This Emulates [Almost] Everything! - Retroid Pocket 4 Pro Deep Dive

how's it going everyone Taki here we have another  retroid pocket 4 video for you today this is going   to be a deep dive into the emulation performance  in the new retroid pocket 4 Pro now this video is   going to be predominantly about emulation but  there are some important points that I didn't   get to talk about in my last video that we're  going to talk about in here and we also have   some updates on some issues that I brought up  in my first look that we'll also address in   this video as a quick recap the retroid pocket 4  Pro comes with the dominity 1100 processor and I   had a few questions on this topic in my first look  video I said this is an 8 core processor and a 431   configuration with three of the a78 cores clocked  at 2.4 GHz if we go by retroid spec page they list   that all of the a78 cores should go to 2.6 GHz  and we can see the same thing on mtk website I   don't know if this is a software issue or if this  is a poorly bin processor that can't reach all of   the max clocks at the same time but the highest  performance profile that we can use on the rp4 pro   will not allow us to lock all of the a78 cores to  2.6 GHz we are left with three running at 2.4 GHz   and a the final one running at the rated 2.6  GHz it's kind of strange and I've never seen   anything like this before but we need to use  that high performance profile for most of what   this can do on the high-end so we're stuck with a  431 configuration with slower clock speeds I hope   that clarifies things in this video we only have  the rp4 pro but a lot of people are wondering how   this compares to the cheaper rp4 model with the  dominity 900 I don't have that model but I do have   another device with that chip that we can look at  for this section of the video it probably won't   be obvious to you because this is not a color  that was officially sold but this is the Odin   light and it comes with the same processor that  comes in the rp4 base model the biggest difference   being that this was a $199 that was discounted  to around $179 if I'm not mistaken whereas the   rp4 base version comes in at $150 now the gaming  experience is completely different between the two   of these and the processor cost has come down  since this originally came out but this was a   very good deal before this Hardware came out and  real compy competition that this had from a price   versus performance standpoint was probably the  pimax portal retro Edition when it went on sale at   $199 but anyway we're going to do some benchmarks  on both of these and we're going to talk about   some of the performance differences that we can  expect to see or we would expect to see between   the dimensity 1100 and the dominity 900 when it  comes to emulation to get a better idea of what's   going on here let's take a look at some benchmarks  between both of these chips before we switching   over to some real world performance tests because  this won't tell you the full story first up we   have a CPU Benchmark and you'll see that there's a  big difference between the single core scores and   the multi-core scores the difference between the  multi-core score is easy to explain because this   chip has more performance cores than this one  does this one has two performance cores and six   efficiency cores but this has four and four but if  we look over at the single core performance we can   see that that's 1126 on the d1100 versus around  900 on the d900 that difference is largely due to   the fact that this has a performance core that can  clock up to 2.6 GHz whereas as this one can only   go to 2.4 when it comes to emulation we care more  about the single core performance because a lot   of the emulators that are out there can only run  on one or two cores so you won't really see a big   difference between both of these processors in CPU  bound emulation this one will have an easier time   playing high-end PS2 games but this one won't be  that far behind but when we go over to the Vulcan   score there's a huge gap here and if you just  looked on the numbers here you'd say okay I can   spend $50 more and get the d1100 and I'm getting  more more than double the GPU performance and yeah   if you just go buy this that's a good tradeoff for  Vulcan we're at 4716 on the d1100 versus 2189 on   the d900 when it comes to open CL we also have  a huge gap here with 4680 on the d1100 versus   2061 on the d900 the problem is this doesn't take  into account a lot of the specific things that are   required for emulation this score doesn't tell  you anything about the compatibility of the GPU   have and whether or not it has the things that are  required from the emulators that you want to run   but more importantly it doesn't scale like this  when you're actually emulating something where you   will see this performance Gap the most is when you  are going to increase the rendering resolution but   if you're capping both of these at 720p or 1080p  then it won't be that easy to notice a difference   all right so now let's look at some real world  performance differences between both of these   processors and just keep in mind this is a $199  device and this is $199 device but the hardware   that's here is sold as an rp4 model for $150 let's  start with the God of War games for PlayStation 2   since those are a decent CPU Benchmark I have  the same Save State on both these devices so   we can get an easy apples apples comparison  on the top we have the rp4 pro using the high   performance profile if we swipe down from here you  can see that it's on high performance I have the   brightness set to Max and we're using the smart  fan and on the Odin light if we swipe down from   there you can see it's on high performance Max  brightness and the smart fan and at Max brightness   the brightness level between both of these is  about the same so that shouldn't be big difference   but let's take a look at what we have here first  on the top you can see that we do in fact have a   43 1 configuration you can see that CPU 7 is the  only one that goes to 