TABLE OF CONTENTS
What makes the best CPU for video encoding?
In this article, I’m going to walk you through everything you need to know about CPU video encoding, including what to look for in a video encoding CPU and how other components may impact your encoding experience.
Let’s dive into it!
The Basics of CPU Video Encoding
What Makes Encoding Different From Rendering and Exporting
First, let’s take a moment to talk about what video encoding actually is since there’s a lot of confusion around the topic.
It doesn’t help that certain rendering and exporting processes actually do involve encoding in and of themselves, either, but for the moment I’m going to focus on encoding by itself.
Basically, video encoding is the process of taking a (sometimes raw, uncompressed) video file and converting it into another, more usable file format for your specific needs.
If the file in question is already encoded in another format, this process is called transcoding, but the essence remains the same.
For example, you may want to transcode an MKV file to an MP4 container file with an H.264 Codec or AVI container file with a WMV9 Codec that’s readily compatible with more devices.
Or, say you have raw footage that you want to use in a video project, but it isn’t compatible with your editing software.
Encoding that raw footage into a usable, compressed format is another common use of video encoding software.
Additionally, whenever you’re using video editing software, there’s a high chance that your final rendering and exporting process will itself involve some form of encoding or transcoding.
For instance, if you’re preparing a video for YouTube, you’ll most likely want to encode it in one of YouTube’s supported file formats.
Video Rendering vs. Exporting vs. Encoding vs. Saving – An overview
All of these mean very similar things, but there are some key differences that we should talk about.
Let’s start with the simplest one:
You can save a project file. This can be a Photoshop File, a Premiere Pro Project File, an After Effects File etc. The common denominator is that you are saving Application-specific settings in this file that usually can’t be opened with other software. Saving is generally a quick process that doesn’t need more than a few seconds.
When you save a Project, you package Application-specific settings into a file and save it onto your Storage. This does not normally involve any form of converting between formats. It can include packaging source material (such as footage in its original form) into that application-specific file, though.
When you export a Project, this typically designates a conversion into a non-(Project-)Application-native format takes place. This is to either make the resulting file compatible with other applications or to flatten and get rid of some of the information that isn’t needed for the file’s intended purpose (e.g. playback, viewing, streaming).
When you export a Project you flatten & strip it of Application-specific information (e.g. layers, effects settings) and/or add compatibility for viewing in other applications, players, image viewers etc.
Exporting a Video Editing Project includes the following two tasks, too: Rendering and Encoding.
Rendering is part of the export process in visually demanding applications and workloads. The act of rendering includes calculating a final image for each video frame by applying effects, transformations, color grading, stabilization etc. to your footage, layering the footage, adding graphics and text, or in more complex cases, ray-tracing physical light bounces to calculate 3D Objects and Environments.
These tasks are very demanding on your hardware and can take a long time to complete. The goal is to calculate any of the footage alterations you have set up in your project and create a final uncompressed image for each frame.
These final frames can then go on to the next step:
Encoding / Transcoding
Encoding is the act of taking the final uncompressed frames (after they’ve been rendered), encoding them, and packaging them into a container file format (such as MP4, MOV, AVI) that then consists of an encoded (H.264, AAC, etc.) video and audio stream.
So, what we’re talking about in this article, is which CPUs are great at encoding, not necessarily at rendering. Thankfully, though, these often go hand in hand:
Why Great Video Editing CPUs are also Great Video Encoding CPUs
As you may be able to surmise, video encoding and transcoding are pretty closely-tied to other video editing tasks.
You’ll often be doing a bit of all of them, especially if you’re working with uncompressed video footage on a regular basis.
Fortunately, all of these tasks share a key trait: they’re fairly CPU-bound, and scale fairly well to multiple CPU cores and threads.
This means that if a CPU is good for video editing, it’ll most likely be pretty good at video encoding and transcoding, too.
Even if you have a great graphics card to help accelerate these workloads, a strong CPU goes a long way in reducing total render times while keeping the final project looking high quality.
So, let’s talk about what makes a CPU appealing for video encoding and similarly CPU-intensive workloads.
