Audio Lag, Drift, Latency, IDK. Audio Issue

[ricmetal]

hello,
i'd like to know what the issue is here.

1. i'm recording video and device audio (not mic) in OBS (some gameplay, or even a youtube video playing in the background has the issue)
2. i'm recording audio with my mic in Audacity (speaking over gameplay, or even not speaking over anything i record (the issue happens weather there's a lot of audio information being recorded, or nearly non at all - if that's even relevant or not) - the only purposeful audio recorded is a start and finish clap for syncing the Audacity-recorded audio file and the OBS-recorded audio/video file, which i do later in Vegas)
2.a. (there is no sync issue with the audio and video recorded in OBS when placed in Vegas)
3. the issue is visible when placing the OBS file (audio and video) and the Audacity audio file into Vegas: the Audacity audio file is smaller in length than the OBS audio/video file, even if i start recording in Audacity first. So the Audacity audio file should be longer in duration that the OBS file. But it's not. After some time The OBS file becomes longer. In a 10 minute recording session, the Audacity file is about 10 seconds shorter, increasing the difference to 20 seconds if i record about 15/18 minutes (values not accurately represented, i don't remember the exact numbers.
The issue is that by the end of the recording session, OBS has recorded a lengthier file than Audacity!

anyone know why this happens? and how can i fix it?

here's a picture with recorded times.

[Attachment 52034 - Click to enlarge]
again, i started recording Audacity first, yet after minutes the duration is shorter than OBS.
i stopped the recording less than one or two seconds in between each other.

things i can tell you.
1. im using a usb mic
2. sample rates in OBS, Audacity and Window (that shared option) is 44.1 khz
3. OBS records the file onto a HDD (not the system)
4. Audacity records to the system SSD
5. Recording for 2:22 minutes (on my phone's stopwatch) actually recorded only 2:19 seconds in Audacity. da hell?

i tried the stopwatch thing on obs. 5 minutes in obs equal 5 minutes realtime.
im assuming this is the usb mic's problem then? would a jack mic solve this issue? or an external audio board?

[Cornucopia]

That’s a strange issue, if what you say about common sample rates is true (the likely problem scenario), but I think you could easily get around the whole thing by recording all audio parts through OBS, since it is certainly capable of that.

Scott

[ricmetal]

i'm having issues with the mic quality in OBS, like audio being cut off for milliseconds, like a stutter in recording every few seconds.

regarding the sample rate, this wouldn't be the issue that causes the discrepancy between the recorded length in audacity and the stopwatch though, right? that doesn't sound like it should be the problem.

[Cornucopia]

If it's misinterpreting the sample rate, it could be.

For example. Let's say you are recording 10 seconds. At 44.1kHz sample rate, 10sec * 44100 = 441,000 samples (per ch). At 48kHz sample rate, 10sec * 48000 = 480,000 samples. If the app expected all samplerates to be the same (in order to properly mix them)*, but didn't interpret them correctly on input, there would be a discrepancy of almost a full second (0.88...) between the two clips (the 48 being longer than the 44.1). The misinterpreted one would also be improperly pitched, so that could be a clue.

*In general with computers, to properly mix, ALL framerates (video) and samplerates (audio) must be the same upon import. This is also true during simultaneous-record time, and I know of no device that works in concert with computers that operates independently WRT those rates (IOW, it must be completely disconnected in order to free-run).

BTW, what USB mic are you using? There should be no stutter, etc (that also sounds like it could be a common samplerate mismatch).

Also, if you are using any time-based USB device, be absolutely sure to NOT have it go through a hub, but be direct to the mobo ports. This will very often screw up due to timing mismatches.

Scott

[ricmetal]

so no, i don't use hubs. mic is connected to the back panel
mic is this one: https://www.amazon.es/gp/product/B071171DBP/ref=ppx_yo_dt_b_asin_title_o03_s01?ie=UTF8&psc=1

[ricmetal]

omg. it's the mic that's crap, or something.
i did the stopwatch test with another jack mic i have here and times are perfectly synced.
slow facepalm....
how do i know what mic won't do this kind of effect?
should i be buying jack mics from now on?
faith in hardware..gone

so maybe that's the cause for the issue when recording in OBS. OBS was trying to keep the audio length correct and the mic couldn't keep up. smh. back to shopping for a mic.

also, plugged into a usb 2 or 3 port makes not difference. still records like crap.

[Cornucopia]

Are you needing strictly a clip-on/lavalier/lapel mic?

These days for simple podcast, I will swear by Samson GoMic. Mountable. ~$40USD. Great sound (for <$100USD), switchable pickup patterns, switchable DNR or AEC.

Yeah, never heard of your mic, and the ad doesn't give much confidence either.

