You still forgot to compare the quality. You can go for a much higher compression on CPU encoding and retrieve the same video quality compared to NVENC encoding. For example, my compressed H.264 copy of Shazam has a size of 2 GB in 1080P including 5.1 audio. And the quality is better than the 12 GB NVENC encoded file. So for me, GPU encoding is only an option if you can’t use CPU encoding for example for live streams…
x265 10bit slow CRF 26 still better than nvenc h265 10bit on CQ 23. And the Nvidia output filesize has more than double the size of the x265 output filesize.
Intel Quick Sync has come a long way since its introduction almost 10 years ago, and seems the most balanced between quality-time-filesize. For their Ice Lake CPUs it had more improvements in speed and it will get better in Tiger Lake/Rocket Lake, apparently.
My summation: if you have time, use CPU encoding. If you're looking to compress a 200+ episode series within days (versus weeks with CPU) use GPU encoding! There's no One Answer. Some people are purists (like me) and want to preserve as much detail versus someone who wants to just 'get in in there!' and move on. Thanks for the vid!
@Some Hairy Italian Guy do u think that ever will change or it will be a program that can do it with the GPU cause that is like outdated just able to use cpu only lots of programs can render using both at the same time not to compress but just render they can use both gpu and cpu at the same time
@Some Hairy Italian Guy but dont u feel like u using uneeded space 2 days ago I dropped a vid that was 100gb to 10gb 4k took 2hrs the best downsize I ever did big difference and my cpu temp maxes at high 80s which is high but drops down quick to 60s I wanna use the gpu one but makes no sense if my sizes will be the same or higher
Awesome! You deserve more views man. But 1 thing that I wish you could confirm is that if these final compressed files give out the same quality. Because when i used gpu encoding the file was super small but quality was potato so I have always sticked with h.264 cpu encoding.
Strange as when I used H.265 vs H.264 to encode a 500MB [4K > 1080p] video in Handbrake, the file was not only smaller using H.265 [20MB vs 33MB], the quality was also better.
Just a thought: it would be nice to have shown the compression ratio instead of the output file size. Source:Output would show in easy-to-read terms how well the specific encoder functioned.
I think the issue is the quality settings needs to be tuned to the encoder. If you look there isn't so much a correlation between filesize and processing time, but rather the algorihm and file size. Everything that was CPU only was about the same size. The two that were NVENC was about the same size, and then the VCE and QSV were also unique sizes (though with just once sample each). If you adjusted the CRF value of each I'd wager you could dial in an approximate file size and then compare all of them at the same file-size and see the relative qualities and times that they each came out to.
This was actually super helpful, thank you. If I need file sizes reduced I usually just drop the bitrate down a bit (thanks for that lesson, Adobe Premiere) so having all the data on hand makes decision making easier. :)
Totally agree but you haven't mentioned the heat factor? CPU encoding is not an option for me as my CPU (11700K) goes straight to 100°C and can easily get damaged! (as per what I've found online) I therefore use GPU encoding (H265 NvEnc on Handbrake) though the file size is much larger. Cheaper to buy another HDD than a new CPU! Someone please tell me if I'm missing something?
How's the quality difference between these? That seems like it would be hard to measure, but there are tools for it - see the VMAF extension for ffmpeg, which will give you a comparative difference between source and result. But I've also heard that CPU quality is far higher than GPU, but I have no idea where QSV lands...
Thank you. This confirms I have found with my tests too. CPU encode for smallest file sizes FTW. BTW my M1 Mac Mini is about 40-50% faster in terms of FPS in handbrake converting to HEVC with VFR compared to my Ryzen 3600.
Good info. The main reason why I use Nvidia NVENC encoding is because of speed and lower temps for my CPU. I have a I7-7700K which runs very hot even with an AIO. While actually encoding a video file it could get as high as 180+ F. Which is still within the max temp for the CPU, but too hot for my liking. When I started using NVENC encoding the temps dropped down to about 140 F when it is actually encoding the file.
@Mik e Good info, but in my case this is an apples to oranges comparison. The CPU I am using is known for having high temps. The only way to really solve the high temp problem would be to delid it. But since I can keep the high temps under control with the AIO I am using right now I don't want to take a chance of damaging my current CPU. Encoding with the CPU, an Nvidia GPU, and using the Nvidia NVENC option brought the temps down lower and even decreased the encoding times. Using NVENC with my current setup brought the temps down like 30 F.
Hello, I just recently go into his field and have been doing lots of testing of Handbrake to compress video for a plex service one day. My GTX 1060 wasn't up for the task, so I tried to get a new one. All the latest gaming ones were out of my reach, so I check custom computer builder websites and found out about Quadro cards. I purchased a Quadro P4000 and once I got it running, wow. Faster encoding and I could use the graphics option finally to encode video. It encodes faster and I can select a higher compression setting than even Handbrake recommends for encoding. I'm barely seeing any loss in quality. Ýou might want to give it a try with some commercial-grade cards. Just make sure you get one that's Pascal architect or later, I almost made that mistake. the original file was 1.2GB and with H.265 and the preset to default and everything else auto I got these results: quality 18 1GB, quality 20 837MB, quality 22 647MB, quality 24 486MB, quality 26 384MB, quality 30 246MB(at this setting I started noticing the quality of the video being watchable, but hair strands weren't as defined as before) It's also a single pass encoding which I'm guessing means something.
What about quality at a given setting? I've heard hardware encoding (NVENC) is lower quality than software (CPU) encoding. Any truth or significance to that? Cheers.
I came for the information, I stayed for the clear and easy to read charts END OF MEANINGFUL COMMENT STOP READING HERE. Seriously, I love the chart formats. And the info was helpful Too, but those charts are top notch. Okay now it seems I'm being sarcastic, but I assure you I am being earnest. Too many benchmark videos out there with slightly different hue/colored bar charts and small fonts and only on the screen for a second... Yeah, these charts are my fav. Charts charts charts charty charty charts.
Does it matter which CPU you are using when when you are using a dGPU for encoding? Does your GPU reaches same results when using a budget CPU vs. a high end CPU?
QSV rullz. That's why I bought 3770K in 2012 and it is still good. Even for encoding 4K video to 1080, I got up to 175 fps in FFMPEG 5.1. I hope 11gen is way faster which I consider to switch.
In the previous video to this one ,you where having trouble with selecting your perferred audio you wanted ,all you have to do is go to HandBreak main menu,next to the video is the audio tab,,select it.you'll see Audio Track,,right under that you'll 2 drop down boxes click on those and select your perferred language that helps.
strange, i use my I7 4790 CPU to compress video files in x265 and i never have above 1gig for DVD resolution DVD resolution is too low to go above 1gig of filesize, and even at this filesize i dont see any loss in quality
I saw that in both videos you converted to an mkv file before encoding and compressing. Is that a required step in the process? because I've been just encoding it straight off the dvd. I'm using a i7-3770 to do this and I tried x264 and quicksync and Quicksync has been longer for me. I don't know about the file sizes tho.