What To Look For In a Video Encoding CPU
High Core Counts with a Modern CPU Architecture
A major boon in favor of video editing and other professional rendering software in comparison to something like a video game is that these applications are extremely scalable. In this context, scalable refers to the ability of an application to utilize more than one processing core.
Just a decade ago, CPUs with core counts of 4 or higher were considered a luxury in the consumer space, and for good reason.
A bit over a decade before then and a CPU didn’t even have “cores” to speak of, since a single processing core used to be the whole CPU.
We’ve come a long way since then, but you would be surprised at how many applications (especially games)…haven’t.
As it turns out, programming an application to utilize more than one CPU core can be pretty difficult, and over in that gaming space, effectively using more than 4 cores at a time is still a tough ask for many game engines.
Here in the video encoding, video editing, and productivity space, though, things are a lot better.
Since these workloads aren’t necessarily real-time in the same way that a game is, it’s easier for tasks to be broken down and divided across pretty much however many CPU cores and threads are available.
When you can put your whole system to work like that, you have a lot more incentive to invest in cutting-edge CPUs with high cores and thread counts, because double the cores can sometimes very well mean double the performance.
Simultaneous Multithreading (SMT) or Hyperthreading (HT)
On the note of cores, it’s also important to talk about SMT or Simultaneous Multithreading. Intel processors call this Hyperthreading in their marketing, but SMT is the proper name of the technology.
So, what is SMT?
Earlier, I mentioned that CPUs used to not have “cores” because a single core was considered a CPU in and of itself.
In the early days of multicore CPUs, multi-core CPUs were essentially considered to be two or more CPUs in one and basically were, since very few applications were made to utilize more than one core.
In this time, a “thread” also corresponded directly to your core count: think of a “thread” as a virtual representation of a core to your operating system.
But with Simultaneous Multithreading, the rule of 1 Core = 1 Thread can be broken.
In fact, with SMT enabled, the rule instead becomes 1 Core = 2 Threads. (Unless you’re a 12th Gen Intel CPU and you have both P-Cores and E-Cores, in which case only the P-Cores have 2 Threads per Core. More on that funny business in its section below.)
Now, SMT doesn’t suddenly multiply a CPU’s processing power by 2, as cool as that would be.
What it does do is enable it to multitask and manage scalable workloads really, really well. So if you’re looking to maximize the encoding or rendering power of a given CPU, getting an SMT-enabled CPU is a really good way of doing that.
High Clock Speeds with a Modern CPU Architecture
Clock speeds are a little more complicated because they’re fairly often misunderstood.
If you aren’t familiar with CPU technology, you might assume that something like 3 GHz is a straight-up measure of speed and that any given 4-core 3 GHz processor is going to perform about the same as another.
Due to the constantly changing nature of CPU architectures, especially across different generations or different brands, this is basically never the case.
High clock speeds are great, don’t get me wrong: but making sure those high clock speeds are being achieved on a modern CPU architecture is just as important.
Current-gen 3 GHz is a much different story from last-gen 3 GHz, not to mention three-gens-ago 3 GHz.
But generally speaking, 3 GHz and higher is good for a CPU, and many high-end CPUs being released today can boost up to 5 GHz without even needing to be overclocked. (If your Motherboard, PSU, and Cooling are adequate)
A Note on P-Cores and E-Cores
If you’re shopping for a modern Intel processor (and who knows, maybe AMD will start doing it, too), you may start seeing mentions of P-Cores and E-Cores.
Basically, these correspond to Performance Cores and Efficiency Cores.
Performance Cores are pretty much the same as CPU cores have been for generations now.
High-performance cores that can double up on their thread count with SMT enabled.
Efficiency Cores are a little bit different. They’re still powerful in their own right, but they’re tailored for lowering power consumption and heat rather than pushing the best raw power possible.
As a result, even with SMT enabled, these Efficiency Cores remain at 1 Core = 1 Thread.
If you’re interested in learning more detailed information about P-Cores, E-Cores, and how they perform, Jerry wrote an excellent guide on the topic.