Scott

[ricmetal]

i chose the lapel mic cause they where the cheapest
and now thinking about it, it's useful to clip them instead of having mic stands.
ill still consider a lapel mic, but a jack mic and see
because im looking for a cheap mic that wont have much noise floor or whatever you call it, so i dont have to do much audio fixing.
perhaps ill have to go a bit higher on the price than my previous mic to get a good silenced mic?

[Cornucopia]

IMO, you won't get ANYTHING good for <~$30USD. While companies may be able to go cheap w the A/D and USB sections, the actual mic capsule is still analog and is the weakest link, and thus the most important to get right, quality-wise. Don't skimp on that part.

Scott

[ricmetal]

is there anything in the mic specifications that can tell us the expected quality of the capsules? or the amount of expected noise?

[Cornucopia]

You COULD check the usual stats:
Freq response within -3dB falloff range
Phase linearity
Self-noise
Thd
These specs have been good objective references since at least the 70s.

But that just tells you part of the story. No mic is perfect - all have coloration (though the B&K testing mics come close). Some coloration is more palatable, some very not so.

You could look at the specs of the best studio condenser mics and work your way down, which could take some time. However, you aren't going to find a "you must be this tall" (to ride the ride) threshold marker of what is good vs. not.

One thing to do is check on boards similar to this that have pro or semi-pro (or possibly even knowledgable consumer) reviews of mics (this board does not). Go with people or groups or companies you know and respect.
I wouldn't recommend these "best widgets of 2020" ad-based gonzo faux journalist lists as they are often totally non-scientific as well as funded by those same companies listed.
I hate to say it but it is quite a bit harder to figure out mics with new connection tech such as usb or bluetooth because the dust hasn't settled yet, and China is undercutting a lot of traditional manufacturing, but it's a very mixed bag quality-wise (usually cheap, but not universally warranted).

I did a quick search, and nothing comes up that is a diamond in the rough. Either regular diamonds, or regular rocks.

Scott

[ricmetal]

speaking of diamonds in the rough, im starting to see people record their audio into their phones with these lapel mics.
im guessing the phone have less noise associated with it, in comparison to the computer hardware, is this true?
its seems very complicated to look for specs for a mic, i see
and about the floor noise, i read most of it comes from electricity/static and stuff, which seems to me there would be less in phones.
would this be the reason to use a phone to record over a computer?
i might go the cheap route again (but jack connection), and try using my phone to record instead.

[Cornucopia]

Not really true. Convenience.
Your phone is a very compact computer. Compactness generates heat, heat often raises the noise level, compactness also means elements are closer than ever and there is a greater chance of interference (= more noise).
What's good about phones is they are battery operated, which means true DC, so certain AC types of interference are not possible. But there can be DC interference, or interference from other components as mentioned above.
Again, overall, the difference between phone & computer digital sections are minor compared to the difference between A/D sections, and that again is miniscule compared to the difference between microphones, their quality and their placement (and environment).
You MIGHT find a lapel mic w/ 1/8" (3.5mm) audio plug that is a better overall level of quality than your USB offerings, for a reasonable price. Partly because there isn't the added onboard A/D & USB digital sections required, and partly because more traditional mic makers are already covering this market.

Scott

[ricmetal]

a'right. appreciated the answers.
just a side question: when you say i wont find anything below 30 bucks +-, this is a threshold, something you said didn't exist (the height to ride the ride). you where basing this statement on experience i presume?

[Cornucopia]

Yeah, experience. I worked for 20+ years as an Audio engineer in production/post. Just what I've seen so far. Could be wrong, esp. if new tech changes the paradigm. But that happens rarely in microphone construction.
If you find that hidden inexpensive gem, do post your experiences here so we all can benefit.

Scott

[ricmetal]

will do. thanks very much for the insight.

[ricmetal]

recorded some audio to use in a video, using my phone, instead of pc/mic and the quality is great compared to the jack mic i have.
just a very very faint hiss that i can remove completely, without damaging the integrity of the voice.

so i'd go ahead and say between a 10 dollar mic and a semi-shitty phone (vdf 610 - not an iphone or something along those price lines), the phone's quality is actually way superior, so it's a clear choice.

just wanted to put that out there!

[babygdav]

1. https://steamcommunity.com/app/689790/discussions/0/1694920442946536734/
2. Video - capturing using constant frame rate?
What standard?
Ought to be 29.97 fps ntsc .... It's complicated, but audio usually syncs to that. If you're recording at let's say 60fps video and import audio, it's trying to sync the audio at 29.97fps.

Project properties.
Ought to match. E.g. 29.97 fps.

Ser everything to 44.1khz, 16-bit.

[ricmetal]

afai understood, any audio setting i use should record 10 seconds, if i record for 10 seconds (timed on my phone's stopwatch) i should get 10 seconds worth of audio, no?
"recording for 2:22 minutes (timed on my phone's stopwatch) actually recorded only 2:19 seconds of audio"