll maybe adding additional parameters or statistical data over time, but that's getting pretty complex at some point or for special investigations. For example a so-called ANOVA-study would really tease me here, but that's a real mountain of work if you want to doing it right. And also needs quite some financial investments...6. How the scenes are selected and why

In this chapter I want to give you more information about the selected comparison scenes, write down important informations for each scene and describe what you should expect in general when using an ENB in this scenario.

Bear in mind, that the scenes 19 to 30 weren't ther since the beginning of this comparison series. Therefore it's possible that your desired preset have only the scenes 01 to 18 in its screenshot series included. I'll be adding these shots later on respectively when the ENB gets an update.

Important: I've taken nearly anything of information about possible suitable enviromental screenshot scenarios from the technically outstanding guide for screenshot comparisons in Skyrim at wikidot.com. Unfortunately I couldn't contact the author of this guide (or the side owner) to express my thanks for the work that must have gone into it. The problem with the side is in fact the nearly total lack of actuality. I mean, the last covered ENB preset is for the dll-version 0120, so this shouldn't be a daring statement. Anyway I've made some (fairly little at the moment) changes to the scenes and surely will making more over time. So go onto it...

Comment on the "Comparibility"-parameter given for each scenery: I've included this annotation for each scene to give an indication how good the screens should be comparable from one preset to another. Take scene 17, where you sitting on a chair next to an NPC and look to the cork of the winebottle. Unfortunately Skyrim don't save the camera view within the savegame and therefore the camera must be realigned for each shot. That leads of course to some amount of "user failure". Nonetheless this shouldn't have an great influence and is almost unnoticeable... but I want to point it out. 100% stands for absolute perfect accordance of ALL image components (this would include things such as ambient fog... you can't find such a scene), 0% would be... yeah, you won't having that here.

Scene 01: City by clear day
Fast travel to Whiterun, than look at the top of the Bannered Mare roof.
Weather: SkyrimClearTU_A, ID 10e1ee
Time: 12:00 PM
Comparibility: 95%
A scenerie that tends to be quite bright. In the "original" guide there was the default SkyrimClear-weather noted (81a). I changed that because most of the presets then were too bright in my opinion. This is a good scene to examine if the preset has a DOF that blurs out the sceen part in front of the focus plane. This should be noticeable when the Bannered Mare is appearing clear, but the houses in front are blurried out.

Scene 02: City by clear night
Wait 12 hours after scene 01.
Weather: SkyrimClear, ID 81a
Time: 12:00 AM
Comparibility: 98%
In contrary to the first scene this here can give (obviously) a much darker experience. This differs in terms of brightness (at least subjectively) greater than the first one from one ENB to another.

Scene 03: Interior lighting (lively places)
Enter the Bannered Mere through the front door and stand still.
Time: 12:00 AM
Comparibility: 92%
One of the main scenes to compare interior lighting in more busy buildings. Quite big differences can be found here, from a relatively dark to a much more cozier feeling.

Scene 04: Interior lighting at day (dungeons)
Fast travel to Bleak Falls Barrow and enter the temple hall.
Time: 12:00 PM
Weather: SkyrimClear, ID 81a
Comparibility: 92%
The only screen in this comparison in which the weather outside has an influence on the screen. It should be considered that this scene is the only one so far that needs some time of "readjusting" the lighting after forcing the weather (at least in the mostly presets).

Scene 05: Exterior lighting at forests at day
Fast travel to Falkreath and l ook to the stone wall behind.
Time: 10:00 AM
Weather: SkyrimClear, ID 81a
Comparibility: 98%
There are several screenshots of areas where trees taking a big potion of the screen. But here there are for sure predominant. Therefore this should the primary shot to examine the tree shadowing.

Scene 06: Yellow and red leaf trees, grass
Fast travel to Snow-Shod Farm and look to the roof of the farmhouse.
Time: 12:00 PM
Weather: SkyrimClear, ID 81a
Comparibility: 98%
A quite generic scene that should represent the looking of areas such as The Rift really decent. Gives a somewhat red tinted vegetation with some of the presets. Also good for giving an feeling how much the ENB is blurring out the elements in the background.

Scene 07: Indoor with multiple lighting sources
Fast travel to Solitude and go into Castle Dour. Then stand still.
Time: 12:00 PM
Comparibility: 90%
This scene can produce a big contrast between bright lightened and dark shadowed areas. Good to examine the interior lighting intensity when additional ambient fog is present.

Scene 08: Sunlight reflections on water and lens flare
Fast travel to Ilinalta's Deep and look towards mountains.
Time: 5:45 PM (changed it a bit from the original guide)
Weather: SkyrimClear, ID 81a
Comparibility: 98%
The only scene (beside no. 12) where water is a really predominant element - besides the sky. A good scene to check the reflections of sunlight on the water. The coloring of the sky at sunset and - if present - also sunrays or lens flares can be observed. Also a good scene to see if a preset has a nearly indefinite focusing distance for the DOF settings.

Scene 09: Rainstorm and fog at day
Fast travel to White River Watch and look to the mountains.
Time: 12:00 PM
Weather: SkyrimClear, ID c8220
Comparibility: 98%
The only scene which shows a weather ID with a rainstorm. The strength of the fog differs greatly between the preset and so does the foresight.

Scene 10: City at noon
Fast travel to Falkreath.
Time: 12:00 PM
Weather: SkyrimClear, ID 81a
Comparibility: 95%
A good scene to examine the SSAO settings that should be visible at the gateway. Also another shot where a DOF that's blurring out the screen in front of the focus plane should be noticeable.

