** Official Recording/Streaming Tips & Tricks Exchange **

[uncle_rufus]

One of the pieces of footage I've imported into my editor timeline has out of sync audio (yes, ShadowPlay variable bit rate). However, I've noticed that after running it through handbrake, part of the footage is now synced properly as desired, but another section isn't.

Do you just run it through handbrake again, or try something else? I don't always find it works 100% first time.

Yikes - that doesn't sound good. I can't imagine running it through again would have any effect... when you say out of sync do you mean by a constant amount? Or is it sort of getting worse as time goes on?

 

[uncle_rufus]

Constant amount. It doesn't get worse, thankfully. In fact, the bit that was detected later on, is only ever so slightly out of sync. I may just end up leaving it as it is. I guess it's that or throw the piece of footage away.

If it's a constant amount can't you just try and identify roughly where it starts and shift the audio yourself? (perhaps masking the transition with a natural seeming cut in the video)... Most editors will let you separate the audio and video and then you can simply slide one or the other to correct things...?

 

[uncle_rufus]

Wish me luck . I'll be trying to put out 4 - 5 videos a week from now on.

Good luck! Don't burn yourself out though; that's a lot of content if you want to do any sort of decent editing on it... (See how you go though, of course!)

It's just coming up on the 1-year anniversary of starting our Youtube channel... lost a bit of steam lately mostly due to lack of games that we want to cover, but broadly speaking still happy with what we've put out...

~50 subs after a year seems pretty low, but at least a decent number of those were organic... A few "successful" videos which got a few hundred views... lots of learning about how to do things, maybe year 2 will be more successful!

 

[uncle_rufus]

Sounds good . I hope you have a good 2nd year of YouTube. I'd be happy with 50 subs it goes to show that people actually like your content.

Just out of interest how many videos do you do a week and what is your upload schedule? I'm trying to plan out mine at the moment and was just wondering what other people did.

I was thinking a video on Monday, Tuesday, Thursday and Friday and if I have to miss one of those I can do an extra video on Saturday or Sunday just in case.

Cheers... Our schedule (and content) has changed a fair bit over time... Initially we were doing daily videos, because our theory was that getting a lot of content out there would help get the channel off to a good start... We managed to maintain this fairly well for a while, but it was quite hard work so we eventually cut it back to just 3 videos a week...

The other main thing is we realised that you don't get much of a response when you cover games that are too popular and/or done to death by bigger channels already... So our content moved more towards trying to cover smaller new games (all of our biggest videos are ones where we managed to cover a brand new indie game pretty much)... Since then we've tended not to bother releasing content unless it's worthwhile (and there haven't been enough couch co-op titles lately to satisfy that)... We're thinking of carrying on Dead Space 3 just because it happens to be halloween and we want to play it

Do we know of any software that allows you to record audio from a specific program, such as If i wanted Discord/TS to have its own audio track without any interference from other programs.

Or is it something that can only be bodged with Virtual Audio Cable?

Something So I can manage audio levels a bit better in Sony Vegas, like one track has the game audio, another track has my voice which was recorded in Audacity and then another track that's only recorded Discord/TS audio.

Hope that makes sense.

There's two different questions you're asking here I think... How to output sound to different audio devices, and how to capture sound from multiple devices. Some recording software covers the latter (DXTory or OBS will let you have multiple audio sources on separate channels)... To split your output sound into multiple devices Virtual Audio Cable is one option, or also a free alternative called VB-AUDIO cable, though I've not used that one

 

[uncle_rufus]

Indeed +1 for OBS Studio... although currently I still use DXTory for a lot of my recording, as I'm waiting on a particular feature to be made available in OBS (at which point I am probably going to switch to OBS exclusively)

FWIW the feature I'm after is simply the combination of two features which currently can't be used together: the custom ffmpeg output option (which allows you to output proper high-quality capture instead of stuff that's already compressed) and the multiple audio streams... ffmpeg itself definitely supports multiple streams so I think it's just that someone needs to plug the two together in the GUI

 

[uncle_rufus]

Hmm. It would indeed be useful to have multiple audio tracks with lossless video on OBS. I did buy DXTory but decided I didn't have the hard drive space to record in lossless which is why I switched over to OBS Studio in the first place. So far though I have to say I am happy with it after using both FRAPS and DXTory it seems to have far more options and there are some really good tutorials out there on how to use it.