2.6 GHz we don't really   need maxed out clocks for something like this but  this is the best way that we can get an Apples to   Apples comparison now everything is equal and we  can see a couple of interesting things here first   if you take a look at the power consumption there  are times when the rp4 pro can get really high up   in the power consumption where this current right  here will get close to 2,000 for example right   there was 2.2 amps which is a lot of power to use  on this chip but the d900 and the Odin light also   can get pretty high right here it's at 1.7 amps or  1.8 amps it fluctuates a bit but it does use a lot   of power for something that's this demanding also  point out that we're not using any underclocking   settings on either one of these this is just  purely 0 for this Benchmark that we're trying to   do but as you can see it's running on the bottom  device with no problems in this situation this   is not going to be Universal across the entire  game but in this area right here it's fine and   obviously the same thing can be said for the  rp4 4 Pro now let's increase the resolution to   2x native to see what that does with these two  chips and see if we can then find where the Gap   lies now we're at 2x native on both of these that  bump in resolution is a lot more apparent on the   Odin light than it is on the rp4 pro but both of  them still look good you'll see now on the Odin   light that we are not hitting a solid 60 FPS in  this setting right here it will dip a bit below   and the fans are also increasing because now we  have a higher load especially on the GPU on the   rp4 pro we're at a solid 60 FPS in this area  running the same track so you can obviously   see that there is more performance to be had with  this processor obviously because we do have that   higher clock but we have more performance cores  now here's where things get interesting so this   is a 00 versus a 0000 comparison where we're not  using any underclock settings but we could use   underclock settings on the d900 to get about the  same experience that we're getting on the rp4 pro   we can set the E cycle rate to minus one and the  E cycle skip rate to two and that should give us   about about the same experience that we can get on  the rp4 pro if you go by this OSD reading on the   top here you can see that we're no longer hitting  a Max of 60fps there are times when it'll go under   but that is 100% of the speed that's needed for  this underclock setting and the game is fully   playable it's just not as smooth as it is on the  rp4 pro with a z00 setting let's check another   game for this we have God of War 2 and you'll see  that we don't have that e cycle skip rate turned   on right now but we are not able to run this  game at 60 FPS with these settings on the d900   we're only running at 42 FPS and it is much slower  than it is on the D100 the D100 is not having any   problems here but it is using a ton of power you  may have just seen that went up to 3 amps which is   an insane amount of power to use on a device that  is this small the d900 in the Odin light cannot   run this game with a z00 setting but it also  does not use anywhere near as much power as the   rp4 pro does and if we just go into those system  settings and set this back to minus1 + 2 yeah we   can play this game it's just not not as smooth  as it is on the rp4 pro so it's not like you're   going from completely unplayable to playable with  a $50 increase in price you're just able to play   a game that is demanding like this without having  to use any of these settings that will lower your   FPS Target while still running at 100% but you're  also using way more power to do so for our next   test we have a GameCube game this is Metroid Prime  2 it's one of the most difficult GameCube games to   run and it needs a lot of CPU power to run well  with the official Dolphin Emulator and the same   settings on both of these we can Run the game at  60 FPS on both of these in the area that I'm in   and the experience is about the same on both of  them but you will see there are times when the   FPS will go a bit under 60 on the rp4 pro the  same thing will happen on the Oden light but it   goes down a bit further if you wanted to run this  game on the d900 you'd probably be better off with   something like MMJ R1 or mm jr2 that would get  you better performance in this but you do have   to work around some annoying hacks in terms of  power consumption both of these are pretty close   together and our final game for this comparison  is a Wii title this is Super Mario Galaxy 2 it's   another very demanding game but for the Wii system  and you can see that it's running on both of these   but the FPS is higher on the rp4 pro so again  this is a situation where the game can run on   both devices but for the d900 if you want to play  it at 2x native resolution then You' be better off   using the mmgr or the mmgr 2 build we talked  briefly about power consumption in that last   section and I also talked about it in my first  look but let's talk about it on the extreme end   and what is going into that power consumption and  if you look down over here on the left side you   can see our wattage now this total package here  can use at most around 12 watts of power which is   really nuts for something that's this small there  are a few things that are going into this both on   the software side and on the hardware side on  the software side we have that fan and we can   see how much that impacts Us by just going over to  a static screen so let's go over to a 64 all right   here we are in Ada 64 I have this on the high  performance mode with the smart fan let's just put   this on turbo or sports mode and you can see when  we're idling with the clocks all maxed out with   nothing running we're using around four watts of  power which is a lot to use if we just go to this   discharge rate here we can see that we're idling  at just under 1 amp which is a lot of power to use   for something like this but let's further isolate  this let's now go over to the standard setting   we're going to set that