Scene 11: Dungeon lighting (1)
Looking at the torches near the bridge inside the Embershard Mine.
Time: 12:00 AM
Comparibility: 92%
One of two screens that should depict the lighting inside dungeons. This is of course also greatly depending on the additional installed mods, such as ELFX.

Scene 12: Sunset coloring
Fast travel to Ilinalta's Deep and look towards mountains.
Time: 6:50 AM (changed that a bit from the original guide)
Weather: SkyrimClear, ID 81a
Comparibility: 95%
Nearly the same scene as no. 09 but with time set nearly an hour later and therefore a darker look. Depending on ENB you may have a red sunset here instead of scene 09

Scene 13: Looking of ice on day
Fast travel to Septimus Signus's Outpost.
Time: 11:00 AM
Weather: SkyrimCloudy, ID 12f89
Comparibility: 92%
The only scene with a cloudy weather setting. And the only one in which ice is absolutely predominant. Therefore it should be the one when want to look at the coloring of ice.

Scene 14: Lights of windows and snowstorm weather
Fast travel to Windhelm, go outside the gate and reenter. Then stand still.
Time: 11:30 PM
Weather: SkyrimStormSnow, ID c8221
Comparibility: 90%
Another screen in the series that's the only one who shows a strong snowstorm weather. Besides the looking of the snowy weathers (obviously) her should also the lighting of the windows in the middle building be comparable.

Scene 15: Reflection and lighting in snowy areas
Fast travel to Bleak Falls Barrow and look towards the ruins.
Time: 10:00 AM
Weather: SkyrimClear, ID 81a
Comparibility: 95%
A good screen to test how bright a ENB will actually look in snow-covered areas (and the only one at the moment). In some presets this scene tends to be really bright.

Scene 16: Water reflections from lighting sources at night
Standing on the stairs of an underwater ladder near the East Empire Company Warehouse at the docks of Solitude.
Time: 12:00 AM
Weather: SkyrimClear, ID 81a
Comparibility: 98%
One shot where water reflections produced from some smaller light sources (e.g. torches, lanterns) can be viewed. Also the darkness of nights in general should be good comparable as the light sources are quite scarce.

Scene 17: DOF and feeling of barely litted rooms
Sitting next to Belrand on a chair in the Winking Skeever (Solitude) and look towards the cork of the winebottle.
Time: 12:00 PM
Comparibility: 85%
In fact the only shot that is focussing on the "normal" DOF settings. Therefore this is absolutely vital in the series. Due to the lacking of light sources in the viewed corner this scene can be quite dark with some presets.
Comment: It should be noticed that Skyrim seems to doesn't save the looking angle in the savegame files when sitting on a chair. Therefore the screenshot excerpt can differ between the series.

Scene 18: Dungeon lighting (2)
Enter the Solitude Catacombs at the entrance near the cemetry (that has changed from the original guide). Then stand still.
Time: 12:00 PM
Comparibility: 95%
Another screen that should depict the interior lighting of dungeons. It can be quite different from a relatively bright view down the corridor to a nearly pitch black appearence.

Scene 19: Dungeon lighting (3)
Wait 12 hours after scene 18.
Time: 12:00 PM
Comparibility: 95%
I'm pretty sure, most of the users will say, that this shot is simply the same as no 18. Yeah, in most of the cases that's right. Anyway ther're some ENBs out there that're rending the same interior at day in another way than they do at night. That's the only case you would see an difference between the two shots.

Scene 20: Water color, steam at day
Standing at some hot springs westwards of the Dragonshrine "Bonestrewn Crest". Then look to one of the half-sunken stones there.
Time: 10:00AM
Weather: SkyrimCloudy, ID 12f89
Comparibility: 90%
A quite nice scene in my opinion. Gives the water there a somewhat greenish appearence with most of the presets. Should also be a good scene to examine the look and brightness of the steam springs.

The following shots are mainly focused on characters or creatures. Therefore the additional texture mods are only used here. Every scene should be a good point to compare the DOF settings (although this is more an accompaniment and not the focus).

Also it should be noted that these scenes are far more difficult to shot then scenerie ones. That's mostly because the added "randomness" portion due to the present NPCs. Normally it's possible to circumvent that by using the "tai"-command in the console (toggle AI). Unfortunately this lets the NPCs then stand in an pose with abducted arms after reloading the corresponding save. Because of that I'm only using this command when standing really near to the focused actor. If "tai" is used or not will be stated by an respective comment.

Scene 21: Female character, indoor, ambient lighting
Inside the Winking Skeever in Solitude. Standing in front of the bard "Lisette".
Time: 1:00 PM
Comparibility: 90%
TAI: not used
The first character screenshot in the series. Here's only the ambient lighting present because there's no real lighting source nearby (besides a few candles maybe). Should give an indication how characters are lightened with really sparsely lighting around them.

Scene 22: Male character, exterior, lateral day lighting
In Whiterun, outside of "Belethors General Goods". Standing in front of "Jon Battle-Born".
Time: 12:30 PM
Weather: SkyrimClearTU_A, ID 10e1ee
Comparibility: 98%
TAI: used
Good shot to examine the character look with from an lateral angle. Also the main shot to compare the AO settings on characters.

Scene 23: Female character, interior, direct lighting
Inside the "Bannered Mere" in Whiterun, standing in an angle towards "Olfina Gray-Mane".
Time: 12:30 AM
Comparibility: 90%
TAI: not used
A shot where are no really shadowed area on the character faces in most of the presets. That's because of the fire in the centre of the room.

