AMD "Record & Stream" option missing on Laptop 16

Framework Laptop 16
1

Hi there,

I received my Laptop 16 with Radeon 7700S a few days ago, and so far everything is going flawlessly, except for one little detail: the “Record and Stream” option (also known as AMD ReLive, apparently) is not showing in the Radeon Adrenalin software.

This is my first AMD card in a loooong time, and I got used to the “Shadow Play” features in the NVidia software. You can record games (or even your desktop) with it, and it also has a hotkey to save your last 5 minutes of gameplay, in case something funny happens that you want to keep.

This AMD ReLive thingy looks to be the red team equivalent of that, but I can’t get it to work. People with the same problem online usually have desktop machines with AMD CPUs, and disabling the iGPU of those seems to fix the problem. I obviously can’t do that. I also tried to completely remove the drivers with DDU and reinstall everything from scratch, but the problem persists.

Is it working for anyone else ? Ideas on how to fix it ?

Thanks,
Gaëtan

2 Likes
2

Hi, just realized that I also don’t have that option in my Adrenaline drivers. I have used AMD GPUs mostly (before FW16 I had PC with RX580 and the record option was definitely there). It is weird, but I think maybe because we use custom AMD drivers for FW16 specifically, Devs just… omitted that functionality so drivers are lighter and you can always different soft to capture screen? Too bad…

3

I suspect it is more to do with DMCA requirements than lightening the drivers.

4

given that @HappyPond had difficulty with it on the mainstream AMD drivers (and this likely being due to having two GPUs in use), I would instead suspect it got disabled in the Framework drivers because it couldn’t be made to work on the Framework laptop.

After all, whose intellectual property would the feature violate?

5

The alternative would be using OBS, but unlike ReLive, there isn’t a way to have it run in the background and save the last 5 seconds of footage.

Unfortunately for me, I use Fedora so there’s no ReLive anyways there, so I would just have OBS running and recording. Again, not ideal, but I do this. :frowning:

6

OBS can run in the background and save the last N minutes. It’s called “Replay Buffer”.

grafik
grafik529×283 4.18 KB

1 Like
7

I do not see this option, and I am on the current version (31.0.0).

8

Netflix, Amazon, Apple+, etc. I can’t see any streaming service wanting any recording method of unencrypted video being allowed.

9

the feature defaults disabled, you have to turn it on in settings

image
image964×329 30.6 KB

2 Likes
10

Thanks! This will come in useful!

1 Like
11
counterargument to the replied-to post

but capture cards are legal to own in the us arent they? not sure since although i have one i live in the uk. and before you mention hdcp it is known that some hdmi 1 input to 2 output splitters will silently strip hdcp out (and also allow you to watch your content while capturing it).

also windows has built-in application window recording

also there are other pieces of software to record the screen

also the amd one can in theory run on other machines so why would framework be different

also you could literally set up a camera to record video from a display (and get it quite good with some effort) and connect a 3.5mm cable to a line in input, then sync them up by playing a bit of video with an audio cue synced to a visual cue to sync this, which you can make in any video editor, and thats if your camera doesnt have a line in input which i reckon some do anyway)

im pretty sure that windows has some flag that applications can apply for not wanting to be recorded (remember copilot recall? it was documented to not record copyrighted content, precisely through this flag. android also has an equivalent flag which prohibits screenshots - apps like signal actually let you enable it manually in their settings to prevent malware from capturing your messages easily but you can also disable it if you want to screenshot your messages yourself and its not on by default iirc, if your phone is enterprise managed your enterprise can also turn it on systemwide, and finally if you have a rooted phone there are apps that can prevent the system from respecting it in any app. i reckon macos and ios also have such flags. i dont think linux has the same thing though - iirc any program with root privileges can just read the framebuffer file, on x11 any app can read any region without any permissions whatsoever, and on wayland most compositors have some xdg-desktop-portal implementation that allows them to request a screen capture and then the user chooses what is to be captured)

saying that the option isnt there because of streaming services intellectual property licenses is like saying that elevators freefall downwards (see postscript for an explanation of why this isnt the case) - despite initially seeming logical it only makes sense if you don’t really think about whether the implications would be logical. even though i wouldnt put banning or attempting to ban screen recording software (even specific screen recording software) past some intellectual property holders, given that the amd one still exists in the generic driver build id say it isnt the case. after all, it seems much more logical to say that since the framework build of the driver is meant to have tweaks to keep it fully compatible with the framework hardware, and that the screen recording present in the generic version of the driver doesnt function on framework hardware, the capability was removed to ensure that nobody relied on something non-functional to work.

