Recording Audio Sounds Worse Than It Should

[SkyBoi]

Hello, I've attached an example:

The audio sounds worse in finished recordings compared to the quality of what I'm actually recording. I've messed with a bunch of settings, but haven't had any luck fixing this. What's going on?

This was done on my HP Envy laptop, running Windows 11.

 

[AaronD]

With such a short example, and no "good" version to compare to, and such high energy (more likely to have effects on purpose), *anything* could be "what it's supposed to sound like". Can you post something longer, and both "good" and "bad" versions of the exact same thing?

 

[AaronD]

laptop

Most laptops are not good for media production. They don't have much cooling, and so they overheat and throttle back. Then the encoding (and thus quality) falls apart. At that point, you're essentially trying to run OBS on a *vintage* machine in terms of actual performance. It's nowhere near "modern" anymore, even if it says so on paper.
The idea behind that design is to be super portable, load an app or a document quickly, and then do nothing but slowly cool off while the user looks at what it just loaded.

Desktop towers almost always have adequate cooling, so they're fine.

If you MUST use a laptop, look at the Mobile Workstation class of machines. They're thick and heavy, because they have an actual cooling system! So they *can* use all of their published specs indefinitely.

 

[SkyBoi]

With such a short example, and no "good" version to compare to, and such high energy (more likely to have effects on purpose), *anything* could be "what it's supposed to sound like". Can you post something longer, and both "good" and "bad" versions of the exact same thing?

The example here does have both. The one with facecam is recorded in OBS, the one without is an export from premiere. Sorry for not making that clear.

 

[SkyBoi]

Most laptops are not good for media production. They don't have much cooling, and so they overheat and throttle back. Then the encoding (and thus quality) falls apart. At that point, you're essentially trying to run OBS on a *vintage* machine in terms of actual performance. It's nowhere near "modern" anymore, even if it says so on paper.
The idea behind that design is to be super portable, load an app or a document quickly, and then do nothing but slowly cool off while the user looks at what it just loaded.

Desktop towers almost always have adequate cooling, so they're fine.

If you MUST use a laptop, look at the Mobile Workstation class of machines. They're thick and heavy, because they have an actual cooling system! So they *can* use all of their published specs indefinitely.

Yeah, I use both. The issue is my recording backdrop just doesn't fit behind my desk setup, so I have a little portable table if I want to have facecam. Long story, and I'm hoping there's a way to get it to work without moving my desktop whenever I want to record.

 

[AaronD]

The example here does have both. The one with facecam is recorded in OBS, the one without is an export from premiere. Sorry for not making that clear.

So...the output from OBS is good, but Premiere is not?
Live -> OBS -> File -> Premiere -> File

 

[SkyBoi]

So...the output from OBS is good, but Premiere is not?
Live -> OBS -> File -> Premiere -> File

No, the opposite. The first clip is the good audio, the second clip is an OBS recording of the good audio.

 

[AaronD]

No, the opposite. The first clip is the good audio, the second clip is an OBS recording of the good audio.

Okay. It's still hard to tell a difference, but maybe the second one is *slightly* lower volume??? Now that you've told me which is which, which makes it no longer a blind test, and therefore less valid to confirm a significant difference. (see above for the blind-test results)

Higher volume tends to be perceived immediately as "better" in a direct comparison of the same thing, and everything else needs more time to find. More than what you gave here. That's a common trick in "audiophile" demonstrations, and part of why the nickname "audiophool" exists.

 

[SkyBoi]

Okay. It's still hard to tell a difference, but maybe the second one is *slightly* lower volume??? Now that you've told me which is which, which makes it no longer a blind test, and therefore less valid to confirm a significant difference. (see above for the blind-test results)

Higher volume tends to be perceived immediately as "better" in a direct comparison of the same thing, and everything else needs more time to find. More than what you gave here. That's a common trick in "audiophile" demonstrations, and part of why the nickname "audiophool" exists.

It's very apparent with good headphones. I produce music so I know quite a bit about audio.

 

[R@de]

Comparing audio is always a very subjective thing. For me there is a night and day difference between the two.

Tells us your OSB settings.

 

[PaiSand]

Like always, an OBS log file from the recording is more than helpful.

 

[SkyBoi]

Like always, an OBS log file from the recording is more than helpful.

Here's the log: https://pastebin.com/BQrdDcFy

Here's the settings that seem relevant, please let me know if you need to see more.

 

[PaiSand]

Would you be kind and attach the log file in here? You can also upload it via the option built for it on the Help menu or directly on the analyzer.
The button to attach is down bellow.

Do not use stream encoder for recording. Select a different one. Even if it's the same NVENC you're using on the stream. This is to make the GPU use a second encoder for the recording.

 

[koala]

Your logfile doesn't contain a recording session. Without a logfile containing an actual recording session that includes the issue it's impossible to say anything. The logfile contains the settings OBS did actually use, so it's the only reliable source for troubleshooting, because it's a protocol of what actually happened, not a description of what should happen.

 

[SkyBoi]

The first attached is the log I linked on Pastebin. The second is a brand new recording I did because I was told the other one somehow didn't contain a recording session, in case this helps.

 

[PaiSand]

In the log file you only use OBS for streaming. You should use CBR instead of CRF.
Please do a test selecting an encoder for recording, not " use stream encoder". You can safely select x264
Or instead of using x264 use the QSV encoder.

Hardware Encoding | OBS

OBS Knowledge Base. Choosing a Hardware Encoder

obsproject.com

If the recording goes thru the plugin you're using then you need to look into this plugin support.
Also take into consideration that Intel's 13 and 14 CPUs have unfixable issues.

Analyzer | OBS

OBS (Open Broadcaster Software) is free and open source software for video recording and live streaming. Stream to Twitch, YouTube and many other providers or record your own videos with high quality H264 / AAC encoding.

obsproject.com

 

[SkyBoi]

Here's the log file using the x264 encoder. I don't believe it's going through any plugins. Just tested, and when I record on my desktop, the audio is nearly indistinguishable from the source. So, do you think the CPU is the problem?