Scene 24: Male character, interior, strong lateral lighting
Inside the "Dead Mans Drink" in Falkreath, standing in front of "Dengeir of Stuhn".
Time: 12:00 AM
Comparibility: 98%
TAI: used
This is a scene with an really harsh contrast between the lightened and the darkened side of the characters face. That's coming from the direct lateral ligthing because of the fire on the left. Contrast on faces can't be much stronger than here.

Scene 25: Female character, exterior, ambient lighting
On the Fellstar farm in Ivarstead, standing with an angle towards "Fastred".
Time: 9:20 AM
Weather: SkyrimClearTU_A, ID 10e1ee
Comparibility: 90%
TAI: not used
Another scene in which the face of the focused character is only lighthened by ambient lighting.

Scene 26: Female character, exterior, direct lighting
In Dragonbridge, standing in front of "Faida" (manually spawned, ID 00019a28).
Time: 9:30 AM
Weather: SkyrimClearTU_A, ID 10e1ee
Comparibility: 98%
TAI: used
Here you will see a tendentially bright character face becaus the sun is standing quite deep.

Scene 27: Female character, exterior, mixed lighting
At the backyard of Jorvaskr in Whiterun, standing in front of "Aela the Huntress" (manually spawned, ID 0001a696).
Time: 12:30 PM
Weather: SkyrimClearTU_A, ID 10e1ee
Comparibility: 95%
TAI: not used
The only character screenshot where probably some sort of "mixed lighting" can be noticed. How much exterior lighting is present, depends of course on the used configuration (but I've seen quite some differences in this till yet).

Now using the console command "tdetect" to disable the detection for enemies.

Scene 28: Skeleton head, interior, ambient lighting
Standing in front of an skeleton inside the Halls of the Dead in Solitude.
Time: 12:45 PM
Comparibility: 85%
TAI: not used
Probably the most "inaccurate" character/creature scene in terms of comparibility. It's impossible to use tai here (even when standing really near to the skeleton) because the glowing eye effect won't be stored in the save (you would then see some unlit eye sockets. Anyway this should give another nice indication how ambient occlusion looks on creatures.

Scene 29: Spriggan, interior, ambient lighting
Inside the Eldergleam Sanctuary, standing in an angle towards an Spriggan (manually spawned, ID 00023ab9).
Time: 11:10 AM
Comparibility: 90%
TAI: not used
Another scene in which the Spriggan tends to look quite vibrant. Also a good shot to examine the general darkness in an dimly lit interior (background).

Scene 30: Atronach, exterior, mixed lighting
At the Bleakwind Basin, standing next to the big fire in an angle towards an Frost Atronach (manually spawned, ID 0003053f).
Time: 2:00 AM
Weather: SkyrimClearTU_A, ID 10e1ee
Comparibility: 95%
TAI: used
A very nice scene because of the big contrast between the cold, blueish looking atronach and the warm red fire in the Background. Tends also to show some vibrancy in it (but not as much as in scene 29).

So... this are the screenshot sceneries at the moment. More will probably follow over time but this will need alot amount of pre-testing. In the meantime I'm pretty happy with the situation, to be frankly.

One additional word on the used tool for taking the screenshots: I'm neither using the integrated screenshot function of ENBs nor the of the eventually avaiable SweetFX-function. That's because none of them seems to taking care of ingame adjustments to the brightness (NOT fGamma, but the Slider under "Display" options). So I'm taking the shots with the well-known FRAPS as it takes every screen exactly as ingame (reference-post for this).


7. How the benchmarks are done

Beside the screenshots the benchmarks are of course the second great focus of this whole work. I now want to describe in a more detailed way which benchmark scene I've chosen and what are good preconditions for adequate benching.

The scenerie I'm using is as simple as it can be - start a new game and go through the intro sequence. This scene has several advantages over "running along a pathway on foot":

• quite (NOT most) demanding in terms of performance
• two really common scenes in one run: alot of trees at the beginning, more village-like structures at the end
• by far the biggest: possibility to (nearly) eliminate a influence of the user to the benchmark results

Of course there are propably better scenes for benchmarks in Skyrim (e.g. Solitude exterior area) but especially the third point makes the difference for me. Maybe I'm able to find a way to add another appropriate benchmark scene, but this will take alot (lot, lot^^) time for testing and verification of the scenery stability.

For actual benching the presets I'm using mainly two tools at the moment: One is MSI Afterburner. It's logging nearly all of the relevant data you can sse in the benchmark spreadsheet. How to interprete the particular parameters is explained greatly detailed in chapter 7. The second tool I'm using is Elys' MemInfo. This is monitoring the RAM usage of the TESV exe. Unfortunately the data can't be logged this way, but maybe there's another solution (that will take quite a long time for sure to implement).


7. Interprete the benchmarks the right way

IMPORTANT: Even I'm describing the data very detailed in my opinion, everyone should bear in mind that benchmarks are of course greatly depending on the used system including (but not limited to) the used settings in your inis, the state of the OS, fragmentation of HDDs, OC/not OC, background processes and much, much more. Due to this statements such as "reliability" should be taken with some caution because they are a semi-objective classification based on my own observations.