on standard and then we're  going to turn the fan on and we're going to turn   the fan on sport so it's using the most power for  the fan with the lowest CPU and GPU profile you   can see that there's not a lot that's changed here  the CPU clocks are still spinning up a bit higher   than they would normally need to be in a situation  like this but we are still using around 1 amp of   power and this is with the fan on Max let's see  what happens when we turn the fan off completely   and everything else stays the same so we just  disabled the fan we're now using about half the   power that we were before we're idling at 2 Watts  which is way better than 4 Watts So based on this   we know that the fan is adding about 2 watts of  total power to this system when you're on the high   end emulating very demanding things that need the  fan running at full speed so that's one thing and   there's not a whole lot that we can do when we're  using that high performance profile where we need   to use the fan on the max setting or the smart  setting because the CPU attemps can get pretty   high in this enclosure but there are situations  where the fan speed is ramping up too high when   it doesn't need to be and in those cases we're  probably using about one watt of additional power   just to have the fan spining up more than it needs  to be the next Hardware thing that we have to deal   with is the screen this screen uses a ton of power  if we have this thing running at Max brightness   like it is right now that is going to add a lot  to our power consumption for example if I change   the brightness to the minimum setting which is  more than you'll need if you want to play this at   night in bed we're using one watt of power and if  you're going to be playing something that doesn't   require a lot of power then your battery life is  going to go a lot further the final thing is both   software and Hardware on the hardware side this  CPU can use a ton of power when you push it all   the way you do get something for that power but  I don't think the juice is worth the squeeze on   the software side the CPU and GPU profiles that we  have on this are not not really tuned adequately   for this kind of solution in this section we're  going to talk more about the screen on the rp4   and the rp4 pro now I read a lot of comments from  people about that footage some people that just   had like normal thoughts about the screen looking  bad and how they hoped that the company would come   in and fix it and then there was a separate group  inside that group that was angry that I was even   talking about this situation to begin with OR  that they were angry that I didn't talk about   that the company was going to fix it which was  something that was being talked about at that   time if a problem exists and it exists while I'm  reviewing it then that's what I report on I don't   do wishful thinking stuff that it's probably  going to be fixed because a lot of people can   have different definitions of what a fix would be  so what I did is I just said here's the problem   and here's how it will be fixed and here are the  potential consequences of the fix typically when   I'm doing a review video for something like this I  don't really have much conversation with a company   that's making the device unless I run into a  problem that I think is significant enough to   bring to their attention the green screen issue is  an example of that but this also extends to other   products if I'm not not mistaken I believe there  was like an RG 351 series device that had like   a d-pad issue that was not registering inputs  correctly where there was another one that had   Wi-Fi interference with the onboard Wi-Fi and  the speaker so they just like removed it before   release then there's the whole RG 552 issue with  the audio delay that it didn't end up getting   fixed before they pushed the release of that  device but some more recent examples would be the   touchscreen issue in the win 600 or the line issue  that I pointed out in my RG 35xx plus video I pass   on these kinds of issues to the manufacturer if  the device is a review unit or if I bought it with   my own money if I think the issue is significant  enough so I was working on my rp4 first look video   there were a lot of other people that were doing  First Impressions looks on it and a lot of them   pointed out that their screens were green when  that feedback got sent back to the company they   thought that that problem was due to the wallpaper  and how Android just colors the UI elements based   on the colors that are in the wallpaper which is a  fair assessment to make because that is something   that happens and that's why I pointed it out in  my first look video while that was going on I was   filming the last part of my first First Look video  at the beginning because the green screen was so   shocking and I wanted to pass that information on  to the company what I was trying to do was to show   them that it was a hardware issue with the footage  that I had and this essentially just looked like   showing them some clips from that last section of  the video that I made while I was filming it to   give them a full overview of what was happening  let's go over some of the ways that this issue   could have been caused both from the software side  and the hardware side before we ultimately get   to the point where we know that it's a hardware  problem and that it's not something that you need   to be like so worried about so from the software  side there's really only two ways that this issue   could have been caused and one of them I already  detailed with the wallpaper if we change the   wallpaper to White and the screen is still green  and all the UI elements are still green then there   must be another reason for that to be the case and  that leaves you with your second option which is   that the screen is misconfigured and a lot of  the messaging around this time was that there   was some saturation issue that needed to be fixed  or the screen wasn't properly configured which is   not technically correct but if we just