Yeah it takes up a lot of space - our 2 recording machines both have dedicated 2TB recording drives so it's not such a big deal to record lossless (well, *nearly* lossless; diminishing returns after a point!). In terms of features and user experience though I really like OBS - the first time we had to use it was for recording Dead Space 3, which DXTory refused to work with (I think the Origin overlay was confusing it)... and I actually found once I got the custom settings right for the ffmpeg output mode the recording was superior to DXTory (from a performance perspective, with the same codec - UTVideo). Luckily for Dead Space 3 we didn't need the multiple audio either because we recorded both POVs and mics, so OBS was fine...

The strange thing is that ffmpeg itself supports all those features that you want. I used to use it on Linux all the time and it is some really advanced media software. Have you tried making a feature request in the OBS Studio forums? That might help your case or at least the developers will be able to tell you that they are not planning on implementing that feature.

Indeed it does - I had a look around on the forums when I first used OBS and did find a thread or two mentioning support for this that suggested it was on the list of planned features. At one stage I was planning on raising a pull request and putting the plumbing code up myself but haven't gotten around to it (hard to justify working on software sometimes in your free-time when you do it for work!)

Personally though I am just using the Nvidia encoder because I have a laptop CPU and it isn't all that powerful so I need to offload as much stuff as possible to my Nvidia 1070.

Nothing wrong with that - the NVENC encoder seems to be an absolute beast performance-wise, and if you try to structure your recordings to require fairly limited editing you can get away with the quality

 

[uncle_rufus]

Does Audacity allow you to record more than one audio track at once? I recorded some game + microphone footage earlier, but the sound of my voice was drowned out at times by the game audio. If anyone knows of a good tutorial for this, then that would be very helpful.

I also tried using ShadowPlay. I increased the volume & boost of the microphone, but it didn't make much, if any, difference.

I think my settings are right, but here's a couple of pics anyway:

Spoiler

I don't think Audacity can multi-track, but you can launch more than one instance of Audacity... I'm not sure that's what you need here - probably better to do something like:

1. Set Shadowplay not to record your mic - this will give you the game audio on its own as part of the video file
2. Record your mic with Audacity
3. Do a manual sync - once both applications are recording but before you start your intro you want to do an action in-game that you can easily match up to something you say; for instance say you're at the main menu, maybe you'll say out loud:

"Okay, I'm at the main menu now - going to be flicking between "New game" and "Load game"... 1, 2, 1-2-3, 1, 2, 1-2-3... and again... 1, 2, 1-2-3, 1, 2, 1-2-3"

As you're doing the above counting move the mouse or hit up/down on your controller so that the menu options change in the same rhythm.

4. When you come to edit that's how you can line up the audio and video in your editor - make sure you always move the game audio with the game footage, as they are already lined up, but find the sync point in the footage and in your mic recording and then carefully line them up. If your editor has a way of displaying the waveform for the mic audio that can usually help as you can line it up partly using the peaks in the audio from your speaking.

Some people like to do a manual sync at the end of their recording session as well, or even mid-way through it - this can catch problems with the sync drifting over time (although fixing them can be tricky if not impossible)

Hope that helps

(p.s. also thinking of your settings - they look fine although you could give it a test record with the mic set to "Mono" - microphones aren't generally stereo devices, unless the AT2020 has 2 drivers inside it or something to allow it to pick up directional sound... and to be honest even if it does you'll probably get a cleaner sounding recording on mono anyhow; you don't want your voice to sound like it drifts from left to right in the balance)

 

[uncle_rufus]

Nice tutorial I can only echo the recommendation to use something like OBS rather than ShadowPlay (I should probably have mentioned that) - ShadowPlay is really only ideal (at the moment) for footage which you don't intend to edit; unless you're using an editor which can cope with the variable-frame-rate it produces (or as Halfmad says; you're happy to process everything you record to make the frame-rate constant)...