Video Encoding Benchmarks
Before we dive into our CPU picks, let’s take a look at some Benchmark numbers that we’ll base our CPU recommendations on:
Techgage ran some Handbrake Video Encoding Benchmarks on both Threadripper and AMD Ryzen CPUs and we can already tell that there seems to be a sweet spot in terms of performance per dollar. The Threadripper 3960X, 3970X, and 3990X CPUs show heavily diminishing returns on adding more cores on the high-end, whereas the 16-Core 5950X seems to scale well thanks to its higher clock speeds.
Pugetbench compares the more recent 12th-gen Intel Alder Lake CPUs to AMDs 5xxx Series CPUs in this PugetBench Benchmark that runs inside Premiere Pro.
We see Intel’s high-clocking CPUs pull ahead considerably. A 10-core 12600k even outperforms the 12-Core 5900X. Do note that these scores involve not just encoding, but also exporting / processing any layers and edits within the project File – so we’ll take these results with a grain of salt.
Intel’s CPUs also like to draw a lot more power than their competitor’s CPUs, so this does not necessarily mean their technology is on par watt for watt.
Tom’s Hardware, too, tested 12th gen Alder Lake and AMD Ryzen 5xxx Series CPUs in Handbrake and LAME, which are pure Audio and Video Encoders.
LAME favors Intel’s CPUs, while Handbrake seems to perform better on AMD’s offerings. The overall scaling is what we would expect, though.
For now, though, let’s get to what you came for: what are the best video encoding CPUs you can get right now?
What Are The Best Video Encoding CPUs Today?
Note: When comparing pricing between Intel and AMD processors, be sure to also consider how much you’ll be spending on a compatible motherboard and RAM.
Best High-End: Intel Core i5-12600K
- Cores: 10 (6 Performance, 4 Efficiency)
- Threads: 16 (12 Performance, 4 Efficiency)
- Clock Speed: Up to 4.9 GHz Turbo
The best pick for the most people will be the recently-released Intel Core i5-12600K.
Boasting 6 SMT-backed Performance Cores and 4 Efficiency Cores capable of running at up to 4.9 GHz without an overclock, the Intel Core i5-12600K is no slouch in the performance department.
In CPUAgent’s Handbrake 1080p60 HEVC benchmark, the 12600K manages one of the highest scores, at a clean 263.
At the time of writing, it’s only beaten in that benchmark by a Ryzen 9 processor and the Intel Core i9-12900K. The 12900K with the top score is only at 275, which isn’t much more than the i5’s at all.
This is a processor that is punching ridiculously far out of its weight class.
By comparison, the last-gen Intel Core i5-11600K only manages a meager 168 in the same benchmark.
Add on the fact that it’s cheaper than every other pick on this list despite performing within spitting distance of all of them, and the Core i5-12600K becomes a pretty much impossible-to-beat value in the current market conditions.
It laughs in the face of last-gen Core i7 and Ryzen 7 processors by outperforming them at a lower price.
If It’s Cheaper Than Best High-End: Ryzen 7 5800X
- Cores: 8 Performance
- Threads: 16 Performance
- Clock Speed: Up to 4.7 GHz Boost
Prior to the 12th Gen Intel CPUs storming the market in all of their great terror, the high-end Ryzen 5000 CPUs seemed pretty much unbeatable.
And truthfully, their performance is still fairly high up there in workloads related to video encoding and transcoding.
A glance at Anandtech’s benchmarks shows a total bloodbath in favor of these processors when pit against Intel’s 10th Gen, and Intel’s 11th Gen performed pretty much the same, if not worse in a few cases.
I would still recommend the Core i5-12600K first, because the processor is actually that good, but if you can find the 5800X for cheaper then it’s definitely an alternative worth considering.
No Compromises: Intel Core i9-12900K
- Cores: 16 (8 Performance, 8 Efficiency)
- Threads: 24 (16 Performance, 8 Efficiency)
- Clock Speed: Up to 5.2 GHz Turbo
If all of this talk of Intel’s 12th Generation superiority in video encoding sounds good to you, the Intel Core i9-12900K is that but more.
However, you’ll be spending a lot more money for the privilege, and you won’t always get an improvement that directly scales to all of the new cores that you got for the upgrade.
This is the No Compromises pick for a reason, though- if you want the fastest video encoding CPU currently on the market and don’t care how much you gotta spend on it, this is the pick for you.