You got my like, since sometimes I do need to record screen, but want it fast and not too taxing (storage, cpu resources) . So Intel win qsv, and Ryzen win the core per price.
Everything has its place. Hardware encoding is great for low latency, like for encoding live feeds, and for testing stuff as rough drafts. I tried an NVENC HEVC encode once, matching the quality of an x265 encode. It was like 10x the file size. People love to argue about this kind of stuff, with no real knowledge of what they're talking about. For software encoding, the fewer threads the better for output quality. Your choice is to balance speed and quality. It's well known, for those who choose to listen to the programmers, that high parallelization reduces quality. If you only intend to do one video at a time, it would be better to buy a CPU with a low core count and high clock speeds. If you can take advatage of batch encoding, you can shoot for a high core count with lower clocks, and give each encoding instance a low thread count.
Very helpful 👍 I would be very grateful if one day you could do a video with Intel QSV using their new Iris Xe iGPU Will the Final File Size be smaller? Will it be faster? Many thanks 😊
I too would like to give that a whirl and see what kind of performance gains, if any, Intel has made with their Iris XE iGPU. Thanks for checking out the video and for the comment.
I've been doing my Blu-Ray collection with the following results. It ain't pretty, although the results are ABSOLUTELY PERFECT !!! So, here goes: On average over 150 movies, the average 1080p MKV off disc was a file size between 18.8 GB to 41.2 GB and took between 8 to 12 hours on an Intel 4400 Series Quad-Core AIO with integrated Intel HD graphics. I used ONLY the H.265 codec for better quality and smaller resultant file size. Most all files concluded around at 1/10th the original file size, and looked and sounded EXACTLY like the 1080p BD Movie ! PERFECT ! So, one movie is Handbraked over night, for a picture perfect watching experience at 1/10th the disc space needed. I kept all the .mkv files in case I want to re-encode to something else, but with VLC and H.265, it ain't really worth the extra trouble. AND, a few movies needed a bit of doctoring, for example "An American Werewolf In London" .mkv from BD was BAD, as in VERY grainy ! A little extra Handbrake FILTERS file settings tweek, and the finished file looked AWESOME cleaned up ! It took an extra 2 or 4 hours, but was worth it for the extra clarity !!! Now, if I could get my laptop to cooperate, a DELL AMD Quad-Core 9800, 16GB RAM, 512 SSD and 6GB video RAM, I could really crank this out...so I thought. DELL somehow closed out any real performance with this setup, and it takes 12-20 hours to do ONE movie from .mkv to H.265. I think I need to reset some of the Handbrake settings on it, but the laptop froze up two weeks ago trying to encode a movie. Here's a few .mkv to H.265 file size comparisons. The .MKV File size is straight off the Blu-Ray disc to the SSD, this file is used for Handbrake to encode to H.265: Movie Title - 3:10 To Yuma .MKV File Size - 27.741 GB Finished H.265 - 3.112 GB Movie Title - The Good, The Bad and The Ugly .MKV File Size - 36.943 GB Finished H.265 - 4.456 GB Movie Title - Iron Man 2 .MKV File Size - 34.969 GB Finished H.265 - 3.161 GB Movie Title - The Punisher .MKV File Size - 19.031 GB Finished H.265 - 2.556 GB Hope this helps as far as CPU encoding, at only 10-12 hours PER MOVIE !!!
@@Physics072 Yes...yes it is !!! I still want to know why my Dell laptop won't do better though ? I mean, 6gb of video RAM, and it was gonna take even longer ? Nope !
@@ntvtexan5015 I found doing a combo of gpu + cpu gives best size and quality but its not a preset type of thing but its much faster than CPU alone. I don't see a need for 4k at all. I have two 65" tvs a Plasma and a Newer 4k Samsung. At 1080P vs 4k at 12 feet on the couch ZERO difference. The Plasma crushes it on black scenes and Sports (600Mhz). The 1080p Plasma still the best over all. 4k really needs a 100+ inch screen unless you sit 4 feet from TV but then you get convergence eyestrain.
I want to see new Intel iris plus igpu quick sync encoding which come out in 10th gen ice lake mobile CPU I7 1056g7. The quick sync encoding time is very impressive and I would like to spend money on that feature rather than a high end CPU.
For the speed of encoding which brand do you think is the best Intel I9 or AMD R9, I currently have a Ryzen 5 5600H on my laptop and I would be interested in buying a mini pc with a powerful processor 32 GB Ram in DDR5 , would there be a real saving of time or not thank you.
I only use CPU encoding. You get a smaller file size which is the point of encoding. Saving time is worthless. The computer is doing the work, not you. So what time did you save?
The reason why I don't use GPU accelerated encoding is two fold. 1) the sub-optimal compression ratios, and 2) sub-optimal video quality. The H.264/265 codecs use lossy compression and the quality of the output will never be as good as the quality of the input using lossy compression. Some encoders do a better job of preserving the quality than others, and currently GPU accelerated encoding does not match CPU encoding for quality at the same bit-rate. So, you just have one more video to make comparing the video quality for the output of each encoder. th-cam.com/video/H0pCpNT4b-Q/w-d-xo.html
Just started using intel QS on a laptop i just picked up (i5-1035G4). I've been using CPU encoding (3700x @4.4GHz/3600MHz ram) and Intel QS is about 2-2.5x faster, 10-20% larger, and slight noticeable reduction in quality overall. Noticed in Handbrake the "constant quality" slider doesn't do much (if anything) and the Encoder Preset slider does a little bit. All this done on 720p files. Maybe it's a different result on 1080p, idk.. but for what I'm using it, it seems to be that CPU is still better but its FAR closer than it used to be. I also have a GTX 1050 2GB on my HTPC, i remember testing it out compared to the VCE on my main PC with a Vega 56. It was better but i can't remember.. probably because i still thought CPU was the best choice at the time. Great video, good info ;-)
You made a correct observation: When you compare CPU (AKA software) encoding vs GPU (hardware encoding), the picture quality is always better with the CPU @ the *same* bitrate. However, if speed is of the essence (like during streaming) the GPU encoder the better choice b/c you cannot have frame drops during streaming.