Even more important: My system tends to run a lot of the ENBs (in "minimal configuration") with an FPS over 60. When VSync is enabled, this would undermine an appropriate and good comparable benchmark series. Therefore I'm always checking that VSync is disabled in the configs or otherwise forcily disabling it. It should be noticed that this approach isn't recommendable in the uttermost of cases. You would probably suffer some weird issues with the havok engine and other things when doing this during "normal gameplay". So as I said, only for the sake of performance comparison.

Comment about repeatability: This is an estimations due to my observations over some multiple benchmark results of the same presets. Should therefore not be confused with statistical terms as "standard deviation", "measuring accuracy" or such.

So in this chapter I'm going about every parameter you can find in the spreadsheet benchmark and describe how it should be compared to your
personal setup.

FPS
Recorded: average and minimum value
Reliabilty: middle
Repeatability: +/-1 FPS for average, +/-3 FPS for minimum
Of course this is the main parameter for comparing the performance of different presets and I'm pretty sure it's the most interesting parameter for the majority of users. The own results should be (compared to others) not too much differing from mine if your hardware is pretty new. For older or weaker hardware you will likely experience greater performance drops.

GPU temperature
Recorded: max
Reliability: low
Repeatability: +/- 2°C, tends to rising over a longer time (I'm not cleaning my fans every week...)
The temperature parameters shouldn't really be compared between different ENBs because they're way too much dependent on the individual system. In fact they're mostly there to show that my GPU isn't anywhere near the temperature where throttling sets in.

GPU workload
Recorded: average and maximum
Reliability: middle to high
Repeatability: +/- 1% over a average workload of 95%, more deviation possible at lower workloads
This parameter should give an feeling how much the preset is taxing your GPU. Most of the ENB presets tend to having a quite high average workload above 95%. The maximum value is in fact only displayd to show the GPU can reach 100% workload (I'm thinking about an alternative evaluation of this). Should be examined in conjunction with GPU clock.

GPU clock
Recorded: average and maximum
Reliability: middle to high
Repeatability: +/- 1MHz over a average clock of 1025Mhz, more deviation possible at lower clocks
Similar to workload, the average value should give an estimation about how much of the time my GPU is running nearly or at 100% clock.

GPU VRAM usage
Recorded: average and maximum
Reliability: middle
Repeatability: +/- 15MB for average +/-25MB for maximum
Probably the most important parameter besides the FPS for most of the users. Of special here may be the maximum value if someone want to exclude lagging due to VRAM overflow.

CPU workload
Recorded: average only
Reliability: low-middle
Repeatability: +/-3%
This is shown mainly to depict that my system isn't running anywhere near the CPU limit. A quite simple rule becomes here obvious (though not proven in an statistical way): the more performance friendly your used ENB is, the higher your CPU workload will become (if VSync turned off of course).

CPU temperature:
Recorded: maximum only
Repeatability: +/-4°C
Pretty much the same as the GPU temperature. Not much more to say here.

RAM usage of TES5.exe (MemInfo)
Recorded: maximum only
Reliability: middle-high
Repeatability: +/-5MB
Here you can find the RAM usage for the TES5.exe in the way it was measured in releases till 1.2.2. It isn't as accurate as it could be (especially because I can't log or analyse the data) and from 2.0.0 on I've found a way to enhance this part. Nonetheless this parameter will remain in the benchmark spreadsheet for quite some time because I haven't updated to the new measuring method yet.

RAM usage of TES5.exe / enbhost.exe (Perfmon)
Recorded: average and maximum for each
Reliability: middle-high
Repeatability: +/- 5MB
I've searched for an reliable way to measuring (and more important, logging) the RAM usage of both the TES5 and enhost since the beginning of this comparison series. It has taken quite some time, but finally I'm able to log this data with the merely known (at least in the desktop segment) Windows-native tool "Perfomance monitor". It can be open by searching for "perfmon.exe". The application will open and it's then possible to log quite a massive amount of data. At the moment I'm only interested in one special parameter for both of the Exe's: the Working Set.

Comment about that: It should be noted that the Working Set is considered as a relatively unaccurate way of measuring the RAM usage of an process, mostly because it's including memoy-mapped files and other values. I would normally track the Private Bytes but decided against that because "Working Set" is the parameter that's viewable through the task manager (and the MEM-value MemInfo is reporting ingame). Maybe that'll change in later releases but it would be a parameter for more advanced users and these ones are not the main audience of such comparisons in my opinion. There's a systematical measurement deviation between the value MemInfo gives an the one from Perfmon (Perfmon is always a bit higher then MemInfo) but it isn't to big (normally in an one-digit-range).

9. What a ENB should have to be covered here
IMPORTANT: The following small chapter shouldn't be treated as a real, strict and unchangeable guideline. But with each "requirement" a preset fulfills the chances for taking it up into the series are increasing. But as I said, also a preset with a download count under 300 or a pretty new one will be benched probably.

Indicators I'm looking at when considering to add a new preset:

• Downloads/endorsement: for sure pretty obvious because this implies how much prevalent the preset is. This doesn't mean a preset with an high download count will always be benched (the best example for this may be Sharpshooters Extreme Graphics ENB...)
• Support by the author: for sure it's better when the author of the considered ENB is actively supporting his work. It should be noted that I'm distinguishing between "not actively supported" and "officially abandoned" - the last category are likely to be not compared after a longer time.
• Used ENB-dll: It's probably not really needed for any preset to use always the most actual ENB dll's. But it's a great advantage if they're doing so (at least in my opinion). I'm not really having a "hard line" of an version number here, but any preset that's using an 1xx dll will surely NOT be included.
• Time since last updated: Related to the first point, this is another parameter. Should be pretty self-explaining.

