OBS Studio - really basic tips for x264 encoder

my interest is in offering a configuration capable of transmitting a single SDR content at 1280x720, 60 fps, on processors over 10 years old, like the Intel Core i5-3230M which was tested for the production of this document.

menu: settings, advanced


color format.: nv12 (8-bit, 4:2:0, 2 planes)
color sapce..: rec. 709
color range..: limited

we are working with content in 8-bit HD SDR and therefore this is the recommended configuration.

live streaming of HDR content content will use different settings not covered here.

menu: settings, video


base (canvas) resolution...: 1280x720
output (scaled) resolution.: 1280x720
downscale filter...........: (none)
common fps values..........: 60 (or 48)

this settings allows for higher image quality and lower processing costs at the same time. which is very rare. but this magic only happens when the resolution is the same in all 3 elements, namely: capture content, canvas and output.

as the base/canvas resolution is the same as the output resolution and as long as the captured content is also in the same chosen dimensions, then, there is no downscaling or any additional processing cost.

because of this we are saving processing that would otherwise be carried out by the GPU, and would appear as a 'rendering' cost by OBS. on the other side, downscaling is in no way heavy for current computers, but for very old and not very powerful computers it is a strategy that we can use.

values ​​below 60 fps also increase performance but at the same time they do not allow capturing some false transparency effects used in retro games, such as the effect of the Rolling Shield weapon when fully charged from Mega Man X on SNES. for games on newer consoles, this type of false transparency effect is no longer used and so 30 or even 24 fps become acceptable values.

intermediate frame rates such as 48 fps provide incredible fluidity for the viewer and significantly reduce the load on the encoder and rendering processes. don't be afraid to try them out.

if the aforementioned scenario is not possible, the alternative:


if the monitor or captured content is larger than 1280x720p, it is best for it to be captured in its native resolution and then make explicit use of one downscaler filter.

base (canvas) resolution...: 1366x768* (native monitor resolution)
output (scaled) resolusion.: 1280x720
downscale filter...........: Area (Weighted sum, 4/6/9 samples)
common fps values..........: 60 (or 48)

OBS Studio currently provides 4 downscaler filters, namely: 

bilinear, area, bicubic, lanczos (listed in order of quality and processing cost.)

for retro content, the area filter is usually good and light enough.

although the bilinear filter is a bit blurry and doesn't bother me for gaming content, it may be bad for reading text displayed, especially in menus and applications, outside of game content.

bicubic and lanczos tend to produce higher quality in general, but they also tend to require more powerful or minimally newer GPUs.

don't be afraid. there are cases where the result of the bilinear filter is completely acceptable. it's just that the area filter is a little more reliable and equally light. and hence my default recommendation.

output resolutions such as 852x480 and 640x360, when properly compressed, also offer good visibility for the viewer. give them a try on more modest hardware as well.

menu: settings, audio

sample rate.: 48 kHz
channels....: stereo

the important thing is that you preferably force your audio card driver and apps to use the same sample rate chosen here to avoid any additional processing being used in conversion.

codecs like Opus use 48 kHz by default and also it is the YouTube recommendation. although 44.1 kHz can also be used, I would prefer to recommend 48 kHz in all location: OBS Studio, audio card driver settings, multimedia players, emulators, etc.

menu: settings, video, streaming (advanced mode)

the heart of this publication. s2 <3

# streaming settings
audio encoder.....: libfdk AAC or any AAC
video encoder.....: x264
rescale output....: disabled
(resolution)......: (irrelevant)

# encoder settings
rate control......: CBR (always CBR for live streaming)
bitrate...........: 6,000 kbps (to the moon!)
keyframe interval.: 2 s (always 2 s for live streaming)
CPU preset........: veryfast
profile...........: high
tune..............: film
x264 options......: bframe=0

audio encoder. libfdk AAC is the best possible option for audio, but it's not that relevant. it may not be present in your installation due to licensing issues and can be obtained by manually compiling OBS Studio. which I don't recommend, but I'm pointing you towards the stone path if that's your interest.