If It’s Cheaper Than No Compromises: AMD Ryzen 9 5950X
- Cores: 16
- Threads: 32
- Clock Speed: Up to 4.9 GHz Boost
A Ryzen 9 5950X won’t necessarily boast the same raw encoding performance as a 12th Gen Intel processor, but it’ll come pretty freakin’ close.
And in other core-dependent high-end workloads, the 5950X remains one of the leading choices in the entire industry from a performance perspective.
Where current-gen AMD CPUs are suffering in comparison to their Intel rivals is in pricing, because a major bump in performance has happened on Intel without as major a bump in price.
It’s been a long time since the CPU market has been this competitive.
If you can find this CPU cheaper than the Intel Core i9-12900K, it’s definitely a worthwhile alternative to consider for your heavy-duty workloads.
What about Threadripper?
Because Encoding needs high clocking CPUs, ample amount of cores, and fast cache access, while Threadripper CPUs do come out on top in many benchmarks, they are considerably more expensive than mainstream CPU offerings that don’t perform that much worse.
If all you need is the highest performance for encoding videos, no matter the cost, and you’re also not planning to do any kind of Editing work or run related workloads, a Threadripper could be an interesting choice.
Because most professionals that do encoding, though, also want to be able to edit smoothly or work in the applications on that same workstation, choosing a mainstream CPU with higher single-core performance is the better choice, even if it has less cores.
What else should I be mindful of for a Video Encoding PC Build?
If you’re going to be doing a lot of video encoding, editing, rendering, and the like, you need to make some concessions with your PC build besides just buying the strongest CPU and GPU within your budget.
Namely, your PC build is most likely going to be running at pretty high temperatures if you don’t invest in plenty of case fans and a good cooler to keep everything running optimally.
Applications like Handbrake and your usual video editing software can very quickly use up your entire CPU if you tell them to or don’t remember to tell them not to.
Running a CPU at full-throttle for a sustained workload is a surefire way to introduce yourself to some thermal throttling if you aren’t compensating with a properly-cooled and ventilated PC build.
Does RAM impact Video Encoding performance?
Like with most RAM-related things, if you’re running in Dual Channel with at least 2 RAM sticks you’re most likely doing fine.
Generally speaking, RAM capacity is going to have the biggest impact (compared to other RAM specs) on how well your workload runs.
Since the purpose of RAM is to hold whatever the CPU is actively working on, you should target having enough RAM to manage your project files without needing to offload them to the much slower paging file on your hard drive or SSD.
Past RAM capacity, increasing RAM speed, and lowering RAM latency are other good ways of improving RAM-related performance.
You generally shouldn’t need to worry too much about this, though, just remember to enable XMP.
Does the GPU impact Video Encoding performance?
Yes, as long as the video encoding or video editing application in question supports GPU acceleration.
The majority of them do, and I particularly recommend Handbrake for anyone looking for a GPU-accelerated video encoding application.
Handbrake does take a little know-how to use effectively, but once you figure it out, it’s free and provides superb encoding performance that makes the most of your hardware.
Over to You
We’ve reached the end of the article.
I hope you have a better understanding of CPU performance and how it relates to video encoding now.
Even if you don’t end up going with one of the recommendations listed above, it’s my sincere hope that this article armed you with the knowledge you need to make an informed buying decision whenever you may need to in the future.
Whenever that future comes, or until then, feel free to leave a comment below or on our forum with any CPU or video encoding-related questions you might have.
30 July, 2022
A huge deal is hardware acceleration of h.265 codecs in 10bit 422. The M1’s from Apple have it, and the Intel 11th and 12th gen CPU’s have it. So many modern cameras use this codec, and without that hardware acceleration editing is a nightmare. I hope the new AMD cpus wil have this built in as well. It makes a monster difference.
31 July, 2022
Thanks for the feedback! I’ll be doing some more fleshed out research of my own, but I’ll definitely be updating the article to reflect this information in the coming days.
11 June, 2022
nice article =)
what would u say about streaem performance? many streamers still see encoding with x264 the best way and the “slow” preset is somewhat of a dream to achieve
11 June, 2022
My recommendations should still apply for an x264 stream encode! Depending on the pricing of the AMD chips, those tend to do extremely well for streaming workloads.