10. Credits/thanks

Boris Vorontsov for ENB
Of course all of the authors from the used presets (listed in chapter 3).

all other authors from additional used mods in conjunction with the presets:
Enbhelper.dll by Alexander Blade
Enhanced Lighting for ENB by JawZ
More Dynamic Shadows with Striping Fixed by mcm2014
Skyrim Particle Patch for ENB by Mindflux

Climates Of Tamriel - Weather - Lighting - Audio by JJC71
Climates Of Tamriel - Weather Patch by prod80 - Aiyen - Saerileth
Dark Dungeons for ENB by Mindflux
Enhanced Lights and FX by anamorfus
ETHEREAL CLOUDS by Gamwich
Natural Lighting and Atmospherics for ENB by Confidence-Man
Post-Process Effects Addon for ENB by ZeroKing
Pure Weather by Laast
Relighting Skyrim by NovakDalton
Remove Interior Fog V2 by rgabriel15
Revamped Exterior Fog by Greg Manthey - gargorias

CBBE - Caliente's Beautiful Bodies Edition by Caliente: Slim, Vanilla, Ultra Textures
SG Female Face Textures Renewal by Hello Santa
XCE - Warpaint and Dirt by Xenius
CMO - Complete Male Overhaul by Mok Chaoticran (Quality version)
High Res Face Maps for Men by Geonox
Frost Atronach HD Retex by Yuril

Author of the guide at skyrim-comparison
Tannin for creating Mod-Organizer
Casmithy for creating ENBManager

in special

jafin16, prod80, mangaclub for extended and relentless support when searching for failure issues and for reviewing the screenshot series of their presets
ao2thend for inspiration on an special scenerie

11. Permissions

That will be pretty short...

If you're an author of an ENB that's compared here you can of course use the screenshots for your description site or similar purposes. No need to ask me for permission. Anyway i would be greatly interested in knowing of it

For the spreadsheets normally the same thing applies - unless there's code included in form of more sophisticated VBA scripts or similar things. In this case you've to ask me first - mostly because some of the code isn't probably mine.

 

 

If you cannot access the download on the Nexus I've set up an alternative link here:

https://drive.google.com/open?id=0B5rtgpvrfoeWfkFRQ2N0RXN4NHhoRXRiMTVjUDQtOTNoUk52NjhHMzZXeDNNWVpFNEd1NFU&authuser=0

Bear in mind, that the archive there might not be updated that often (in comparison to the Nexus archive).

 

[Kastagir]

Any thread topic with the words "ENB" and "benchmarks" in the title is a lost cause.  To evaluate Skyrim ENB presets on performance is an exercise in ignorance.  ENB binaries are a Direct3D wrapper that runs outside of the tesv.exe process.  It must do this because Boris doesn't have access to the source code and as such, it will never be optimized for performance.  All efforts by the author of ENB binaries have been aimed at visual quality and this is the only valid metric for evaluating ENB presets.  Any presets that claim to be performance friendly simply have effects disabled.  You can do this yourself with any preset.  ANY.  PRESET.  You would be much better off looking at how individual presets address image quality in different conditions in-game (weather, time of day, interior/exterior, etc.) and provide a detailed guide on tweaking any preset to achieve a desired performance level.

[ao2thend]

Any thread topic with the words "ENB" and "benchmarks" in the title is a lost cause.  To evaluate Skyrim ENB presets on performance is an exercise in ignorance.  ENB binaries are a Direct3D wrapper that runs outside of the tesv.exe process.  It must do this because Boris doesn't have access to the source code and as such, it will never be optimized for performance.  All efforts by the author of ENB binaries have been aimed at visual quality and this is the only valid metric for evaluating ENB presets.  Any presets that claim to be performance friendly simply have effects disabled.  You can do this yourself with any preset.  ANY.  PRESET.  You would be much better off looking at how individual presets address image quality in different conditions in-game (weather, time of day, interior/exterior, etc.) and provide a detailed guide on tweaking any preset to achieve a desired performance level.

 

No.  

 

First of all, there are plenty of ENB's that don't hit performance that badly and have all the effects enabled.  While the d3d9.dll will remain the same, each ENB has a multitude of different .fx files, different lighting and weather plugins, and different settings in the enbseries.ini.  ENB's can also use SMAA and SweetFX in addition to the ENB itself.  All of these things effect performance differently and can, in some cases, effect performance very drastically.  

 

But don't believe me, test it for yourself (because its painfully obvious that you haven't really done any real testing or any work on the subject whereas Gunthral and myself have both done extensive ENB comparison archives and therefore have a better understanding then you).  

[...0...]

I must agree with kasagir on this one.

 

Ive tested also alot and it depends on so many variables its almost imposible to copy it.

 

Oh btw all who have high end systems 99% these days dont play on 1080 displays.

 

I got 40-45 fps 1440p super heavy modded with almost everything 4k textures.

 

Nice project btw and your time invest still enjoyed it.

 

Guys don't take it always so personal its just forum where you reply with your opinion that's all its after all a open forum where we have some freedom to express our views on the matter right:)

[Guest]

 

 

Ive tested also alot and it depends on so many variables its almost imposible to copy it.

 

useless blabla

 

if i take a screen without enb, and another one with k enb, i will have less fps

and it will be the same for you, your so many variables can't change that

 

almost everything 4k, so what?

 

 

does it look bad?