video encoder. for this recommendation, we will use x264, which is processed by the CPU. potentially, depending on the settings used, the x264 encoder is capable of providing better image compression than any native GPU H.264 encoder.

however, this comes at a cost. achieving such high levels of compression is extremely resource-intensive and requires a new and incredibly powerful CPU to perform it in real-time during online broadcasts.

for offline video conversions where conversion time is not a concern, x264 can be used to achieve better compression.

however, since we are talking about live conversion, using GPU encoders typically has less impact on computer performance while yielding equally satisfactory results. and so, GPU encoders are recommended for live streaming, and mainly for 1080p resolution or higher.

in this text, the user's computer is expected to be quite old and lacking a GPU capable of video conversion. therefore, the recommendation is to use x264 for resolutions up to 720p.

rescale output. here again the "rescale output" option appears and it is different from the original option in another menu already listed previously.

before, the operation was considered by OBS Studio as "rendering" and done by the GPU. now, on the current screen, the operation becomes "encoding" by OBS Studio and is done by the chosen encoder. in the case of x264, it is done by CPU. in the case of others encoders, normally done by GPU.

generally speaking, you won't need to use this feature almost ever. perhaps when you need to output resolutions not supported by the GPU or in the extreme case the GPU is overloaded. so, don't worry about it.

as we are doing live streaming in h.264/avc, we need to transmit the content in CBR. it happens that the x264 encoder fluctuates a lot even when choosing CBR. one way to tame this fluctuation is through the use of "custom buffer", but we won't do that here.

as my concern here is image quality and not bitrate control, I recommended not making any changes to the buffer. x264 CBR fluctuation is expected by streaming sites. there is nothing to worry about.

bitrate. bitrate is the most important metric. as a general rule you will use as much as the platform allows within the chosen resolution, and preferably no more than 50% of your actual, tested, upload bandwidth.

it is important that the bitrate used is not greater than your internal network transfer capacity, as it is expected that packets will be lost, and when lost they will be retransmitted. so, you need to have enough bandwidth to retransmit what was lost and still provide the current content at the same time.

for the vast majority of streamers, the upload capacity of their internet is much greater than the maximum bitrate allowed by the streaming platform and this is not a problem nowadays.

however, there are cases of high-speed contracted internet, in the gigabit range, but with poor internal management in the house, mainly when using simpler or generic wireless repeaters, which normally operate below the capacity of the local network and recurrently provide a measly 30 mbps or less, with a lot of interference, etc.

therefore, prefer to broadcast live using wired internet. avoid WiFi, especially WiFi 2.4GHz and WiFi from repeaters. when using wired connection you will have the possibility of using all the bandwidth allowed by the streaming platform.

otherwise, the local network will be a bottleneck and this will harm the quality of the stream or the lightness of the encoder.

finally, the sufficient bitrate for image quality depends on the content to be transmitted. scenes with rain, snow, confetti, and many small distinct and independent dots produce scenes that are difficult to compress and thus demand much more bitrate for image quality. static content, with low fps, and rare transitions almost never require a higher bitrate.

however, these behaviors change during the playback of content. a game, movie, or any multimedia content contains some segments that are easier and other that are harder to compress. an encoder configuration might produce a great-looking image on the character selection screen, but a terrible one once any action begins. 

keep in mind that the difficulty of image compression is not linear, but rather depends on the content present at any given moment. content that is constantly changing, even within the same game.

if something needs to be clear, let it be: the need for bitrate varies according to the content. even if the resolution and fps are the same, the image content may require more or less bitrate. so always try to use as much as possible.

keyframe. keyframe is always 2 seconds in live streaming. YouTube accepts up to 4 seconds, and Twitch accepted (when I was doing live there) up to 10 seconds for resolutions below 720p.

however, the recommended and safe time for live streaming is 2 seconds. use it. for professional films higher values ​​are used. for offline videos recorded on YouTube, typically lower keyframe are used. the recommendation here is for live streaming, use 2 seconds and you won't have any problems.