and now for the tangential postscript on the causes of elevator freefalls that was mentioned in the above counterargument

elevators freefall upwards because the counterweight is heavier than the elevator cab, so that when the elevator is at about half capacity the weight of the car and its load is about equal to the weigth of the counterweigt. this means that a lower power motor can be used. since elevators are designed so that they cant easily be overloaded due to the physical dimensions of the car, this means that an elevator is usually under half load. since policymakers generally arent informed about specific implementation details and rely on assumptions, they generally assume that elevators freefall when the cables break, despite these cables having multiple redundancy and being able to support significant overload (so that most of them have to break for a freefall downwards to start) and thus only mandate prevention of this. in practice, elevators generally freefall when the motor is powered down while the elevator is at a floor, due to brake failure (in normal operation, these brakes are only used for this purpose, although they are also used to stop the lift if the safety circuit disconnects (the safety circuit is only connected when all landing doors are closed, the car doors are closed, and the emergency stop button is not pressed) while the cab is moving fast enough or if the cab is far enough away from a landing, and also they are used to stop the elevator while it is in inspection mode if the controls are released - but in normal operation the lift is decelerated using the motor as it approaches a floor). if the brakes fail while the elevator is under half capacity, it will freefall upwards (and in the uk, it will generally continue to do so until it hits the buffers because most elevators here have no means to properly detect and respond to this failure condition, since protection against upwards freefalls was only mandated for elevators installed after the lift regulations 2016 came into force). if the lift was to freefall downwards instead, the governor would trip and activate the track brakes (separate to the other brakes) i.e. blocks of metal are shot into the tracks that the elevator rides on (these prevent it from swinging around in the shaft and only allow it to move along a straight line under other circumstances), at which point you’ll be going nowhere. (after the lift regulations 2016 came into force in the uk, elevator governors also had to trip for upwards overspeeds as well as downwards ones).

12

I might actually use this, thanks for telling me that this is a thing!

1 Like
13

glad to have been of service. i’ve never actually used the feature myself, so i don’t personally know how competent it is.

if it’s terrible, i’d like to hear how so i can adjust my recommendations to others accordingly XD

2 Likes
14

Here are some heads-ups you might find useful, and that I wish I knew earlier:

Places to enter your screen resolution (there are multiple, and for me it works if all of those are the same):

  • Settings -> Output -> Streaming -> Streaming Settings -> Rescale Output
  • Settings -> Output -> Recording -> Recording Settings -> Rescale Output
  • Settings -> Video -> General -> Base (Canvas) Resolution
  • Settings -> Video -> General -> Output (Scaled) Resolution

Don’t forget to set Settings -> Output -> Replay Buffer -> Maximum Replay Time. I set it to 600s.
Here are the Settings -> Output -> Recording settings that work for me on my Framework 16 Laptop:

grafik
grafik983×762 33.6 KB

You’ll need to do some experimenting with the bitrate. The maximum I can select is 100000 Kbps, but 70000 seems to be enough for me. At 3440x1440 this typically produces files slightly below 5GB at slightly below 10 minutes length.