Also worth mentioning though that the VAC / voicemeter stuff is only really needed if you need to split the output from one device into 2. By default your speakers or normal sound listening device is already a separate device to your mic, so you can record these separately with any program that can record multiple tracks. Where the virtual cable stuff comes in is if you've got 2 separate programs (for instance a game and the sound of your mates talking on skype / teamspeak / ventrillo / mumble)... By default both sounds are coming from the same device (e.g. your speakers)... the virtual cables let you re-route the two separate bits of audio so that they appear to be coming from 2 separate devices (which a different program can then record as separate tracks)

 

[uncle_rufus]

The problem now was that it sounded like my voice was drowned out again. I can only assume that it was the export settings that did that.

I guess one option is to look into what PP can do to enhance my voice, which it may well be able to do, but I haven't tried that yet.

What processing (if any) did you do to the audio before exporting it? I know you mentioned noise removal and "cleanup" but what about things like compression, normalisation, EQ, ducking? If you check the earlier pages in the thread I'm sure there were some discussions about this you could have a look at but you'll always get better results with some proper treatment of the audio... I'll try to remember to update this post with a link when I get home (I'll have a look at your screenshot too; damn work firewall blocks out imgur etc )

Edit: what about this post: https://forums.overclockers.co.uk/showpost.php?p=29347480&postcount=170

 

[uncle_rufus]

That's a pretty good deal!

Has anybody else noticed that Youtube now has some sort of in-built end-slate / end card tool? I'm not sure how long it's been there for but I'm tempted to give it a go and see how it works... anybody tried it?

 

[uncle_rufus]

Your hard-drive will take most of the strain (particularly if you're recording lossless), but this is dependent only on the dimensions and quality of the output file - a 1080p output recording with an embedded DSLR "face-cam" in the corner shouldn't tax the drive anymore than a 1080p output recording with no "face-cam"... but if you want the recorded DSLR footage to be totally separate then I think you want it going to 2 different drives/machines (as you've indicated) because you're recording to 2 independent files...

As for the audio source I'm not familiar with Mac versions of software but on the PC side most recording programs can handle the idea that you might have more than one possible audio source attached and you only want to record audio from one specific source. Easiest way if you've got all the gear is to dedicate a few hours to trying it out and see what happens

 

[uncle_rufus]

Essentially I think that's what I'm going to have to do.

I was planning to record the DSLR separately. Pro tools will be recording to an external HDD (as I always do). I might try recording the DLSR straight onto the rMBPs SSD.

I expect that'll be okay... the only remaining thing you'll have to do is ensure you have a way to sync up the recordings... I think what I'd do is start recording the audio first, then clearly say into your mic "I'm pressing record on the DSLR in 3,2,1,now" and then do the same again for the game recording software, but make sure your mouth is visible in the DSLR shot when you do the countdown for the game recording. That will give you some clear points in the audio waveforms and in the DSLR to get everything sync'd up

If you want to be extra thorough you can then do a more traditional in-game sync (i.e. move the mouse between two defined points on the screen whilst vocalising it, like "new game - quit - new game - quit - new game" on a main menu or something). Hope that makes sense

 

[uncle_rufus]

I was just going to clap each time we hit record so it can be synced up afterwards in adobe premiere.

That's another way to do it and I know quite a few people do it that way as well... I prefer a vocal cue because a clap can be quite a huge spike in the audio, and I find a series of rhythmic vocal peaks easier to line up...

However additionally in your case how will the clap help you to line up the game audio? Or do you mean you'll hit record on the audio, then clap as you hit record on the camera, then clap again as you hit record on the game? How many hands do you have? (or do you have a helper?)