CPU Preset. the x264 encoder has anomalous behavior, in this case, the veryfast preset often produce the smallest video file even when compared to the veryslow preset. it doesn't mean that the image quality of veryfast is greater than that of veryslow, but there is this anomalous behavior of the x264 encoder that is good to be aware of.

there are 10 presets available in x264, namely:

ultrafast, superfast, veryfast, faster, fast, medium, slow, slower, veryslow, placebo

listed in order of processing cost and image quality, from worst to best, lightest to heaviest.

ultrafast is extremely light but does not respect the recommendations of any platform. if you use it, the transmission will probably be accepted, but it will hardly produce good image quality in more sophisticated content since the bitrate ceiling for online transmissions is low.

if you wanted to make offline recordings, and being able to significantly increase the recording bitrate, then, when using ultrafast you will have low encoder impact on the content to be captured.

but, don't use it for live streaming. it could be useful for capturing content that will then be edited and processed in non-linear video editing programs like kdenlive, DaVinciResolve, etc, later.

superfast is already compatible with the encoder recommendations of most streaming platforms, as it provides CABAC, for example, but still produces low compression. at high bitrates for retro content, its result if combined with qpmax below 34 is not bad. and remains quite light.

when using around 4,000 kbps, try using the following in x264 options:

bframe=0 ref=3 qpmax=32

it remains light, with a result that is quite pleasing for a retro style content.

veryfast is by no means fantastic, but it is light and satisfactory for the scenario proposed here: compression of retro content in 720p.

faster is better than the previous one, but also heavier on the CPU.

fast is already starting to demand a newer and more powerful CPU to use.

medium is ffmpeg's default compression, and probably requires a CPU that has twice the power of the CPU tested here for live streaming.

slow, slower, veryslow are much heavier. they produce better images in low bitrate scenarios compared to the previous ones, but a new and powerful cpu is necessary for this type of preset.

placebo is for testing or debugging purposes and has no practical use. it is extremely heavy and does not produce any significant image quality improvement over its predecessors.

level. level appears on some encoders and should never be changed from auto. there are no advantages to changing it. it must be correctly chosen and that's it. it is an "inert" option in relation to image quality. its function is to indicate which device is compatible with the content produced. the streamer doesn't have to worry about this or define anything. use auto, done.

profile. profile can be understood as the "age" of the encoder. there are 3 options available: baseline, main and high. always use high. there is no more hardware on the market that is so weak that it is not compatible with high.

at home, for private testing, if you want to send video to the first Android cell phones, you may need to use baseline. but for devices with access to current streaming platforms, high is the minimum. don't worry.

tune. normally tune in (none) works fine. but here we want to extract the best image quality in a gaming scenario, and the film option has proven to be more reliable in this scenario.

x264 Options. bframe is the intermediate frame with the highest possible compression. however, it is also the most expensive to process, and also the one that tends to blur the image most easily. in scenarios with better presets and more computing power available, bframes can be useful. 

in the context of this guide, I am recommending not using it. so, use bframe=0

using ref=2 with veryfast and ref=3 with superfast can slightly improve bitrate stability without significantly impacting processing cost. using a qpmax value below 34 can also improve image quality by forcing the encoder to use more visually pleasing image compression, at the cost of greater bitrate fluctuation.

when using qpmax=34 you should use at least 2.500kbps to avoid high bitrate fluctuation. qpmax values ​​lower than 32, such as 30, quickly become uncontrollable and fail to adhere to lower bitrates. they are initially safer at higher bitrates, such as 6,000 kbps, however, you will experience spikes in bitrate usage the lower the value chosen for qpmax is.

so, be careful when using it. and you likely won't need to lower it below 30.

menu: settings, output, audio


audio bitrate: 128

128 kbps when using AAC should be sufficient. and this value is compatible with all streaming sites. higher values ​​may or may not be accepted.

values ​​such as 160 and 192 if accepted by the streaming platform can produce higher audio quality in some scenarios. but it is unlikely that more than this will be necessary when using the AAC encoder. other encoders like Opus can use much less and preserve the same audio quality. AAC is much better than the old mp3, don't worry, 160 is an extremely safe value.

menu: settings, output, replay buffer

the feature is a kind of local clip, but it drains resources on less powerful computers. disable it in general.