If you like to have your microphone, maybe a game, and other sound recorded to different audio tracks (so you can choose e.g. whether you want your voice audible if cut out a snippet), then I recommend the plugin “win-capture-audio”, that allows you to whitelist and blacklist specific programs when recording to different audio tracks.
Go to Settings -> Output -> Recording -> Recording Settings -> Audio Track and check all numbered checkboxes. Then in Settings -> Output -> Audio give those tracks appropriate names. I highly suggest something like “Mixed audio” for the first track, and then I have “Game”, “Teamspeak”, “Desktop”, “Microphone 1” and “Microphone 2” for the other tracks.
Go to Settings -> Audio -> Global Audio Devices. I don’t really know why this exists in the first place, but here I have “Mic/Auxiliary Audio” set to one microphone (external) and “Mic/Auxiliary Audio 2” to a secondary microphone (laptop-integrated), and everything else “Disabled”.
At Main Window -> Scenes, click the plus button, name it something like “Gaming on external screen” and confirm with “OK”.
At Main Window -> Sources, click the plus button, select Application Audio Output Capture (this is the plugin), select “Create New” and name it e.g. “TeamSpeak Input”. In its settings, add the executable “ts3client_win64.exe” to the list of executables (either manually, or by selecting that executable on the bottom and clicking “Add executable”). If you are using Discord, it’s probably called “discord.exe”. You can add multiple here (e.g. Parsec, Skype, etc.). Confirm with “OK”.
Create another input and name it e.g. “Game Input”. Add the executables of the games you play (e.g. “r5apex.exe”, “factorio.exe”, etc.). Confirm with OK.
Create another input and name it “Rest of Desktop Input”. Add all the executables from “TeamSpeak Input” and “Game Input”, then make sure to enable the checkbox “Capture all audio EXCEPT sessions from the selected executables”. Confirm with “OK”.
At Main Window -> Audio Mixer, click on the gear button. I recommend reducing all sources to 30%, because if you are shouting, your friends in TeamSpeak are shouting, and the game is loud, the “Mixed Audio” track might be clipping. You should experiment with this. In the “Tracks” section, enable track 1 for all sources, so that track (which is you named “Mixed Audio” in the settings) will receive all audio. And for each source, check the corresponding single track. For me, this looks like this:

grafik
grafik1324×749 42.7 KB

“Profiles” are a bit weird at first but mostly make sense once you get it. Profiles include “most settings”, according to the docs, which definitely includes the output video and canvas resolution. If you plan to capture different screens (e.g. gaming on an external screen while at home vs gaming on the primary laptop screen when away), then you need to create one profile for each screen resolution. You cannot switch profiles while the replay buffer is running, because that could change the video resolution, which you can’t change in the middle of a video. I recommend setting everything up for one screen and making sure it works, then creating the first profile, cloning that, and modifying the clone. And create one scene for each screen that you want to record, add the same audio sources (you can add existing ones; note that they are not cloned when doing so, but changes you make to one source while in one scene will be reflected in other scenes) but different video sources.
So for recording the entire screen in one scene: At Main Window -> Sources, click the plus button, select Display Capture, select “Create New” and name it e.g. “External Screen Capture”. In its settings, you’ll need to experiment. For me, “DXGI Desktop Duplication”, with “Capture Cursor” enabled and “Force SDR” disabled (not running HDR anyways) works for the external screen, but the laptop’s internal screen has the yellow border around it and you might not be able to get rid of that because all modern software sucks.
It’s probably also be possible to record multiple screens in one video file by setting the canvas and output resolutions appropriately and placing multiple video sources in one scene, but I haven’t had a need for that yet. Capturing just specific windows should also be possible, but I don’t know how that behaves when the window size changes. YMMV.

You can set Settings -> Hotkeys -> Replay Buffer -> Save Replay`. I use Alt+F10 because I’m used to how it was in the software formerly known as ShadowPlay.

If you are forgetful like me, edit your link to obs64.exe and append --startreplaybuffer to the target path (including a space, so it’s e.g. "C:\Program Files\obs-studio\bin\64bit\obs64.exe" --startreplaybuffer), so it automatically starts the replay buffer when you open it.

If you are forgetful like me, you’ll likely lose recordings of your voice because the selection of Settings -> Audio -> Global Audio Devices keeps changing on you if you ever unplug your microphone. Always double-check when you open OBS that when you speak into your microphone, the corresponding “Mic/Aux” thing in the Main Window -> Audio Mixer shows activity. Selecting “Default” instead of a specific microphone does not save me from this nightmare. YMMV.

When I talk loudly, the recorded microphone sound is clipping, despite everything being reduced in the audio mixer. It’s clipped to that level. My microphone is not particularly high-quality, but TeamSpeak and Discord do not seem to have this problem. Again, YMMV.

Close OBS before hibernation/standby. If I don’t do that, it will suck an entire core of CPU usage upon wake, and I need to kill it with task manager.

1 Like
15

with the recorded microphone sound clipping, you could try turning down the gain manually in windows settings or pavucontrol (linux) or equivalent

windows has automatic gain control, but i dont think it works properly when multiple applications are using the microphone (and its possible that obs doesnt implement it to prevent issues in other applications using the microphone).

edit: you could also try turning the gain down in obs.