 Hey guys, welcome back to my YouTube channel. This is Daniel Rosal here. So I've just gone ahead and upgraded to the Kaden live 21 decimals zero four decimal three. I'm not sure what I was running before but this is something I Put the PPA on to my Ubuntu 2010 machine This is running Alex DE as a desktop environment One of the reason I'm mentioning that is I think that is actually the cause of the bug or So says the wonderful people on reddit. I have a book here the shut down computer after renderings Button which would actually be very useful to me is a great out there, unfortunately So in any way in any event what I do this video for is because the options in rendering have changed Instead of saying GPU accelerated they've changed the description slightly to hardware accelerated Still marked as experimental, but I've now got a three profiles ABR VBR and ABR again For the H.264 and the H.265 codecs Respectively, so I've got nvenk H.264 ABR nvenk H.264 VBR nvenk H.265 ABR now it was previously CBR and VBR constant bit rate and variable bit rate now There's obviously there's still a variable bit rate Profile under H.264, but there's now average bit rate So I just plug these into Google because I have to be honest. I'm not sure exactly Do I want to make sure these were actually different and it's not just a different word for the same thing So from what I can gather that is the case. It says in the Wikipedia's average bit rate article at a given bit rate VBR is usually higher quality than ABR, which is higher quality than CBR So something's changed is all I can say now what I wanted to do this time if I can get back to my Caden life here. Nope. Ah, I can actually show this is the old Caden life Still have opening and they just close it down and this is the you can see slightly nicer So what I've done before is as before I've grabbed a random 10 second clip of test footage of of pixels I've got some free background music and I've added little this is a test rendering test overlay so you can see my beautiful composition here so it's Not in line to win any Oscar awards, but it's a 10 second clip now I'm not sure if comparing things over 10 seconds if it would be significantly different if I did a five minute test rendering project, but For the sake of quickness and simplicity now what I've done before is I'm going to just Ring this PC man FM window here This is where the files are going to build out in and what I also did is I opened up my N video Just so I can see when it's running. I saw this on another youtubers video. This is the graphics card. I'm running It's the g g e-force GTX 1050 Ti It's got four gigs of memory as you can see here And it's video engine utilization is currently at 13 and eight someone else on YouTube Made the same observation usually for me. These two are like almost zero I am recording this in OBS. So that would be my guess as to why we're seeing those numbers the way they are and That's basically it. So I thought I just kind of keep this here so that I can see When it's running if it's running. So firstly just as a benchmark I'm going to start off with the normal in inverted commas profile, which is going to be You know the mp4 h264 and AAC and I'm just going to go ahead and put this in a logical place on my file system test render and I'm going to go for rendered test render project mp4 and that should be building itself in this directory So let us watch as the magic unfolds. I'm going to go for more options I'm going to go as before for maximum on the encoding speed Quality is normal for project video audio as default. I'm not using the proxy clips to render And I'm not doing anything like rescaling it to pass is also great out and parallel processing So I'm using all the options that I can see that are going to speed this up. Let's go so we can see it started to build Actually getting a bit more options bit more information compared to this frame Telling you what frame it's on was not something in the previous KDN live. I was using so that's an innovation This is pretty quick. So let's see almost done And 15.5 megabytes 15.5 completed in 20 seconds. Let's just see the exact duration nine seconds And 20 let's just see here. Yeah Nine point. I'm not sure that's too far. That's the frame kind of thing, right? So nine seconds footage rendered in 20 seconds. So now Let us go in to the GPU accelerated. So let's start with nvenc h264 codec codec average bitrate. You can see the actual The melts commands being used here f equals mp4 v codec equals nvenc underscore h264 vb equals percent symbol bitrate plus k in quotation marks a codec audio codec equals aac a b equals Quotation marks audio bitrate plus k again in so again, I'm using the same thing here in coderspeed to max and let's go and render this and I Don't want to overwrite it. I'm gonna actually call it test render to test render project going to go underscore nvenc H264 a b iron is all separate files here rendering Now let's just quickly check the GPU utilization Where it's still at about 13% so ah, sorry 15% so it's just gone up a notch 20% now on video engine utilization and drop back dropping back. Yep. It's just about finished Seeing those numbers fall back now. So just confirming that it is using the GPU and 18 seconds so it shaved two seconds off the previous profile which was supposed to be CPU only rendering and Let's just take a look at what that file blew up to blew up to 27.3 So it's actually a smaller Overhead than I got previously on my previous set of benchmarks, which was something like two or three fold So let's try now to go for a 264 variable bit array. I'm going to just speed up this process a little bit here replace a br with vbr And let's go again product encoder speed max parallel processing rendering 2018 so the question is can we beat 18 seconds on the rendering time and can we beat? 27.3 as the file size We got 11.7. Let me just make sure I'm looking to write file here 11.7 is the file size Yeah, 11.7 is the file size so actually lighter than The original mp4 and 17 seconds. So that's the best performance so far Looks totally fine. Now. Let us just wrap this up by Going into the h265 codec average a bit array profile f equals mp4 file I put it is mp4 the video codec is hvec underscore nvenk Vb equals percentage bit rate plus k audio codec aac a b like this So this is h265 and parallel processing is turned on coding speed maximum Everything else on default quality is the same as it was on the previous ones rendering. Whoops. I want to just give this its own file page 265 codec and This is a br not vbr boom Rendering off to a flying start Checking checking utilization up to 25 Two seconds left and we should see this drop down if it follows working. Well finished Whoa, I dropped all the way down so you can see it's definitely using the GPU and again I'm using the GeForce GTX 1050 TI I am very very far from a graphics card expert. This is I actually bought this before I got into video editing I bought this because it had enough you can see I'm using three monitors. It actually has four. I was looking for a GPU with four outputs that could fit within my psus Pyrid delivery, so I don't know if this is the best graphics card in the world It wasn't that wasn't definitely not high-end graphics card. Hopefully a tiny bit above budget. It's been really great Runs my displays. I didn't know that people actually rendered video on GPUs So this is what I'm using but I can just looking at those not at this in action I it looks to me for sure that it's really using the GPU and Just get that out of the way that rendered in 16, which is our quickest so far, but in terms of the file size 27.4 so we got an overhead that was more like the original Go so h.264 vbr I would say given given that the time was much of a muchness. We only shaved off a small amount on each profile I would go for the The speed plus file of of h.264 vbr So I'm actually going to just do another one just to make sure that wasn't a fluke So let's go for h.264 vbr and let's call this file h.264 vbr underscore 2 Now if I wanted to do this more sophisticated I could you know do three passes on each on each render, but let's just see if we can replicate that result here looks like we're going to come in a bit similar so 11.7 again, and we also got 18 seconds Versus let's just make this big so there you can see all the details here So just to reiterate my results. So using the mp4 profile rendered in 20 seconds time results 18 seconds on 264 a br 17 seconds on h.264 vbr 16 seconds. So the quickest was on h.265 a br and again 18 seconds on h.264 vbr And in terms of the file sizes versus the rendering times, I'm just gonna put this here put this here in Terms of the rent the output sizes versus the rendering time So as you can see the best one was actually h.264 vbr because it actually Was a smaller file than the default mp4 profile and the time was 18 versus 20. Sorry the time was H.264 vbr 17 seconds and on the second run that came in as 18 seconds versus 20 seconds on mp4 So a bit quicker as well. Hope that video is useful. These are this is my little quick batch of test renderings on Kaden live 2104 decimal 3 using my graphics card