accept that  last Point as true that the screen is just misc   configured then we can find out that that is not  the case by just doing a simple test and just so   we're on the same page when we're talking about  a screen being configured Inc correctly from an   Android standpoint we're talking about something  like this where the user has full control over   this like menu here that they can customize the  color temperature to be whatever they want so for   example I could make the screen have a pink tint  to it or I could make it blue if I really wanted   to ruin the screen but that's what we mean we  don't have this kind of UI on retroid but it is   possible that this UI does exist hidden and it  was configured wrong if you're in a situation   where you have Hardware that was configured  incorrectly in software then you can usually   find out if that's the case by looking at how  the screen looks before Android is fully booted   the good thing for us is we have a boot screen on  all of these devices that has a logo in the middle   and we can use that to judge the white balance of  the screen so what I did for the rp4 pro was boot   up the device and then look at the boot screen  before the Android system is fully initialized   and I saw that that was green and the only way  that that would be green is if the screen was   actually native green by Hardware so this is what  I mean right now Android has not started up yet it   is in the process of booting up so right now what  you are seeing is how the screen is configured by   default from a hardware level and this logo here  is white it was just green there but you get the   point so after I had that I went on to film the  rest of the video because I knew basically there   was only two ways that they could fix that issue  they can either buy a new screen that doesn't have   the same LED backlight that's green or they can  try to fix it in software replacing it outright   is not realistic for any company of this size so  the only option is really for software and I knew   that that would turn out in the same way that the  Win 4 turned out because it works in the same way   so I finished my first look video by showing  what would likely happen to this screen once   it's properly configured in software to correct  the hardware problem so let's boot up the device   now and the first thing that you'll notice  is this screen is still green because right   now Android hasn't started up but you will see  when Android starts to start up that this will   change because the software fix is being applied  it's very obvious see it just changed that is the   software fix and if we change the wallpaper  to White you can see that that screen is not   green at all anymore but I hope that gives you  some better insight to how I approach reviews   and like where I feel my obligation is to bring  stuff to the attention of a manufacturer before   I publish a video here on YouTube I never want a  company to be like completely blindsided I want   to give them as much of a heads up as possible in  this case we had a hardware problem that can be   fixed easily with software I think there's an over  Reliance to just say things are software problems   when they're really not software issues this is  an issue that can be fixed with software but it   is Hardware in nature this would never happen  if the backlight LEDs were not green like they   were before this fix was applied there are some  important implications for the software fix that   they applied here for the lifetime of this device  that I think the company would be better off just   like being open about I don't feel like it's my  place to describe any of that and for some people   that were like really making fun of the company  about how this happened to begin with I will just   say stuff like this can easily happen if you're  a small company for many years it was part of my   job to worry about small details like this and  it still is a large part of the real job that   I do outside of just making YouTube videos as a  hobby even when you have the best intentions to   worry about something like this and keep an eye  on it you can still get screwed by your supplier   and that's if you care about consistency I could  take out nine Miu minis right now from my studio   and they all have completely different screens and  completely different screen characteristics now if   you want to know how this looks for yourself and  you have an rp3 then you'll pretty much know what   to expect this is the rp3 screen that replaced  the blue one that I don't think a lot of people   have from the total lifespan of this device this  new screen with the configuration looks similar   to this but it has a bit more of a magenta tint  especially on the left half of the screen when   you're looking at a pure white background there's  also some magenta fringing in some situations but   largely this is way better than the first screen  was anyway here's the Benchmark data we'll start   off first with how it was before it was configured  and remember I pointed out three things that could   probably be impacted by a software solution these  three things would be the color gamut the maximum   brightness and and the contrast ratio for our  new software fix for this screen we don't have   an impact on our color gamut it's still basically  the same but our maximum brightness did take a hit   from the 700s down to the 500s when it comes  to the contrast ratio that took a noticeable   hit but you will only notice that in Darker games  in regular settings it's not that obvious we've   covered a ton of stuff in the first few sections  of this video but we have a lot more it's finally   time to do a deep dive into what this can do  from an emulation standpoint you are probably   looking at the Timeline right now wondering how we  still have more than 30 minutes of footage left in   this video and that's for a good reason there  are over 140 games in this emulation section I   think this is the most coverage that I've ever  done for a handheld and I hope this answers any   questions that you may have