Would you mind doing a quick and to the point video on how to use Handbrake? I just started going through my video discs to rip them and would like some insight on: DVD Movie/Video (Is there a difference and does it affect how I set up Handbrake to encode?) DVD Standard vs Widescreen - I have at least one disk that is double-side, one side is standard and the other is widescreen. I'm not sure how to setup Handbrake for such instances. I have a DVD labeled Pan & Scan, I have no idea what that means with respect to format and if it requires any special settings for Handbrake Dolby Stereo/Dolby 5.1 (or any other types of audio settings you can think of) Blu-ray Again, I just want to be able to preserve as much of the original format (video and audio) as possible without selecting options that are overkill and have no affect on the result.
Oh boi i didn't know this... using my 3060 to encode and my 10900K with 630 is just laying low. Final size is the key for me. Now question is, which one output best video quality, CPU or GPU, was there any visible difference in term of grain and such ? Awesome video
interesting how the different CPUs end up with nearly identical file size! But why do the other encoders end up with different file sizes? Do they use all different presets? Is it possible to test them with identical presets? Maybe than we could decide which H.254 encoder is the fastest/best What I´m still confused about: the compressed file of the 5700XT is around double the size of e.g. the R5 3600. What does the extra size mean? Higher visual quality or just poor compression?
Im very new to this and I have a question. Im using a 3700x and 6700xt and I went to convert a h265 file to h264 using Handbrake. CPU load was around 100% and GPU was only around 8%. Is there a way to make the GPU work harder to speedup the process?
Ok, mate! Now you just have to perform the final task. 1. Video Codec -> h.265 (sw/cpu, NVENC, QSV, VCE) 2. instead of Constant Quality, set Average Bitrate; 3. Framerate -> Same as source; 4. Instead of Peak Framerate, set Constant Framerate; 5. Average Bitrate = 4800 kbps; 6. 2 pass ON for sw/cpu encoding with Turbo first pass; 2 pass OFF with NVENC, QSV, VCE encoding) 7. Tune -> None 8. Profile -> Auto 9 Level -> Auto In this way, the final files will all have the same size and you can compare their quality (especially during the fastest motion takings and the explosions).
Using the placebo settings, for best quality and smallest file size, a 44 minute show took 6:55:05, that's six hours, on cpu encoding. The same file took 0:2:12 using NVEnc on the slow setting. The cpu encoding made the file size, the original was 1.73 gb, a meager 730.78 mb while the gpu encoded file was 1.19 gb. Is a six hour and fifty four minute time save worth the added 460 mb? Depends entirely on what the end goal is but what I'm doing, detelecineing files for upscaling, I think I'd like to save nearly seven hours of wait time since upscaling itself will take between four and five hours. Or I could just stop being insane and not use the placebo speed...
This is why I don't bother with GPU encodes. I use CPU tweaked settings to give me the smallest file at the best quality. With 1080p screen recordings hours long I go from 5GB to 100MB with no perceptible quality loss. GPU encoding is a sales gimmick for most the part
GPU encoding is good for live streaming and maybe doing rough drafts of renders. I don't understand the taboo of letting a final render take possibly a whole day by using a software encoder to maximize quality versus the file size.
Unfortunately, this is all based on a Constant Quality setting of '20' based on the Handbrake 'HQ1080p30Surround' preset that you used. I hate to be 'that guy' but the test that you used is subjective based on platform used to Encode. Using either NVENC or pure CPU, I rip BluRays to my media server to a Constant BitRate of anywhere from 8,000kbps to 10,000kbps in h.265 and when I compare jobs, I use the same Bitrate Target. If you rip the same file, once in h.264 and once in h.265 at the same bitrate, the end file size will be the same. The quality of the h.265 file however will be 2x the quality, but may take a bit longer. Constant Quality of 20 means something different to you, me, Nvidia, AMD and Intel unfortunately. I can bet that NONE of your files are even close in final bitrates, especially when comparing the largest to the smallest. A constant Bitrate is translated the same on each platform for an equal comparison of both quality and time, as file size for each h.265 file should be within a percent or 2 of each other, same when comparing the corresponding h.264 files to each other. File size is a terrible comparison when using a subjective quality target preset, as I said, each platform will have it's own interpretation of what '20' means. I like your content though and it's too bad that my less-than serious tone doesn't come across in a txt post from a stranger who seems to be complaining about your work over the Internet with nothing else to add...LOL.
How is this even possible to do this kind of comparisons without any reference , other than the "preset" used ? You need to measure the quality of the encoded videos with SSIM, PSNR, or any standardized video quality tool. There's no way you can give any trust to a preset when different encoding methods are used, especially if gpu is used in some of them.
What are 480p videos? I just remember weird VGA back in time. I always wondered. For comparison: Here in Europe, DVDs were/are sold with true 720x576p quality. Where is 480p? Do you suffering on weak quality on old DVDs or what could be the reason?
Yeah, CPU encoding takes longer (I use a 5900x) and can even take hours if it's a 20GB file you recorded from a game.. But I like to keep the quality high, keep it 1440p and do it in 10-bit for HDR because I record with OBS with settings to capture HDR. But the 6700XT 12GB I use will do it in 20 minutes or less. But then I end up with an 8GB file instead of a 2GB h.265 file. And the quality is worse using a GPU vs CPU. Max compression and quality = CPU encoding.
And it's a fraction of parameters to look at. Weird you don't compare quality. What about audio? It's brave you dive into this topic but I bet it leaves you with more questions than answers. For my movie server it is important that the encoding used can be played back on all devices. Transcoding is not an option. I have succeeded finding a balance of file size, device support and ability to stream up to ten movies simultaneously from a single passive cooled small Linux box.
AFAIK, it's CPU encoding only, using 'very slow' preset, tweaked to taste. I would be surprised if they had a 'one size fits all' recipe. More probably, settings vary from movie to movie depending on format ratio, whether it's color or B&W, grainy or not, animation or not etc.
First you tested encoding speed only without checking file size or quality? And then you tested file size and speed but not quality? This is beyond an amateur mistake at this point. You should go read a beginner how to encode guide to learn the basics. You wasted a lot of time trying to learn very simple things.
Nvidia NVENC is in file size, quality and speed about the same as Faster settings with a 5900x of 12 cores and 24 hypers (h.265). The key is NOt to compare quality settings, but to have fixed bitrates (meaning same output size), then quality and speed are the only two factors to tweak. NVENC have only two modes, HQ and HP, both are almost the same speed and quality, but not quite. CPU however, have ultrafast to placebo, both quality and speed will vary with each step. Ultrafast with a AMD 5900x is actually faster than NVENC, however lower quality. Placebo better quality but a lot and do mean a lot slower. CPU gives you flexibility over NVENC, but the 5900x is my first CPU to match the speed and quality of NVENC, and those with less CPU power will still benefit from NVENC. NVENC also consumes less power, make less heat and noise over CPU processing. So even on my rig, I still use NVENC most of the times, I guess I would need a 5950x to truly overtake NVENC.