 

that chair don't have 4k textures (but it's high poly, i am still waiting for the huge fps impact some told me about)

it's just 2k

2k again

rosenspine is 4k (forgot to resize it)

maybe i'll make a 4k, 2k and 1k comparaison later

 

 

 

no wonder so many don't bother releasing what they do, if when someone do it, the only comments are to say it's crap

 

What are you even talking about?

Evasia was only saying that this is a nice project and he appreciates the time OP spent to do the comparison, but he doesn't believe a certain ENB behaves the same way on different specs. Why bringing your 4k textures and outfit ports to the discussion?

And you haven't release any of your outfit ports or share them, just saying...

[yatol]

 

 Why bringing your 4k textures and outfit ports to the discussion?

 

it's not me that have almost everything 4k

I got 40-45 fps 1440p super heavy modded with almost everything 4k textures.

 

 

 

And you haven't release any of your outfit ports or share them, just saying...

 

 

bad guess, look for the unp top model thread...

it's my crappy conversion to demonica you can't have (i don't like krista fanboys)

[Rayblue]

All right. Something I thought of -- I believe that some users use several presets depending on the mood or if it's for normal gameplay or for super-high-resolution screenarchery; the former as they dial down the quality and disable certain effects, the better to be able to fight, have more responsive gameplay and still have decent graphics; the latter (using, say, any of the K presets), their FPS is literally cut down to several notches but have superior visuals for screencapping.

 

It's rare to have both unless one runs a terribly ultra-high-end gaming machine.

[Fuzaki]

No need to get so hostile guys.

[Guest]

 

 

 Why bringing your 4k textures and outfit ports to the discussion?

 

it's not me that have almost everything 4k

I got 40-45 fps 1440p super heavy modded with almost everything 4k textures.

 

 

 

And you haven't release any of your outfit ports or share them, just saying...

 

 

bad guess, look for the unp top model thread...

it's my crappy conversion to demonica you can't have (i don't like krista fanboys)

 

Cool. I got things mixed up there, and I'm sorry for that. But still..

The guy was obviously bragging about his rig, but does he insult the OP? No

Does he state that everything less than 4k is crap? No

Does he call out your outfit conversions? No

Then what is the point of those images? You didn't even show your fps.

[yatol]

 

Then what is the point of those images? You didn't even show your fps.

 

why do you want to see my fps if fps comparaison are useless?

and what is the point of your answear?

 

i have to choose when i convert a 3000*3000 jpg to dds

and bigger don't mean better, most of the time i pick 2k

 

want more useless images to see why?

 

 

 

 

 

on one screen, battlemaiden is 150 or 200 mo of textures, on the other one it's just 32 mo (did a little more than resize down the textures)

[Gunthral]

Hello guys,

 

thanks for the feedback and also the "complaints" (I'm personally seeing them more as constructive criticism.

I want to answer to some of the issues and objections that were given.

 

Any thread topic with the words "ENB" and "benchmarks" in the title is a lost cause.  To evaluate Skyrim ENB presets on performance is an exercise in ignorance.  ENB binaries are a Direct3D wrapper that runs outside of the tesv.exe process.  It must do this because Boris doesn't have access to the source code and as such, it will never be optimized for performance.  All efforts by the author of ENB binaries have been aimed at visual quality and this is the only valid metric for evaluating ENB presets.  Any presets that claim to be performance friendly simply have effects disabled.  You can do this yourself with any preset.  ANY.  PRESET.  You would be much better off looking at how individual presets address image quality in different conditions in-game (weather, time of day, interior/exterior, etc.) and provide a detailed guide on tweaking any preset to achieve a desired performance level.

 

I personally can't really agree in that matter with you - at least in some matters.

Yeah, of course ENBs will never give any user an "FPS boost" or anything like that. Simply because of the technical circumstances you're describing. But does this really mean the can't be compared? I certainly don't think so. I'm saying that due to various thoughts:

1. At least they can be compared relatively to another. The spreadsheets have to be read in the right way: So the way should not be saying "K-ENB runs at 20FPS", "Noir ENB needs 1800MB VRAM" or anything similar. Instead it should say: "K-ENB is more performance-hungry than RealVision." So the judgement should be based on relatively differences in performance, not the absolute values. Even the relative differences should be treated with some care because poorer hardware will likely give a greater performance drop. But I'm already writing this in the description (maybe not clear enough, will change that).
2. How would you compare the visual quality? That's simply not possible and by far even more than performance comparisons. Why? Yeah, performance comparisons are at least objective (relatively). I'm taking the data outside my logs and simply displaying them/maybe calculating some depending parameters. How should a visual comparison take the quality into account? Simply from the activated effects? Therefore a ENB with a few disabled effects would be automatically inferior to an "full featured" preset? How about users that're simply not liking DOF effects, lensflares or anything other (just an example)? That surely won't never be my scope with this series. Simply because I want to show the presets in an way that the corresponding author has thought for them. But in the same time preserve the installation as simple as possible (I'm trying to stick to the authors description and installing only the "really needed" mods). That leads me to...
3. Of course it will be possible to tweak any of the presets out there to your personal likings. Even maybe by writing them new from scratch. But would someone need a comparison like this? Prette sure he won't... I'm totally aware of that and therefore logically not doing this mainly for the advanced users but instead for the "casuals", users that're probably looking for ENBs the first time, having somewhat older hardware or similar things. Although I'm knowing that exactly these users are most likely to interpret the data in the wrong way... but I'm working on some things that hopefully should help with this a bit.
4. Also a guide for tweaking presets? Sorry, but there are guides out there that're by far more superior than anything I could probably write and... looking in point 3

 

I must agree with kasagir on this one.