about what this can  do we will start with the fifth generation of   consoles with Sega Saturn everything before this  can run with no problems on this chip and quite   frankly there's no real reason to buy this model  if you care more about things that are older than   Sega Saturn I will point out that I'm using  the high performance profile for all of this   footage that isn't because we necessarily need  it for every system but it will eliminate any   performance issues that we would have from the  operating system now even though Sega Saturn is   very old it still requires a decent processor to  run well with the rp4 pro we can use the beetle   Saturn core and ra to run just about anything  with ease there are other emulators for Saturn   that have some more options than this but this  one is free without ads and it is plug-and play next up from here we have PlayStation  1 for this I'm using the duck station   Standalone emulator and I have the  rendering resolution set to just   about max out the panel with PG XP turned  on to fix some of the 3D scenes this one is a smaller section but this is Nintendo  64 with the Mupen64 plus core in RA I have the   rendering resolution set to 720p and I'm using  the wide adjusted setting to get a 16x9 image In this section we have Dreamcast  with the flycast core in ra just like   Nintendo 64 we have the rendering  resolution set to 720p and we're   using widescreen hacks to get a  16x9 image for compatible games like most of what we just covered PSP is a  system that ran just fine on the rp3 plus for this   we're using the PPSSPP Standalone emulator and  have the rendering resolution set to 3x native I showed off some Vita footage in my first look  and I wanted to include some more in this video   the vaita 3K emulator is still getting a lot  of updates with the most recent one dropping in   the last week even though some games run well  on this there are a lot of graphical issues   in games that run much better on Snapdragon  processors we have five systems left for this  video and the first four have a ton   of games in each one of them the first one  is 3DS with the Citra emulator I have the   rendering resolution set to 2x native by  default for all of these titles and only   change it to 1x if the game does not run  well at 2x if going to 1x doesn't change   anything at all I will only have footage for  the game running at 2x native our next big system is GameCube and for  this I decided to use only the official Dolphin   Emulator I have the rendering resolution set  to 2x native if a game doesn't run well at 2x   I'll drop the rendering resolution and try  again for almost all of these titles I use   the Vulcan back end but there are a few titles  like Mario Kart Double Dash that need open GL to work now it's time  to test some Wii emulation   on this I made sure to include a few  heavy titles for this video but this   handheld was able to chew through  most of what I threw it at with no issues I believe this is the biggest section  in this entire video this is going to be an   extensive look at PS2 performance on the  rp4 pro if you saw my first look you may   remember I talked about putting together a  systematic way to classify PS2 performance   in handhelds that work is still ongoing and  it has now transformed into what will be a   crossplatform benchmark that I'll be able  to use to test out all of the PS2 capable   handhelds that come out it's not ready yet but I  made sure to include a lot of heavy PS2 games in   this section as well as some viewer suggestions  obviously this processor cannot run the entire   PS2 library but that isn't only because this  is a mobile chip or that this has a Molly GPU   there are actually some games in this section  that run way better on weaker Molly gpus that   are newer like the one in the RK 3588 if you  can believe it still this thing holds up very   well but battery life is going to take a dive on  heavier titles that need all of the power that we have switch is the last system for this video  but the vast majority of titles for this system   will not run well or at all on this handheld so  my advice is to just think of this as a bonus   even when a game looks like it's running fine it  can still crash which probably isn't worth the   hassle in the long run now occasionally I get  questions on input lag in handhelds and I have   an entire video covering this topic with an older  retroid pocket model I only ever comment on input   lag if it falls outside of a normal range when I  do experience excessive input lag it is usually   a combination of a game running at a very low  native FPS or an issue with the emulator out of   all the titles that I tested in this video hollow  night for the Yuzu emulator was the only one that   had very high input lack this is strange because  the game shows it's running at 60 FPS but this   is not how this game feels natively on a PC  on the switch or on other handhelds using the   Yuzu emulator it's a bit strange but I wanted  to leave this comment to sum up my experience   with input lag across all of the titles that I  filmed for this video that's about all I have   for this deep dive video if you found this  helpful at all consider subscribing to the   channel there are no affiliate links tied to this  video for the rp4 pro so if you want to support   work like this subscribing is the best way as you  can imagine with all the games that we tested in   this video this was a ton of work to put together  I hope it helps happy gaming everyone Taki out

Channel: Taki Udon

Views: 231,998

Keywords: taki udon, taki, retroid pocket 4 pro, retroid pocket 4, rp4 pro, rp4, retroid pocket 4 review, retroid, retroid pocket 4 pro review, retroid pocket 4 taki udon, emulation handhelds, android gaming handhelds, retro emulation, ps2 handhelds, gamecube emulation, wii emulation, Retroid pocket 4 pro emulation, RP4 Pro Emulation, RP4 Pro emulation review, RP4 Pro PS2 emulation, RP4 Pro Switch Emulation, RP4 Pro Gamecube emulation, RP4 Pro Wii Emulation

Id: W2jSGUdtrZY

Length: 53min 42sec (3222 seconds)

Published: Mon Jan 15 2024