Cpu is better. Just leave it encode over night and you will get smaller file size and better quality. This is my opinion. And another thing encoding on any cpu will get the same quality the time will differ ( powerful cpu less time), if you use gpu the qualiti will difer from intel to nvidia to amd to the series of that gpu so you will have an unconstant quality movie collection
I've been converting my own DVDs and Blu-Rays to digital for some time now. I've used both QSV from a old Haswell processor computer and a Ryzen 3600. From what I've learned from this experience is a few things: 1. Your optical drive matters. Using a internal drive with 16x speed for DVD and 6x Speed for Blu-Ray will make a noticeable difference when compared to a external drive that can only do 8x DVD and 2x Blu-Ray. Even if you don't have space for a internal drive get a adapter to turn it into a external drive. It'll be the difference between waiting 30-60 minutes for the DVD to rip and encode, vs only waiting 8-16 minutes. (Note: Some DVDs and Blu-Rays take longer or shorter depending on the content and how data was put on the disc) 2. Size difference between QSV and CPU encoding is usually redundant. (I'll list my comparisons further down or in a reply) The size difference I usually see is as low as 50MB to as high as 200MB. Again seems like its dependent on the disc, but if all your doing is DVDs then the difference won't matter that much. I've converted over 200 DVDs ranging from movies to entire series and its yet to reach 1TB. Blu-Rays are a bit more tricky since they can range from 3GB up to 16GB after encoding and downscaling to 720p. 3. I see a lot of people mention quality control as a reason to prefer CPU encoding over QSV. CPU encoding can deliver better quality then QSV, but with DVDs I don't notice much of a difference myself, and sometimes regardless of which you use to encode the DVD or Blu-Ray it'll just encode badly. Nusicaa: Valley of the Wind is an example I've see personally. It looks fine when encoded, but you'll notice some weird graphical effects with in-movie text (Not subtitles) and weird blurring and colors in some scenes. Putting in the DVD reveals a much sharper and cleaner movie. I've reencoded the DVD multiple times and yet the issue persists so in conclusion. If your gonna maintain as much quality as possible and time isn't a issue for you then go ahead and use CPU, but if your like me and have to work 5 days a week and only get a few hours a day to work on ripping and encoding your collection QSV will get more done quicker at around the same visible quality. 3.1 Also a thing to note is that when you stream movies, even over a home internet connection, it is possible for the quality to naturally degrade when compared to moving it onto a USB or Hard Drive and watching it. Also dependent is the TV you have. Some TVs handle low-resolutions from DVDs a lot better then others. 4. In my experience leave the filters at default because they barely have any effect, that I've seen so far, and it would just be better to use a video editor that shows your edits in real time then waiting 8 minutes to see how sharpened or denoised the image is. Everything else is preference.
Not that I have been able to notice. The final output file size from CPU encoding is smaller though; so if you're wanting to fit as many movies as possible onto a home media server or something CPU encoding is the best... which is what I said in this video so I didn't need to say that... Sorry, I've made several videos dealing with CPU and GPU encoding and it gets difficult for me to remember what exactly I talked about in each video until I go back and look closely at which one you left your comment on. 😅 The TH-cam Studio shows me new comments made on videos and the video thumbnail it was made on but I don't always remember what I talked about in that video from the thumbnail alone. 😅
Hey dude, I have been having mic issues on all my games my game chat isnt working. I took it to best buy they said to delete and redownload games but i factory reset my computer and everything, but also I have these popping noises on my asus laptop, and I changed a setting for my speakers to be 16 bits 4800hz and It worked but the audio quality wasnt the same, I changed it back to the regular setting to see if that in all fixed my mic, but it didnt, I tried to change it back, but it isnt working anymore, plz help, TIA
That happens when you set the output quality too high, keep it at the default value and set encoding speed to slow or lower and the size will be smaller
if you want clossest quality to source file use the HEVC 10 bits color at slow preset at crf 22 lowest file sizes at great quality also use the Avg bitrate (kbps) at 6500kbps at 2 pass encoding you'll see how quality you'd get
I was compared the Intel QSV vs NVENC , QSV has the worst quality with noticeable detail loss compare to the NVENC. CPU still the best in term of the Size and Quality.
You still forgot to compare the quality. You can go for a much higher compression on CPU encoding and retrieve the same video quality compared to NVENC encoding. For example, my compressed H.264 copy of Shazam has a size of 2 GB in 1080P including 5.1 audio. And the quality is better than the 12 GB NVENC encoded file. So for me, GPU encoding is only an option if you can’t use CPU encoding for example for live streams…
Exactly! It's beyond moronic to just compare numbers without looking at the actual output! Unbelievably stupid by the OP!
x265 10bit slow CRF 26 still better than nvenc h265 10bit on CQ 23. And the Nvidia output filesize has more than double the size of the x265 output filesize.
If you think NVENC quality is bad, wait until you see how much worse AMD's crap is xDDDDD
Exactly. And the only serious way way to compare the quality is to force the encoding to a fixed average bitrate. He's a rookie! :-)))
Appreciate the video! I also appreciate that you can admit you made a mistake and that you went back and did more research!
Just wanted to let you know
Thanks for spending so much times in running these encoding tests. I truly appreciate it.
Intel Quick Sync has come a long way since its introduction almost 10 years ago, and seems the most balanced between quality-time-filesize. For their Ice Lake CPUs it had more improvements in speed and it will get better in Tiger Lake/Rocket Lake, apparently.
My summation: if you have time, use CPU encoding. If you're looking to compress a 200+ episode series within days (versus weeks with CPU) use GPU encoding! There's no One Answer. Some people are purists (like me) and want to preserve as much detail versus someone who wants to just 'get in in there!' and move on. Thanks for the vid!
Thank you very much! Really helpful benchmarks which I couldn't find anywhere else. Also, real world scenarios are always best as baseline.
CPU encoding is the best 🔥
which is weird
@Some Hairy Italian Guy but it's still weird how the gpu render is faster but bigger in size why is it not a way to have smaller size with the gpu
@Some Hairy Italian Guy do u think that ever will change or it will be a program that can do it with the GPU cause that is like outdated just able to use cpu only
lots of programs can render using both at the same time not to compress but just render they can use both gpu and cpu at the same time
@Some Hairy Italian Guy but dont u feel like u using uneeded space 2 days ago I dropped a vid that was 100gb to 10gb 4k took 2hrs the best downsize I ever did big difference
and my cpu temp maxes at high 80s which is high but drops down quick to 60s I wanna use the gpu one but makes no sense if my sizes will be the same or higher
@Some Hairy Italian Guy but dont u feel u using uneeded space
Awesome! You deserve more views man. But 1 thing that I wish you could confirm is that if these final compressed files give out the same quality. Because when i used gpu encoding the file was super small but quality was potato so I have always sticked with h.264 cpu encoding.