menu: settings, output, recording

recording format.: matroska video (.mkv)
video encoder....: (use stream encoder)
audio encoder....: (use stream encoder)

the OBS Studio team invented a multimedia container apparently because many users wanted to have a file ending with ".mp4" even though it didn't respect the standard.

use ".mkv" and be happy. avoid later problems.

tip: if you use the same settings for live streaming and recordings, you can know the result that will be obtained during live-streaming even when recording the local video. except for possible problems with internet fluctuations that do not occur when recording offline.

however, assuming there will be no transmission fluctuations during the live broadcast, the video sent to the streaming platform and the video recorded using the same settings produce exactly the same result.

Twitch gives viewers the option to watch the live stream via the "(source)" feed, and in this case the retransmitted content is exactly the same as the content sent by your OBS Studio.

all other video options, however, both on Twitch and on other platforms such as Kick and YouTube, present content re-encoded by transcoding.

all the platforms listed produce high-fidelity transcoding, especially when the streamer uses recommended settings, and, so, it is worth to respect the requirements of each platform. in the case of Kick, however, the quality of transcoding varies depending on the user's account stage. as the channel grows, so does the quality of transcoding.

therefore, when doing the first lives, the quality of Kick's transcoding may not be wonderful, but eventually the quality will be the same as the competition.

in summary, you can use the offline recording function to know the quality that will be displayed in live streaming when the viewer chooses the best quality available on each platform. 

for viewers using video player resolutions lower than the streamer's native broadcast resolution, the video will, naturally, be of lower quality, as it is produced via transcoding in inferior properties. but the best resolutions tends to be faithful and identical to the offline recorded version.

in any case, when using local recordings you can find the limits of the hardware capacity without needing to be live.

menu: settings, general, general


check: open stats dialog on start up

OBS Studio's statistics window is probably the most important tool available for knowing what's happening with the multimedia adventure as a whole.

it may not be as intuitive, it may be a little confusing, but it is still fundamental.

after activating the feature in the first checkbox on the screen, re-launch OBS Studio so that the new window opens automatically.


average time to render frame. for 60 fps content, the value needs to be at least less than 16.6 ms. values ​​above this will certainly produce frame loss. try to create a configuration that produces values ​​below 8.0 ms to have a good safety margin at 60 fps.

frames missed due to rendering lag. "rendering" in OBS language and shown here means 3 different things: capturing images from various sources, compositing these images in a single frame, and downscaling these images when the output and the canvas are of different resolutions.

so, if you are dropping rendering frames, it means you are either using a very heavy downscaling filter, or have too many simultaneous captures, or are applying too many compositing effects to the frame, such as chroma key and other effects provided by OBS Studio.

skipped frames due to encoding lag. it means your encoder is unable to complete the task of converting all the frames generated by the rendering process in a timely manner.

therefore, it is necessary to lower the encoder preset, or the output video resolution, or the stream's frame rate.

for Windows users with Game Mode enabled, if a game utilizes more than 95% of the GPU, Windows stops honoring priority boosts for background applications, including OBS Studio, leading to issues where the encoder fails to function.

as options, the user can disable Game Mode and restart the machine, though many fear they might lose FPS by doing so. or run OBS Studio as an administrator, which is not a good security practice, especially if the user employs third-party plugins, etc.

as there is no "game mode" on Linux. it is not a concern for users of the system.

dropped frames (network) - stream / recording. if network frames are being lost, it could be either a transient issue, such as simple, momentary interferences or a sign that the network is unable to handle the selected bitrate.

it is important to test internet speed using sites like Speedtest, or local network communication capabilities using programs like iperf3. all of which are free. and as mentioned, do not use a bitrate higher than 50% of the maximum upload speed tested and supported.

updating drivers, reconfiguring the modem, changing Wi-Fi channels, forcing the exclusive use of IPv6, or conversely, forcing the use of IPv4 only, depending on the service provider, switching DNS servers, and even simply restarting the network tool suite can fix problems. the WiFiAnalyzer app from the open-source F-Droid store allows Android devices to scan for channels with less interference in real time.

finally, if the problem relates to disk writing, it could be due to file system corruption, inconsistencies in the privilege table (especially when switching between administrator and limited user accounts), or even a physical disk defect. programs for checking and testing the SMART function, surface testing, and file system repairs may be required.

tip: disable preview


it might seem trivial, but when dealing with very old computers, even disabling the preview helps save resources.

it makes no difference on new and powerful hardware... but with older gear, every bit of processing power counts.

final considerations

[a] it is important to realize that I am talking about configurations using only a single capture source and nothing else.

adding extra sources, such as a camera, browser source or some alerts, increases system load, and such configurations may not work on identical hardware or when running more demanding games.