 

Ive tested also alot and it depends on so many variables its almost imposible to copy it.

 

Oh btw all who have high end systems 99% these days dont play on 1080 displays.

 

I got 40-45 fps 1440p super heavy modded with almost everything 4k textures.

 

Nice project btw and your time invest still enjoyed it.

 

Guys don't take it always so personal its just forum where you reply with your opinion that's all its after all a open forum where we have some freedom to express our views on the matter right:)

 

Thanks for feedback

 

I'm pretty there with you when it comes to comparibility of the benchmarks to different PC configurations.

But I can assure you that at least for 3 configurations (my own of course, another config with GTX 480 ane a laptop with GTX680M SLI) the benchmarks are pretty good comparable relatively to another. Yeah, the 480 will give a greater relatively drop than my actual 290X, but the "order" of relatively performance-light ENBs to the heavy ones were looking pretty the same (I'd already did something similar a year before on the older two systems, although by far not so extensive).

 

All right. Something I thought of -- I believe that some users use several presets depending on the mood or if it's for normal gameplay or for super-high-resolution screenarchery; the former as they dial down the quality and disable certain effects, the better to be able to fight, have more responsive gameplay and still have decent graphics; the latter (using, say, any of the K presets), their FPS is literally cut down to several notches but have superior visuals for screencapping.

 

It's rare to have both unless one runs a terribly ultra-high-end gaming machine.

 

Exactly that. Some users maybe want to have a decent looking preset for shooting screens. But in the same time they might want another ENB that's more suitable for regular gaming. For both of the cases I want to give an decision guidance, although I'm not really facing towards pure high quality shots. How much do you see of an particular ENB when there are dozens and dozens of texture mods are on it? That would change the whole scope of this and I'm not willing to do extensive installations for every single ENB... also the benchmarks... is this drop now from the ENB itself or mods such as SFO, Grass mods or similar?

 

Anyway, as I said, i really appreciate the feedback and I'm aware of some of the thoughts that were brought up here. I'll try to support especially inexperienced users with some advanced features to interpreting the benchmarks. Although this will take alot of time.

 

Greetings Gunthral

[pinky6225]

Should add Minou's Glossy ENB 1 http://www.loverslab.com/files/file/1334-minous-glossy-enb/ as its on LL in the downloads

[Gunthral]

Should add Minou's Glossy ENB 1 http://www.loverslab.com/files/file/1334-minous-glossy-enb/ as its on LL in the downloads

 

Thanks for suggesting that

I'll probably consider to adding it although then it won't be on the Nexus.

Also I've to admit that ENBs outside the Nexus aren't my priority at the moment (but as I said, thanks).

[...0...]

No need to get so hostile guys.

 

I wasn't, just point some things out about topic, others here seems upset, as always sadly:(

[pinky6225]

 

Should add Minou's Glossy ENB 1 http://www.loverslab.com/files/file/1334-minous-glossy-enb/ as its on LL in the downloads

 

Thanks for suggesting that

I'll probably consider to adding it although then it won't be on the Nexus.

Also I've to admit that ENBs outside the Nexus aren't my priority at the moment (but as I said, thanks).

 

 

As your posting on the LL forum and not nexus it would make sense to include the ENB thats only available from the LL forum imo

 

That ENB also can be copied and pasted into the install directory (unlike other ENB's from the nexus where you have to go get the binary file from Boris website as he didn't want anything hosted on nexus) and be ready to use so its most likely the easiest ENB to install so would be good for new mod users.

 

[yatol]

 

I wasn't, just point some things out about topic, others here seems upset, as always sadly:(

 

 

it would be usefull to do some real testing to know what you are talking about then

 

oh well, let's crush the 4k urban legend, and the high poly urban legend

 

 

 

 

here's the screens with the fps, as blabla is just useless blabla (i don't even have 40 fps without enb, and i don't care about it)

 

 

 

 

 

 

if someone is looking at this, this is why the decimated version cost much less fps

well... didn't decimated everything, small things need high poly to look as good, so i delete the decimated version before exporting the nif

but i forgot to put back the undecimated version in the decimated nif (that part is on the arms, arms eat a lot of fps)

 

 

 

 

high poly kill fps? nope, it's the number of bones attached to those verticles that have an impact on fps

 

angeloi is a 20 mo nif and a 17 mo nif

the 20 mo nif eat about 1 fps (just the dress)

the 17 mo nif eat about 2.5 fps (legs and arms, arms have a lot of bones, so less verticles have a bigger fps impact)

 

 

the 110 mo difference in textures, that's 4k resized to 2k, and some 2k resized to 1k, and everything compressed

someone can tell what is what?

 

 

 

 

 

 

 

 

 

 

 

 

 

there is a difference, but the nif surface is so small it's really hard to see it without zooming

so using 4k for almost everything... /facepalm if someone really do that

[mindqontrol]

Any thread topic with the words "ENB" and "benchmarks" in the title is a lost cause.  To evaluate Skyrim ENB presets on performance is an exercise in ignorance.  ENB binaries are a Direct3D wrapper that runs outside of the tesv.exe process.  It must do this because Boris doesn't have access to the source code and as such, it will never be optimized for performance.  All efforts by the author of ENB binaries have been aimed at visual quality and this is the only valid metric for evaluating ENB presets.  Any presets that claim to be performance friendly simply have effects disabled.  You can do this yourself with any preset.  ANY.  PRESET.  You would be much better off looking at how individual presets address image quality in different conditions in-game (weather, time of day, interior/exterior, etc.) and provide a detailed guide on tweaking any preset to achieve a desired performance level.