Strange as when I used H.265 vs H.264 to encode a 500MB [4K > 1080p] video in Handbrake, the file was not only smaller using H.265 [20MB vs 33MB], the quality was also better.
Just a thought: it would be nice to have shown the compression ratio instead of the output file size. Source:Output would show in easy-to-read terms how well the specific encoder functioned.
or pecentages
I think the issue is the quality settings needs to be tuned to the encoder. If you look there isn't so much a correlation between filesize and processing time, but rather the algorihm and file size. Everything that was CPU only was about the same size. The two that were NVENC was about the same size, and then the VCE and QSV were also unique sizes (though with just once sample each). If you adjusted the CRF value of each I'd wager you could dial in an approximate file size and then compare all of them at the same file-size and see the relative qualities and times that they each came out to.
This was actually super helpful, thank you. If I need file sizes reduced I usually just drop the bitrate down a bit (thanks for that lesson, Adobe Premiere) so having all the data on hand makes decision making easier. :)
Looking forward to see if RDNA2/VCN (big navi) has substantial encoding improvement over RDNA1
RDNA3 does and so does the older VCE
Realy great video. nice the first but with this you are a hero!
Totally agree but you haven't mentioned the heat factor? CPU encoding is not an option for me as my CPU (11700K) goes straight to 100°C and can easily get damaged! (as per what I've found online) I therefore use GPU encoding (H265 NvEnc on Handbrake) though the file size is much larger. Cheaper to buy another HDD than a new CPU! Someone please tell me if I'm missing something?
How's the quality difference between these? That seems like it would be hard to measure, but there are tools for it - see the VMAF extension for ffmpeg, which will give you a comparative difference between source and result. But I've also heard that CPU quality is far higher than GPU, but I have no idea where QSV lands...
great effort, is there an update on this.
Man thanks for this video, it really helped me.
Thank you. This confirms I have found with my tests too. CPU encode for smallest file sizes FTW. BTW my M1 Mac Mini is about 40-50% faster in terms of FPS in handbrake converting to HEVC with VFR compared to my Ryzen 3600.
Good video, but I am looking for lowest file size with best quality possible irrespective of time. I wonder whats VMAF rating for each.
@@Known2God yeah kind of, I tried using FFMetric to test n find out best quality to size ratio for my different presets .
Good info. The main reason why I use Nvidia NVENC encoding is because of speed and lower temps for my CPU. I have a I7-7700K which runs very hot even with an AIO. While actually encoding a video file it could get as high as 180+ F. Which is still within the max temp for the CPU, but too hot for my liking. When I started using NVENC encoding the temps dropped down to about 140 F when it is actually encoding the file.
@Mik e Good info, but in my case this is an apples to oranges comparison. The CPU I am using is known for having high temps. The only way to really solve the high temp problem would be to delid it. But since I can keep the high temps under control with the AIO I am using right now I don't want to take a chance of damaging my current CPU.
Encoding with the CPU, an Nvidia GPU, and using the Nvidia NVENC option brought the temps down lower and even decreased the encoding times. Using NVENC with my current setup brought the temps down like 30 F.
Compression ratio or percentages would be nice
Hello, I just recently go into his field and have been doing lots of testing of Handbrake to compress video for a plex service one day. My GTX 1060 wasn't up for the task, so I tried to get a new one. All the latest gaming ones were out of my reach, so I check custom computer builder websites and found out about Quadro cards. I purchased a Quadro P4000 and once I got it running, wow. Faster encoding and I could use the graphics option finally to encode video. It encodes faster and I can select a higher compression setting than even Handbrake recommends for encoding. I'm barely seeing any loss in quality. Ýou might want to give it a try with some commercial-grade cards. Just make sure you get one that's Pascal architect or later, I almost made that mistake.
the original file was 1.2GB and with H.265 and the preset to default and everything else auto I got these results:
quality 18 1GB, quality 20 837MB, quality 22 647MB, quality 24 486MB, quality 26 384MB, quality 30 246MB(at this setting I started noticing the quality of the video being watchable, but hair strands weren't as defined as before)
It's also a single pass encoding which I'm guessing means something.
What about quality at a given setting? I've heard hardware encoding (NVENC) is lower quality than software (CPU) encoding. Any truth or significance to that? Cheers.
Using the settings I did in this video I haven't noticed a difference at 1080p
@@Maraksot78 Cool. Thanks.
All these tests were about comparing different computers. .. .. I want to know about different handbrake settings compared
Can u follow up with av1?
I'd be like... "Hey Daniel, remember when i gave you that GPU? well I need you to do me a favor!"
I came for the information, I stayed for the clear and easy to read charts END OF MEANINGFUL COMMENT STOP READING HERE. Seriously, I love the chart formats. And the info was helpful Too, but those charts are top notch. Okay now it seems I'm being sarcastic, but I assure you I am being earnest. Too many benchmark videos out there with slightly different hue/colored bar charts and small fonts and only on the screen for a second... Yeah, these charts are my fav. Charts charts charts charty charty charts.
But hey! What about video quality?
Thank you!
Does it matter which CPU you are using when when you are using a dGPU for encoding? Does your GPU reaches same results when using a budget CPU vs. a high end CPU?
QSV rullz. That's why I bought 3770K in 2012 and it is still good. Even for encoding 4K video to 1080, I got up to 175 fps in FFMPEG 5.1. I hope 11gen is way faster which I consider to switch.
In the previous video to this one ,you where having trouble with selecting your perferred audio you wanted ,all you have to do is go to HandBreak main menu,next to the video is the audio tab,,select it.you'll see Audio Track,,right under that you'll 2 drop down boxes click on those and select your perferred language that helps.
An updated version with 3080 pr 3090 maybe, when they ll be available.
strange, i use my I7 4790 CPU to compress video files in x265 and i never have above 1gig for DVD resolution
DVD resolution is too low to go above 1gig of filesize, and even at this filesize i dont see any loss in quality
I saw that in both videos you converted to an mkv file before encoding and compressing. Is that a required step in the process? because I've been just encoding it straight off the dvd.
I'm using a i7-3770 to do this and I tried x264 and quicksync and Quicksync has been longer for me. I don't know about the file sizes tho.
You got my like, since sometimes I do need to record screen, but want it fast and not too taxing (storage, cpu resources) . So Intel win qsv, and Ryzen win the core per price.
Everything has its place.
