I assume the following setup
- there is the 'optimal place' (topic of endless debates) and there is your back too.. Put the mic on a good sturdy mic arm and adjust the place. You can record yourself with video to see your own movements.
- Good Mic placement beats badly placed expensive microphoses as well: a stable, but not looud enough signal is easier to 'fix later in the chain' than one that is 'very dynamic' (has silent and loud parts).
- Mics record in specific directions. Get in that manual of the mic and check the 'characteristics' it has. Some can switch some can not. You can to look up 'cardioid characteristic' and find out how to point the mic.
- we completely ignore the room. They seem fine, i hardly can remember anything else than your voices from many hours. My guess would be John has a louder environment.
- Nutshell: Make something louder adds all kind of 'noises'.
- The basic noise in your room, the noise the devices create etc.
- If you 'crank it up' different types of distortion happen. Think E-guitar vs. Accoustic.
- In the early stage you want to be clean.
- Room has a noise floor. Just listen to your room: PC, open windows with street
- What is noise? Nutshell: sounds that consists of all frquencies. it is hard to filter, once it is loud.
- The SM7B mics need so called 'hi gain' amps. Those adding about 60 DB.
Your amp
- good decision to go with the same setup for each of you.
- its amplification seems to be fine. Use google translate on this article
- It should do the job.
- Its kind of barely making the cut (you are using 56 db of possible 60 if I remember the image correctly)
- The spark plug comes with the software. its adds about 3.5 ms latency to everything but does a lot of magic.
Shortcomings to consider
- a device that rests on the desk and has the mic directly plug in makes sense
- a LED strip that shows how 'loud' you are works wonders.
- USB powered: at least invest one of the expensive ports and buy a very good cable. I assume you run windows desktops
- you maybe want a button to adjust monitoring. In your case maybe not relevant
- 'get the level of the recording right' is a big discusion in the larger audio community.
- You are not recording in the studio, but doing live broadcasts. In studio technology, these 2 areas have 2 different types of mixers for example. In your case this results in the need to get levels a bit more right than for example a studio recording, that is not so much dependent on this.
- Deactivate everything and just record normal audio into audiacity (should be possible when recording a cast to have the real deal audio material).
- The better you hit the right level between -infinite (absolute silence) and 0DB (it is about to break everything).
- It makes sense to aim for a 'stable result'. You cant fix mumbling in post and a good mic will just record mumbling even better.
- For levels, about -10 to -12 dBFS is about right for voice / narration.
- You ears tell you if it is good. You are professionals at all, so apart from a lot fo gatekeeping on how you do it right. Here is some creative freedom. Just be aware that all steps that come afterward are built on this data.
- Take notes on the levels of your normal audio tools as well. Not every meter is the same kind of meter and they might show different things.
AI MIxing
- will give you a good idea, but hardly will be perfect. You want to know how your system reacts to specific things (annoying birds, neigbours kids).
Noise gates
- set audio level to zero when below a certain level (threshold)
- Are a last ressort.
- Deactivate where ever you can, or keep at low levels. See it like this: when you have someone pointing a gun at you, you want that person to get the difference between 'cunt' and 'stunt'.
Hi Pass filter
Is your friend. Removing low frequencies, not being part of the frequency spectrum of the voice helps all later processing. 75, maybe even 150.
Limiter
- Basically a protection on the upper part of the level.
- When being very excited, screaming at people, this will make it possible I can watch the stream and still have the toddler sleep next door.
- Having a limiter engaged might be the difference of hear loss or not
- A limiter should not be limiting normally, so it is fine to activate. would be fine if you yould configure it.
After the limiter you can be sure, the signal has the maximal gain you choose. This is important for livestreams. You need to find out what the expected maximal db is for each service and serve just a tad less. This is kind of impossible to get right at this stage, so securing a specific maximal level with a limiter is fine.
EQ
- This is where you add color by adding things ..... only thing is ... most times you dont add things
- Rule of thumb: for everything you add somewhere, you should remove somewhere else.
- remove something and amplify the rest
- Lot of people dies in the internet wars about the right eq curve. You need to do a lot of things right before to have something to eq to
- handle with care.
Compressor
- in a nutshell makes the quiet parts louder
- can be used to make pronounciation easier to understand etc.
- seems to be mostly presets in your case: these are hit and miss. One might get a good sound with a preset, another one might have the broken fan, road, angry bird or whatever and get different results for subtile differnt inputs. Even using the same hardware, with the exact same settings.
- Dial it in. I did not listen to everything and jumped the first minutes, but later when George talekd about the complaints, it was fine for me. Was close enought to the original.
- My guess would be Jon has more noise in his environment.
- George has dialed it up a notch, but sometimes comes across a bit distorted. Might be the color of the voice.
- I dont know how the 2 voices are mixed back together. I am sure there is magic in discord that tries to make things right and handle the slightly differnt pitches in the voices differently when mixing the 2 voices together.
- Sounds weird: consider a vocal coach. Before the sore throats and other stuff starts. You will learn about the type of voice and this is honestly what you need to know to eq it.
- That Broadcast voice: no this setup wont make it. The mic is something that should stay with you for a while, but in reality every 'famous' voice has a very dialed in setup. Those people have audio engineers.
- All the DSP stuff in the spark plug is rather limited. Dont fall for GAS (gear acquisition syndrome) and use it to its fullest extend. It is a simple tool, but master this, before you click for a more expensive option.
- I guess something with much more dsp power (e.g. custom custom plugins) would be an option. I would refrain from processing audio on a heavy loaded streaming pc. even if it is possible. The spark plug and more extensive 'audiointerfaces with dsp' will still do all the magic while the pc is eating up that unoptimized alpha of a survival game
- There is no shame in getting professional help: I am just a onbnoxious dilletant myself, so there is a limit to what I can do in my homestudio, but if I were a streamer with that many hours of 'online talking' i'd just get it dialed in by a pro and work from there. All that expensive gear is for nothing, if not set up right. Once it is done: no need for further services. maybe not now, but the next stage for sure. Paid Services, that even can be done remotely are widely sought after and even local classifieds can get you somewhere (maybe there is a studio around the block).
- The moment you will get jons voice on the level of georges, some other challenge will pop up. But giving the final mix 2, not super dinamic signals (well spoken - which you do - well recorded) with a limited maximum db is a very good start.
- Owning a cloudlifter and so called di-box (kills signal hum) gets you very far.
- Where do you mix your 3 signals: sound of game, voic jon, voice george. Could be another topic
What is special about a dynamic mic? Why do I need a 'cloudlifter' for the sm7b Why not just crank the amp?