 

[uncle_rufus]

Ah okay I had assumed you were recording it to be like a "face-cam" / live commentary type thing in which case you would want your reactions etc. to things happening in the game to be as sync'd up as possible. You're right as well that you can just sync up the DSLR and mic audio by looking at your lips moving etc. but it's deceptive how tricky it can be sometimes to work out roughly where the "start" is if you get what I mean...

 

[uncle_rufus]

I've got a C920 as well; great webcam

 

[uncle_rufus]

When exporting audio (Audacity), how come the wav file is so much shorter than the original recording? It's like 95% gets lopped off.

Think I've sorted it. It seems it doesn't like the 32-bit wav file, though I could have sworn I selected that for a previous video. Anyway, this time I tried 16-bit and all seems fine.

This sounds pretty odd... I usually export to 32-bit and haven't noticed any problems; what format was the original recording?

 

[uncle_rufus]

do any of you back up the rendered project/keep the original video files even once your work is finally uploaded onto your channel?

I don't back up the original files because they are massive if you record lossless... but I do keep copies of the rendered files (I guess if I ever decide to make a compilation/montage type video it'll be marginally better quality to use the originals rather than downloading from Youtube)

Recently upgraded monitor to a 3440x1440 and am noticing a few oddities with shadowplay. At the moment im getting them with BF1 but will test more games at weekend. Set to record in game resolution and 40% of my recordings are in the shadowplay folder as 1280x720? And some of them are 3440x1440.

I have tried changing a few things around like using Alt + F9 to manually record instead of capturing the last bits of footage. Still happens there. It seems random, any ideas?

Hmm... Shadowplay can be quite tricky at times - you could try using the same NVENC encoder but on a different program (like OBS) as it may just be Shadowplay the application is messing up (I haven't used it in a while but it felt quite basic compared to other recorders out there)

 

[uncle_rufus]

If i continue to get issues at weekend then i will look at OBS some more but the recording i got from it was simply awful and the performance it even with NVENC was significantly more than shadow play even when at a lesser bit rate. ( i could have set it up wrong ) ((Probably definitely set it up wrong)) but yeah i would prefer to use shadow play if possible as i love the shadow play last 20 minutes feature. It's a god send.

Yeah that's a really nice feature, would love to see a buffer feature in the programs I do use to record (but I expect the memory requirements when recording at higher qualities prevent this from working)...

It sounds like you've found a workaround for BF1 which is good - your explanation sounds plausible as well... you do tend to find when trying to record different games they all have different quirks and things; sometimes in the most extreme cases the way a game works means it can't be recorded at all by certain programs (depending on how they work) - I found this with Dead Space 3; DXTory wouldn't record it no matter what I did, I think maybe the Origin overlay was interfering with it somehow, so in the end I had to use OBS...

I get what you mean about OBS/DXTory... they can be a lot more fiddly to use; but it's generally a one-time sort of thing; once you've got them setup to your liking you don't tend to have to go through all the setup again every time

 

[uncle_rufus]

Can you elaborate on what you want to end up with? When you say "edit them together" do you mean with cuts... picture-in-picture... how will the two bits of footage be merged exactly, and to what end?

 

[uncle_rufus]

Essentially, 2 of us will be playing together (One on the PC, one on the MAC). We'll record our gameplay together on separate HDDs then edit together cuts of the 2 recordings. Mostly clean cuts between the 2, maybe the odd picture in picture.

Ah okay - that's pretty much what we did with our Dead Space 3 videos - although in our case it was both on PCs playing at 1080p...

Still I'm not sure that 2 different resolutions would look too bad as long as the aspect ratio is the same... It sounds like if anything the Mac might be a higher resolution scaling down to 1080p which I imagine would be fine (although your rendering may be slower since every frame has to be scaled down)... you could perhaps use a tool like Handbrake to reduce the resolution of the Mac footage first prior to editing? Are you sure there's no way to output at 1080p on the Mac? What about running the game in windowed mode, will that allow you to set the resolution?