Hardware encoding is great for low latency, like for encoding live feeds, and for testing stuff as rough drafts.
I tried an NVENC HEVC encode once, matching the quality of an x265 encode. It was like 10x the file size. People love to argue about this kind of stuff, with no real knowledge of what they're talking about.
For software encoding, the fewer threads the better for output quality. Your choice is to balance speed and quality. It's well known, for those who choose to listen to the programmers, that high parallelization reduces quality. If you only intend to do one video at a time, it would be better to buy a CPU with a low core count and high clock speeds. If you can take advatage of batch encoding, you can shoot for a high core count with lower clocks, and give each encoding instance a low thread count.
Very helpful 👍
I would be very grateful if one day you could do a video with Intel QSV using their new Iris Xe iGPU
Will the Final File Size be smaller?
Will it be faster?
Many thanks 😊
I too would like to give that a whirl and see what kind of performance gains, if any, Intel has made with their Iris XE iGPU. Thanks for checking out the video and for the comment.
@@Maraksot78
I thank you !!
I learned a lot from it :D
Now, do the decoding time! To compare for the remote play, which is best? AMD, Nvidia or intel?
Now go back and do the test again because you forgot to misure the quality of the encoding!
I've been doing my Blu-Ray collection with the following results. It ain't pretty, although the results are ABSOLUTELY PERFECT !!! So, here goes:
On average over 150 movies, the average 1080p MKV off disc was a file size between 18.8 GB to 41.2 GB and took between 8 to 12 hours on an Intel 4400 Series Quad-Core AIO with integrated Intel HD graphics. I used ONLY the H.265 codec for better quality and smaller resultant file size. Most all files concluded around at 1/10th the original file size, and looked and sounded EXACTLY like the 1080p BD Movie ! PERFECT !
So, one movie is Handbraked over night, for a picture perfect watching experience at 1/10th the disc space needed. I kept all the .mkv files in case I want to re-encode to something else, but with VLC and H.265, it ain't really worth the extra trouble. AND, a few movies needed a bit of doctoring, for example "An American Werewolf In London" .mkv from BD was BAD, as in VERY grainy ! A little extra Handbrake FILTERS file settings tweek, and the finished file looked AWESOME cleaned up ! It took an extra 2 or 4 hours, but was worth it for the extra clarity !!!
Now, if I could get my laptop to cooperate, a DELL AMD Quad-Core 9800, 16GB RAM, 512 SSD and 6GB video RAM, I could really crank this out...so I thought. DELL somehow closed out any real performance with this setup, and it takes 12-20 hours to do ONE movie from .mkv to H.265. I think I need to reset some of the Handbrake settings on it, but the laptop froze up two weeks ago trying to encode a movie. Here's a few .mkv to H.265 file size comparisons.
The .MKV File size is straight off the Blu-Ray disc to the SSD, this file is used for Handbrake to encode to H.265:
Movie Title - 3:10 To Yuma
.MKV File Size - 27.741 GB
Finished H.265 - 3.112 GB
Movie Title - The Good, The Bad and The Ugly
.MKV File Size - 36.943 GB
Finished H.265 - 4.456 GB
Movie Title - Iron Man 2
.MKV File Size - 34.969 GB
Finished H.265 - 3.161 GB
Movie Title - The Punisher
.MKV File Size - 19.031 GB
Finished H.265 - 2.556 GB
Hope this helps as far as CPU encoding, at only 10-12 hours PER MOVIE !!!
ONLY? LOL 10-12 hours is crazy slow.
@@Physics072 Yes...yes it is !!! I still want to know why my Dell laptop won't do better though ? I mean, 6gb of video RAM, and it was gonna take even longer ? Nope !
@@ntvtexan5015 I found doing a combo of gpu + cpu gives best size and quality but its not a preset type of thing but its much faster than CPU alone. I don't see a need for 4k at all. I have two 65" tvs a Plasma and a Newer 4k Samsung. At 1080P vs 4k at 12 feet on the couch ZERO difference. The Plasma crushes it on black scenes and Sports (600Mhz). The 1080p Plasma still the best over all. 4k really needs a 100+ inch screen unless you sit 4 feet from TV but then you get convergence eyestrain.
I want to see new Intel iris plus igpu quick sync encoding which come out in 10th gen ice lake mobile CPU
I7 1056g7.
The quick sync encoding time is very impressive and I would like to spend money on that feature rather than a high end CPU.
I don’t do as small as possible because the quality isn’t as good with a one with a higher file size cuz there is higher bitrate
For the speed of encoding which brand do you think is the best Intel I9 or AMD R9, I currently have a Ryzen 5 5600H on my laptop and I would be interested in buying a mini pc with a powerful processor 32 GB Ram in DDR5 , would there be a real saving of time or not thank you.
what was the output bitrate of the bluray with h265 cpu encoding?
I only use CPU encoding. You get a smaller file size which is the point of encoding. Saving time is worthless. The computer is doing the work, not you. So what time did you save?
And I want to see quick sync streaming quality on low end games like cskgo .
The reason why I don't use GPU accelerated encoding is two fold. 1) the sub-optimal compression ratios, and 2) sub-optimal video quality. The H.264/265 codecs use lossy compression and the quality of the output will never be as good as the quality of the input using lossy compression. Some encoders do a better job of preserving the quality than others, and currently GPU accelerated encoding does not match CPU encoding for quality at the same bit-rate.
So, you just have one more video to make comparing the video quality for the output of each encoder.
th-cam.com/video/H0pCpNT4b-Q/w-d-xo.html
Just started using intel QS on a laptop i just picked up (i5-1035G4). I've been using CPU encoding (3700x @4.4GHz/3600MHz ram) and Intel QS is about 2-2.5x faster, 10-20% larger, and slight noticeable reduction in quality overall. Noticed in Handbrake the "constant quality" slider doesn't do much (if anything) and the Encoder Preset slider does a little bit. All this done on 720p files. Maybe it's a different result on 1080p, idk.. but for what I'm using it, it seems to be that CPU is still better but its FAR closer than it used to be.
I also have a GTX 1050 2GB on my HTPC, i remember testing it out compared to the VCE on my main PC with a Vega 56. It was better but i can't remember.. probably because i still thought CPU was the best choice at the time.
Great video, good info ;-)
You made a correct observation: When you compare CPU (AKA software) encoding vs GPU (hardware encoding), the picture quality is always better with the CPU @ the *same* bitrate. However, if speed is of the essence (like during streaming) the GPU encoder the better choice b/c you cannot have frame drops during streaming.
you should show CPU usage too that makes a huge difference on using your PC to do other things while waiting for your encode to finish
Would you mind doing a quick and to the point video on how to use Handbrake? I just started going through my video discs to rip them and would like some insight on:
DVD Movie/Video (Is there a difference and does it affect how I set up Handbrake to encode?)