[b] there are three distinct video codecs used in online streaming, namely:

H.264/AVC, H.265/HEVC, AV1

H.264 Advanced Video Coding (AVC) is very old, very lightweight, supported on all platforms, and suitable for content at 1080p or lower.

H.264 can usually be converted using the CPU and at up to 720p 60 fps, it isn't all that demanding.

for resolutions like 1080p at 30 fps or higher, I consider it more advantageous to opt for hardware-based compression using dedicated circuitry, typically found in high-end discrete graphics cards or newer Intel processors with updated QuickSync.

higher resolutions involve a large number of points, and the general architecture of a CPU is not efficient enough to perform this type of conversion in real time.

the x264 encoder, when utilizing the CPU, is capable of producing better video compression than the dedicated circuits found even in the most powerful consumer-grade GPUs; the challenge, however, is doing so in a timely manner. specifically, in real-time during a live stream broadcast.

at resolutions such as Full HD or higher, GPUs typically have more power for this type of real-time task, without significantly impacting the machine's performance. and, at the same time, in these scenarios, other video codecs become more attractive because they are more efficient.

H.265 High Efficiency Video Coding (HEVC) represents an evolution over the previous codec in terms of compression efficiency, but its advantages only become apparent at higher resolutions. it is too resource-intensive for CPU processing, requiring a dedicated GPU. 

it is not supported by Kick, and, apparently, on Twitch, you need to have some level of partnership on the platform to have it available for use.

AOMedia Video 1 (AV1) is the most efficient codec available for home live streaming and is currently supported only by YouTube, as far as I know, or among the platforms I follow.

it is ridiculously heavy on the CPU but supported by newer, powerful discrete graphics cards from AMD, Intel, and NVIDIA. and once again, its major advantage over H.264 lies in resolutions such as 1080p and higher.

for lower resolutions, such as 720p at 60 fps or below, the H.264 codec is completely adequate, lightweight and sufficient.

[c] there is one more important difference between encoders that use the CPU and those that use the GPU.

CPU-based encoders are improved through software updates and these updates are generally compatible with all CPUs, although the performance is significantly slower on older ones.

GPU-based encoders are not improved by software, but rather by hardware upgrades. consequently, the quality of the NVENC encoder on an NVIDIA GeForce GTX 750 Ti comes nowhere near the quality of the "same" encoder on an NVIDIA GeForce RTX 5090.

thus, the quality of the hardware encoders in current graphics cards is vastly superior to that of hardware encoders from previous generations. 

since the encoder's quality depends on the specific ASIC created, the hardware itself and the implemented design, it cannot be improved except through a hardware upgrade.

[d] however, encoders can be hybrid, and in some cases, native functions mix tasks between the CPU and GPU simultaneously. 

depending on the settings in use, in command-line programs like FFmpeg, the graphics card can be used to offload frame processing from the CPU, or, on the other side, the graphics card can enlist the CPU's help to pre-process certain frames.

in general, however, combining encoders for live streaming tends not to yield satisfactory performance. in systems with two or more GPUs, using a secondary GPU for encoding tends to reduce live encoding speed; similarly, using an iGPU alongside a discrete GPUs often produces the same detrimental result.

for best performance, opt to use the encoding circuit of the GPU handling the game as the first option, or use only CPU encoding without anything else.