 

Well you're kind of right, performance would actually be the least relevant thing to choose a preset on, if people knew how to tweak config files to get FPS back (why is actually both easy and fast once you know how, that's the funny thing). But they don't, since there's little to no guides on the interwebs for that, so people don't know, and they're stuck to using whatever runs okay on their set-up. (I almost shed a tear everytime I see "I tried K ENB/Kountervibe/Serenity/any good-but-heavy preset that's act and I went back to RealVision because of FPS").

 

 

Any thread topic with the words "ENB" and "benchmarks" in the title is a lost cause.  To evaluate Skyrim ENB presets on performance is an exercise in ignorance.  ENB binaries are a Direct3D wrapper that runs outside of the tesv.exe process.  It must do this because Boris doesn't have access to the source code and as such, it will never be optimized for performance.  All efforts by the author of ENB binaries have been aimed at visual quality and this is the only valid metric for evaluating ENB presets.  Any presets that claim to be performance friendly simply have effects disabled.  You can do this yourself with any preset.  ANY.  PRESET.  You would be much better off looking at how individual presets address image quality in different conditions in-game (weather, time of day, interior/exterior, etc.) and provide a detailed guide on tweaking any preset to achieve a desired performance level.

 

No.  

 

First of all, there are plenty of ENB's that don't hit performance that badly and have all the effects enabled.  While the d3d9.dll will remain the same, each ENB has a multitude of different .fx files, different lighting and weather plugins, and different settings in the enbseries.ini.  ENB's can also use SMAA and SweetFX in addition to the ENB itself.  All of these things effect performance differently and can, in some cases, effect performance very drastically.  

 

But don't believe me, test it for yourself (because its painfully obvious that you haven't really done any real testing or any work on the subject whereas Gunthral and myself have both done extensive ENB comparison archives and therefore have a better understanding then you).  

 

Honestly you're not really refuting much of what he said, most of the FPS difference between a heavy ENB and a light one using the same ENB binary version indeed comes from effects being enabled or disabled and their respective qualities (and almost of them are controlled in enbseries.ini), and depth of field quality (which you can tweak easily too in enbeffectprepass). The additional use of optional ENB shaders like lens, sunsprite is minor stuff, same for SweetFX, the only "heavy" thing in SFX is SMAA (so that's like 3-5 fps ? depending on settings used), there's also Bloom and HDR I believe but people rarely ever use them because you can do that better in ENB already.

 

 

 

I wasn't, just point some things out about topic, others here seems upset, as always sadly:(

 

 

it would be usefull to do some real testing to know what you are talking about then

 

oh well, let's crush the 4k urban legend, and the high poly urban legend

 

 

 

 

here's the screens with the fps, as blabla is just useless blabla (i don't even have 40 fps without enb, and i don't care about it)

 

 

 

 

if someone is looking at this, this is why the decimated version cost much less fps

well... didn't decimated everything, small things need high poly to look as good, so i delete the decimated version before exporting the nif

but i forgot to put back the undecimated version in the decimated nif (that part is on the arms, arms eat a lot of fps)

 

 

 

high poly kill fps? nope, it's the number of bones attached to those verticles that have an impact on fps

 

angeloi is a 20 mo nif and a 17 mo nif

the 20 mo nif eat about 1 fps (just the dress)

the 17 mo nif eat about 2.5 fps (legs and arms, arms have a lot of bones, so less verticles have a bigger fps impact)

 

 

the 110 mo difference in textures, that's 4k resized to 2k, and some 2k resized to 1k, and everything compressed

someone can tell what is what?

 

 

 

 

there is a difference, but the nif surface is so small it's really hard to see it without zooming

so using 4k for almost everything... /facepalm if someone really do that

 

No idea what you were really trying to prove there (measuring polycount impact on FPS would be better done by comparing SMIM vs no SMIM in RIverwood or something than by just comparing with and without a single massive daz armor mesh tbh) but the bone stuff was still interesting lol

[yatol]

 

measuring polycount impact on FPS would be better done by comparing SMIM vs no SMIM in RIverwood or something than by just comparing with and without a single massive daz armor mesh tbh

 

 

i don't think so

 

if you want a smim vs no smim comparaison, do it yourself

[mindqontrol]

lol was just making a remark, yeah if I cared so much I wouldn't wait for somebody else to make it and I would do it myself.

[Gunthral]

Uploaded a new update yesterday, featuring the new Serenity ENB - Extended FX and 8 updated ENBs.

 

Overall are 31 presets updated to the new 30-shots-series.

There won't be any update over the weekend because I'm not at my desktop.

Anyway it should be said that the mentioned "Minou's Glossy ENB" will be added in the next version (as a separate, nexus-externally download, obviously).

 

Greetings Gunthral

[yatol]

a small test with 0 esp

 

 

 

enb on

 

 

no textures

 

 

there's no mod, don't know what is eating the fps when you coc whiterunexterior01

enb off

 

 

with some lods and... river

 

 

that water is only 2k, but the mesh is so big... huge fps impact

and you are loading it even if you can't see it

 

if some are wondering why riverwood eat so much fps, the river mesh is 3 or 4 time the size of the river

 

 

if you load the game without some mesh

 

 

well... that was stupid, you fall when there's nothing under your feets^^