DVD Standard vs Widescreen - I have at least one disk that is double-side, one side is standard and the other is widescreen. I'm not sure how to setup Handbrake for such instances.
I have a DVD labeled Pan & Scan, I have no idea what that means with respect to format and if it requires any special settings for Handbrake
Dolby Stereo/Dolby 5.1 (or any other types of audio settings you can think of)
Blu-ray
Again, I just want to be able to preserve as much of the original format (video and audio) as possible without selecting options that are overkill and have no affect on the result.
Oh boi i didn't know this... using my 3060 to encode and my 10900K with 630 is just laying low. Final size is the key for me. Now question is, which one output best video quality, CPU or GPU, was there any visible difference in term of grain and such ? Awesome video
nice vid 😁
interesting how the different CPUs end up with nearly identical file size!
But why do the other encoders end up with different file sizes? Do they use all different presets? Is it possible to test them with identical presets?
Maybe than we could decide which H.254 encoder is the fastest/best
What I´m still confused about: the compressed file of the 5700XT is around double the size of e.g. the R5 3600. What does the extra size mean? Higher visual quality or just poor compression?
Mostly it's just poorer compression.
Dude all those file sizes are way too big. You need to make some adjustments to Handbrake settngs. What encoder CODEC you using?
Im very new to this and I have a question. Im using a 3700x and 6700xt and I went to convert a h265 file to h264 using Handbrake. CPU load was around 100% and GPU was only around 8%. Is there a way to make the GPU work harder to speedup the process?
Ok, mate! Now you just have to perform the final task.
1. Video Codec -> h.265 (sw/cpu, NVENC, QSV, VCE)
2. instead of Constant Quality, set Average Bitrate;
3. Framerate -> Same as source;
4. Instead of Peak Framerate, set Constant Framerate;
5. Average Bitrate = 4800 kbps;
6. 2 pass ON for sw/cpu encoding with Turbo first pass; 2 pass OFF with NVENC, QSV, VCE encoding)
7. Tune -> None
8. Profile -> Auto
9 Level -> Auto
In this way, the final files will all have the same size and you can compare their quality (especially during the fastest motion takings and the explosions).
Using the placebo settings, for best quality and smallest file size, a 44 minute show took 6:55:05, that's six hours, on cpu encoding. The same file took 0:2:12 using NVEnc on the slow setting. The cpu encoding made the file size, the original was 1.73 gb, a meager 730.78 mb while the gpu encoded file was 1.19 gb. Is a six hour and fifty four minute time save worth the added 460 mb? Depends entirely on what the end goal is but what I'm doing, detelecineing files for upscaling, I think I'd like to save nearly seven hours of wait time since upscaling itself will take between four and five hours.
Or I could just stop being insane and not use the placebo speed...
This is why I don't bother with GPU encodes. I use CPU tweaked settings to give me the smallest file at the best quality. With 1080p screen recordings hours long I go from 5GB to 100MB with no perceptible quality loss. GPU encoding is a sales gimmick for most the part
which settings u use
GPU encoding is good for live streaming and maybe doing rough drafts of renders. I don't understand the taboo of letting a final render take possibly a whole day by using a software encoder to maximize quality versus the file size.
Unfortunately, this is all based on a Constant Quality setting of '20' based on the Handbrake 'HQ1080p30Surround' preset that you used. I hate to be 'that guy' but the test that you used is subjective based on platform used to Encode. Using either NVENC or pure CPU, I rip BluRays to my media server to a Constant BitRate of anywhere from 8,000kbps to 10,000kbps in h.265 and when I compare jobs, I use the same Bitrate Target. If you rip the same file, once in h.264 and once in h.265 at the same bitrate, the end file size will be the same. The quality of the h.265 file however will be 2x the quality, but may take a bit longer. Constant Quality of 20 means something different to you, me, Nvidia, AMD and Intel unfortunately. I can bet that NONE of your files are even close in final bitrates, especially when comparing the largest to the smallest. A constant Bitrate is translated the same on each platform for an equal comparison of both quality and time, as file size for each h.265 file should be within a percent or 2 of each other, same when comparing the corresponding h.264 files to each other. File size is a terrible comparison when using a subjective quality target preset, as I said, each platform will have it's own interpretation of what '20' means. I like your content though and it's too bad that my less-than serious tone doesn't come across in a txt post from a stranger who seems to be complaining about your work over the Internet with nothing else to add...LOL.
How is this even possible to do this kind of comparisons without any reference , other than the "preset" used ? You need to measure the quality of the encoded videos with SSIM, PSNR, or any standardized video quality tool. There's no way you can give any trust to a preset when different encoding methods are used, especially if gpu is used in some of them.
What are 480p videos? I just remember weird VGA back in time. I always wondered. For comparison: Here in Europe, DVDs were/are sold with true 720x576p quality. Where is 480p? Do you suffering on weak quality on old DVDs or what could be the reason?
Ntsc video, aka in the west. Europe and PAL use a res of 576 whereas we're stuck with 480i.
Great video!!!!!
Yeah, CPU encoding takes longer (I use a 5900x) and can even take hours if it's a 20GB file you recorded from a game.. But I like to keep the quality high, keep it 1440p and do it in 10-bit for HDR because I record with OBS with settings to capture HDR.
But the 6700XT 12GB I use will do it in 20 minutes or less. But then I end up with an 8GB file instead of a 2GB h.265 file. And the quality is worse using a GPU vs CPU. Max compression and quality = CPU encoding.
And it's a fraction of parameters to look at. Weird you don't compare quality. What about audio?
It's brave you dive into this topic but I bet it leaves you with more questions than answers.
For my movie server it is important that the encoding used can be played back on all devices. Transcoding is not an option.
I have succeeded finding a balance of file size, device support and ability to stream up to ten movies simultaneously from a single passive cooled small Linux box.
Now I know to skip using my rtx2080 and use QSV instead. Thanks!
Handbrake presets are GARBAGE!
We should ask pirated movie encoder how they encode those movie. they produce very small files with amazing quality for FHD movie
AFAIK, it's CPU encoding only, using 'very slow' preset, tweaked to taste. I would be surprised if they had a 'one size fits all' recipe. More probably, settings vary from movie to movie depending on format ratio, whether it's color or B&W, grainy or not, animation or not etc.
Pirates using FFmpeg that you have total control of the compression 😅👌 More effective and more stable because of the raw code.
Damn it, I want to keep my cake and eat it. I'm gonna scweam and scweam until I'm thwick!