[e] each streaming platform has its own quirks; in the case of YouTube, it is a good idea to manually enable the ultra-low latency setting on the site to ensure a more responsive chat during live streams.

bônus: matrixview

on Linux, install Streamlink and MPlayer, then change the command below, which is just a single line, to a live channel:

CHANNEL="https://kick.com/ofernandofilo" ; streamlink --player mplayer --player-args '-cache 8192 -vo matrixview:cols=656:rows=160' "$CHANNEL" best ;

Windows has the same programs with the same functions, but the commands might be slightly different, and I am unable to test them right now.

bônus: single thread points


you can obtain some important and reliable benchmarks from the PassMark Software website and make decisions based on them.

it is important to note that every benchmark is flawed, and the PassMark benchmark typically fails to capture the advantages of increased L3 cache.

that said, the important line in the context of this text is:

Single Thread Rating

according to the image, the Intel Core i5-3230M scores 1,500 points; the Intel N100 scores 1,800 points; and the AMD Ryzen 5 5600 scores over 3,000 points.

there are other important information to understand the capacity of a CPU, such as support for SIMD (Single Instruction, Multiple Data), but I won't go down those rabbit holes right now.

in short, a newer computer with a single-thread score of over 1,500 should be able to handle the settings and workloads recommended here.


bônus: emulation

read Emu Gen Wiki, the best source of information on emulation on the internet.

if the criticism about the Emu Gen Wiki is that it isn't perfect, nothing is.

if the criticism is that there is something better, show the evidence.

the suggested emulators are usually the best in terms of emulation fidelity and library compatibility. 

this means they aren't the most lightweight options. on the contrary, they tend to be the most demanding ones. emulation comes at a cost, a high cost, and I’m not going to sugarcoat it.


bônus: recommended operating systems

for a retro gaming scenario, without modern titles or games that use kernel-level anti-cheat, and also running on modest hardware, consider the following operating systems:

  • Windows 10 IoT Enterprise LTSC 2021 [21H2] build 19044 with security updates up to 2032.
  • Windows 10 IoT Enterprise LTSC 2019 [1809] build 17763 with security updates up to 2029.
  • rolling release linux: CachyOS, EndeavourOS, PikaOS, TuxedoOS
  • alternative rolling releases: ArchBANG, Artix, RebornOS, Solus, Void
  • point release linux: MX Linux, Ultramarine Linux, Zorin OS Core
the IoT versions of Windows, for those who just want to play retro games or do some light video editing, seem like a good option for anyone who isn't familiar with Linux.

for all those frustrated with Windows who want to switch platforms, knowing that a new system entails years of new experiences, and who do NOT want to dual-boot, I list here a few options I find interesting.

CachyOS is the highest-performing distribution at the moment. and that doesn't matter. it is friendly enough, and that matters.

EndeavourOS is a user-friendly, general-purpose distribution. one I actually prefer over CachyOS, as CachyOS swaps out certain programs by default in ways I don't really like.

PikaOS is a distribution aimed at gamers. and also friendly.

TuxedoOS is a distribution that I imagine could become the most important distribution for home use in just a few years. the distribution has now abandoned Ubuntu LTS as its base in order to migrate to “Continuous Debian” using the Rolling Testing model. friendly and up-to-date, a thing of beauty.

ArchBANG, I love the minimalism. very beautiful and lightweight, but not as user-friendly as the other options listed.

Solus is probably the most user-friendly of the alternative distributions listed. it is wonderful, though slightly more challenging because it is independent rather than being based on Arch, Debian, or Fedora.

MX Linux is amazing for really old computers, more than 15 or 20 years old. it’s a point-release distro, which bothers me, but it’s suitable for very modest computers.

Ultramarine Linux is Fedora done right. I have no idea how anyone has the nerve to recommend Fedora for home use when Ultramarine exists.

Zorin OS Core, like Linux Mint, is based on a distribution that, for me, is a thing of the past: Ubuntu. Ubuntu used to be good and important, but it has become the Windows of open source. intolerable, unpalatable, unbearable. invasive, authoritarian corporate junk. I didn't ditch Windows to use Fedora, openSUSE, SteamOS, or Ubuntu. Zorin and Mint still have their appeal, but they are based on the wrong distro.