I ripped a h265 4k movie 6 gigs to 1080p h264 1.4 in windows on VideoProc Converter ai took 1 hour vs 4 for handbrake in Linux and better quality.
Intel Xeon's are pretty good at CPU encoding.
they the best not just good even in rendering in any other programs most people dont have them tho
apples M1 is king
First you tested encoding speed only without checking file size or quality? And then you tested file size and speed but not quality? This is beyond an amateur mistake at this point. You should go read a beginner how to encode guide to learn the basics. You wasted a lot of time trying to learn very simple things.
Nvidia NVENC is in file size, quality and speed about the same as Faster settings with a 5900x of 12 cores and 24 hypers (h.265).
The key is NOt to compare quality settings, but to have fixed bitrates (meaning same output size), then quality and speed are the only two factors to tweak.
NVENC have only two modes, HQ and HP, both are almost the same speed and quality, but not quite. CPU however, have ultrafast to placebo, both quality and speed will vary with each step. Ultrafast with a AMD 5900x is actually faster than NVENC, however lower quality. Placebo better quality but a lot and do mean a lot slower.
CPU gives you flexibility over NVENC, but the 5900x is my first CPU to match the speed and quality of NVENC, and those with less CPU power will still benefit from NVENC. NVENC also consumes less power, make less heat and noise over CPU processing.
So even on my rig, I still use NVENC most of the times, I guess I would need a 5950x to truly overtake NVENC.
Cpu is better. Just leave it encode over night and you will get smaller file size and better quality. This is my opinion. And another thing encoding on any cpu will get the same quality the time will differ ( powerful cpu less time), if you use gpu the qualiti will difer from intel to nvidia to amd to the series of that gpu so you will have an unconstant quality movie collection
I've been converting my own DVDs and Blu-Rays to digital for some time now. I've used both QSV from a old Haswell processor computer and a Ryzen 3600. From what I've learned from this experience is a few things:
1. Your optical drive matters. Using a internal drive with 16x speed for DVD and 6x Speed for Blu-Ray will make a noticeable difference when compared to a external drive that can only do 8x DVD and 2x Blu-Ray. Even if you don't have space for a internal drive get a adapter to turn it into a external drive. It'll be the difference between waiting 30-60 minutes for the DVD to rip and encode, vs only waiting 8-16 minutes. (Note: Some DVDs and Blu-Rays take longer or shorter depending on the content and how data was put on the disc)
2. Size difference between QSV and CPU encoding is usually redundant. (I'll list my comparisons further down or in a reply) The size difference I usually see is as low as 50MB to as high as 200MB. Again seems like its dependent on the disc, but if all your doing is DVDs then the difference won't matter that much. I've converted over 200 DVDs ranging from movies to entire series and its yet to reach 1TB. Blu-Rays are a bit more tricky since they can range from 3GB up to 16GB after encoding and downscaling to 720p.
3. I see a lot of people mention quality control as a reason to prefer CPU encoding over QSV. CPU encoding can deliver better quality then QSV, but with DVDs I don't notice much of a difference myself, and sometimes regardless of which you use to encode the DVD or Blu-Ray it'll just encode badly. Nusicaa: Valley of the Wind is an example I've see personally. It looks fine when encoded, but you'll notice some weird graphical effects with in-movie text (Not subtitles) and weird blurring and colors in some scenes. Putting in the DVD reveals a much sharper and cleaner movie. I've reencoded the DVD multiple times and yet the issue persists so in conclusion. If your gonna maintain as much quality as possible and time isn't a issue for you then go ahead and use CPU, but if your like me and have to work 5 days a week and only get a few hours a day to work on ripping and encoding your collection QSV will get more done quicker at around the same visible quality.
3.1 Also a thing to note is that when you stream movies, even over a home internet connection, it is possible for the quality to naturally degrade when compared to moving it onto a USB or Hard Drive and watching it. Also dependent is the TV you have. Some TVs handle low-resolutions from DVDs a lot better then others.
4. In my experience leave the filters at default because they barely have any effect, that I've seen so far, and it would just be better to use a video editor that shows your edits in real time then waiting 8 minutes to see how sharpened or denoised the image is. Everything else is preference.
Movies I converted comparison between QSV and CPU with H.264:
(Smaller file is CPU unless stated otherwise)
Thin Red Line - 3.49GB
shazam-sama we kneel
Is there a difference in perceived video quality tho between GPU vs cpu?
Not that I have been able to notice. The final output file size from CPU encoding is smaller though; so if you're wanting to fit as many movies as possible onto a home media server or something CPU encoding is the best... which is what I said in this video so I didn't need to say that... Sorry, I've made several videos dealing with CPU and GPU encoding and it gets difficult for me to remember what exactly I talked about in each video until I go back and look closely at which one you left your comment on. 😅
The TH-cam Studio shows me new comments made on videos and the video thumbnail it was made on but I don't always remember what I talked about in that video from the thumbnail alone. 😅
*Maraksot78* Oh Yeah !! Thats the real sound.please also check my list and review it..Thank you
Hey dude, I have been having mic issues on all my games my game chat isnt working. I took it to best buy they said to delete and redownload games but i factory reset my computer and everything, but also I have these popping noises on my asus laptop, and I changed a setting for my speakers to be 16 bits 4800hz and It worked but the audio quality wasnt the same, I changed it back to the regular setting to see if that in all fixed my mic, but it didnt, I tried to change it back, but it isnt working anymore, plz help, TIA
ripping dvd's? DId you also rip your cassette collection?
i did nvenc encoding of 650mb file and it turned it into 1.5gb file. it was fast but no use if it does the work opposite.
That happens when you set the output quality too high, keep it at the default value and set encoding speed to slow or lower and the size will be smaller
Just get an AMD Ryzen Threadripper 3990X.
Also, the hardware assisted encodes are always shittier quality... Which you forgot to show :P
Boy, I love you Bryan, but this is for egg heads or brainiacs only🤦🏽♀️❤️🙋🏽♀️love Aunt Linda
Thanks for watching it anyway Aunt Linda. Love you too. 😀
if you want clossest quality to source file use the HEVC 10 bits color at slow preset at crf 22 lowest file sizes at great quality also use the Avg bitrate (kbps) at 6500kbps at 2 pass encoding you'll see how quality you'd get
So basically cpu encode overnight... Ahahah
For best quality and smallest file sizes, yes. 🙂
1
I was compared the Intel QSV vs NVENC , QSV has the worst quality with noticeable detail loss compare to the NVENC.
CPU still the best in term of the Size and Quality.
Still no dear god
This video is a bit misleading
Do you have Facebook or discord because I got a gaming laptop with a Intel i7 and navida 3060 sorry if navida is spelled wrong.