TL; DR: I don't like a lot of text, long explanations, nothing like that.

on hardware that scores 1,500 in the PassMark single-thread test, it is possible to stream SNES games using the configuration below:

menu: settings, general, general
    ....check.........: open stats dialog on start up

menu: settings, video, streaming (advanced mode)
    # streaming settings
    audio encoder.....: libfdk AAC or FFmpeg AAC or any AAC
    video encoder.....: x264
    rescale output....: disabled
    (resolution)......: (irrelevant)

    # encoder settings
    rate control......: CBR (always CBR for live streaming)
    bitrate...........: 6,000 kbps (to the moon!)
    keyframe interval.: 2 s (always 2 s for live streaming)
    CPU preset........: veryfast  // or better as: fast or medium or slow
    profile...........: high
    tune..............: film
    x264 options......: bframe=0  // when using heavier presets on powerful CPUs, disable this option.

menu: settings, output, recording
    recording format.: matroska video (.mkv) // or FLV
    video encoder....: (use stream encoder)
    audio encoder....: (use stream encoder)
menu: settings, output, audio
    audio bitrate: 128 // or 160

menu: settings, output, replay buffer
    ..UNCHECK....: enable replay buffer

menu: settings, audio
    sample rate.: 48 kHz
    channels....: stereo

menu: settings, video
    base (canvas) resolution...: 1280x720 // or Native Monitor Resolution
    output (scaled) resolution.: 1280x720
    downscale filter...........: (none)   // or Area
    common fps values..........: 60       // or 30

menu: settings, advanced, video
    color format.: nv12 (8-bit, 4:2:0, 2 planes)
    color sapce..: rec. 709
    color range..: limited

menu: settings, advanced, stream delay
    ..UNCHECK....: enable

menu: settings, advanced, network
    ..UNCHECK....: dynamically change bitrate to manage congestion

menu: settings, advanced, sources
    ....CHECK....: enable browser source hardware acceleration

done!

sources and research content:


OBS Studio - Guides


salivity.github.io - How OBS Captures Exclusive Fullscreen Games on Windows


salivity.github.io - How OBS Studio Monitors WASAPI Device Handoffs


salivity.github.io - NvFBC Plugin for OBS Studio on Linux Explained


salivity.github.io - High DPC Latency and OBS Studio Micro-Stuttering


salivity.github.io - Why Use SRT Instead of RTMP in OBS Studio



manned - x264


Codec Wiki - x264


VideoLAN - Documentation:Modules/x264


MeGUI - x264 Settings


/dev/beandog - Encoding presets for x264

FFmpeg - H.264 Video Encoding Guide


FFmpeg - Advanced Audio Coding (AAC)


FFmpeg - Hardware Acceleration Intro


FFmpeg - Encoding for streaming sites


FFmpeg - Streaming Guide


FFmpeg - FFprobe Tips


 FFmpeg - Documentation


XMedia Recode - H.264 Encoder Settings | Tips and Techniques


Multimedia Media Mash - a lot of benchmarks



gstreamer - x264enc
NVIDIA - NVIDIA NVENC OBS Guide


YouTube - Choose live encoder settings, bitrates, and resolutions


YouTube - YouTube recommended upload encoding settings (for content recorded offline!)

Twitch - Broadcasting Guidelines


Twitch - Recommended Software for Broadcasting


Twitch - Getting Started on Twitch - Mobile Streaming

https://help.twitch.tv/s/article/getting-started-on-twitch

Twitch - 2k Streaming on Twitch


Twitch - Enhanced Broadcasting with Multiple Encodes

https://help.twitch.tv/s/article/enhanced-broadcasting


Kick - How to stream on KICK.com


Kick - Why 1080p may not be available on your channel


Kick - Settings and stream quality on the KICK - Go Live app



ReStream - Supported Social Platforms



I thought it would only take me 2 hours to write this text, but it took me more than 7 hours, even though I wrote everything from memory. =]

if the text helped you, that’s great. use it however you like. claim it as your own and resell it! feel completely free to do so, as you should always feel.


_o/

Comments

Popular posts from this blog

free and independent multistreaming

OBS Studio - dica rápida de performance e qualidade