AI vs Human mastering

[excuse me please]

oh my, bitwig 7 years old?
time flies eh...

Well, I'm old anyway

[cryophonik]

For me, some of the best value I've had from working with human ME's has been their feedback on my mixes. My home studio has far from ideal acoustics, so my mixes are generally problematic. The few ME's I've worked with have given me some excellent advice on specific fixes to make to my mixes before they started the mastering (rather than turd-polishing) process. You won't get that with an AI mastering service, so that alone was worth the cost and effort of using a real ME.

[ghettosynth]

The next most interesting thing is that the $5/track service from CloudBounce, received 4/5 stars on audio quality, the same as the human based studio that you liked and a better rating than Abbey Road.

You can get a lifetime plan from PB for £95 atm.

https://www.pluginboutique.com/manufact ... loudBounce

I tried a song in it, but didn't like the result. I preferred my masters from Ozone. It only lets you try 1 song!

TBH, I don't think that any of these AI sites are doing anything magical. Ozone tells us that it's not a tough problem to solve anymore. I'm not arguing that Ozone is a better value really, but I am saying that if you're going to buy some bundle and it comes with one of these simple A/I mastering tools then there's not really any point to using these services.

Frankly, if I didn't have these tools I'd just use BandLab for everything. If you don't have a reference to compare it to then I don't see what difference it makes?

I see them going away in a few years. As I said, this will just become a part of the rendering process in your DAW. It's an analysis and rendering process that takes almost no time and relies on fairly well understood processes.

So much about mastering is based in FUD. I think that the second pair of ears argument has merit if you are serious about your music. I also suspect that having a name on your demo might grease the red tape a bit, but I might be completely wrong there, I know jack shit about the business side of music.

For most of us here though who get excited about single digit listens on bandcamp, there's absolutely no point, nobody is going to care and you shouldn't spend a nickel on mastering. If you have Ozone, use the assistant to tell you roughly where the maximizer threshold can be set and then compare that to some presets with the maximizer threshold adjusted similarly. Choose the one you like the best and call it a day. If you want an easier approach, upload to BandLab, download, upload to Bandcamp. You are done and I doubt that any of your listeners will ever think twice about your mastering.

[ghettosynth]

For me, some of the best value I've had from working with human ME's has been their feedback on my mixes. My home studio has far from ideal acoustics, so my mixes are generally problematic. The few ME's I've worked with have given me some excellent advice on specific fixes to make to my mixes before they started the mastering (rather than turd-polishing) process. You won't get that with an AI mastering service, so that alone was worth the cost and effort of using a real ME.

Yes, that's what I think several people are saying. The second pair of ears argument is the main reason that you should choose an actual Mastering Engineer.

Or, you could use BandLab, upload to some site for preview, and post here or on GS and get feedback on your mix.

[Hermetech Mastering]

Got an album worth of unreleased tracks and have been looking for mastering when this question comes to the for.
Sites like LANDR with AI powered mastering services rather than the traditional mastering engineer. I am going back and forth in my head on this one. Thoughts, options, etc.

I guess it depend upon how much value you place upon your music. Nothing like getting mix feedback and building a long term, musical, working relationship with a human being who wants to help you achieve the best sound you can possibly get.

[Goldwell]

A human mastering engineer can also give you feedback on your mix. Although a good mastering engineer will also sonically beat an AI service, there is the added benefit of having someone check if the track is ready for release. Imagine a track with a vocal drowning in reverb. A real mastering engineer might tell you that, but not the case with Landr.

[ultravioleteverlong]

Ozone seems to be more or less the same as the online services.

[Constructed Identity]

EPILOGUE:
After finding out my #1 choice was no longer mastering and my #2 choice never responded I bought Ozone 9 to do it myself. The 'Master Assistant' will be my guide. This is a cheap choice since I got a loyalty sale price of $99

For me, some of the best value I've had from working with human ME's has been their feedback on my mixes. My home studio has far from ideal acoustics, so my mixes are generally problematic. The few ME's I've worked with have given me some excellent advice on specific fixes to make to my mixes before they started the mastering (rather than turd-polishing) process. You won't get that with an AI mastering service, so that alone was worth the cost and effort of using a real ME.

This is true if you have someone you trust for the particular music genre you make. Ideally, someone with great credits that you want to emulate/aspire to. Probably 99 percent of people making music would benefit from advice. Strictly speaking this is consulting, not mastering and if you can find someone who will do this for you that's great.
I have to admit I am a weirdo who wants to blend genres and try new sonic ideas though so I don't think I am in the majority here.

[TribeOfHǫfuð]

I am learning in the music cafe. And I use Reason only. It is a very social experience compared to just using the automaster options of our soundcloud pro (whose results I do not like anyway). I am a control freak and hungry to learn. So forth, I have made some bad mistakes but they were all corrected in the music cafe with help from more experienced members. If you want to master fast, you are missing a lot, imo, but it is seemingly possible.

[Constructed Identity]

I am learning in the music cafe. And I use Reason only. It is a very social experience compared to just using the automaster options of our soundcloud pro (whose results I do not like anyway). I am a control freak and hungry to learn. So forth, I have made some bad mistakes but they were all corrected in the music cafe with help from more experienced members. If you want to master fast, you are missing a lot, imo, but it is seemingly possible.

I learn by doing. I have already made multiple masters of the same track to figure out the proper levels etc. There is a lot new to Ozone 9, but it is the same general workflow as Ozone 1/2 in which I have years experience. iZotope really has done a lot to improve the user experience in the last 15 years and I am very happy with the purchase.
I still need to find out more about using a reference track and I will be searching for use guides.

[Hermetech Mastering]

My fellow ME Justin Perkins updated his article on the topic recently, it's a great read:

https://theproaudiofiles.com/what-autom ... o-for-you/

[BONES]

This is true if you have someone you trust for the particular music genre you make.

This was always a huge problem for me when I went into a studio. All the owners/engineers ever dealt with were rock bands and at the time even I didn't now that what I did had its own genre. To me it was just what I did but I had very strong views on how it should sound and those guys were of little or no help. That said, they always did a great job, it just wasn't always the result I had envisaged. At the time I couldn't see that, I was always disappointed, but when I look back now I can hear that they had actually done a good job with the shit material (me) they had to work with.

[Burillo]

My fellow ME Justin Perkins updated his article on the topic recently, it's a great read:

https://theproaudiofiles.com/what-autom ... o-for-you/

The true test is this: Run a song through an automated stereo bus processing service, and then send the result back through again. If a human mastering engineer with some intuition and sensibility was sent a song they already mastered to master, they would do no additional processing because they already did exactly what the song or project needed the first time, right?

this quote irked me as this is obviously a very bad argument. sending the same mastered track to the same mastering engineer is not the same as sending the track to automated mastering service twice. the real test would've been to send it to a different mastering engineer (since we cannot make the mastering engineer "forget" that they have already mastered that song), not to the same one.

the article in general is not very well written IMO, because it sets up a main premise (that mastering isn't just bus processing), but spends like ten paragraphs meandering on all the ways that it isn't "just bus processing", without actually telling what it is. it got so boring that i started scrolling through the article without reading it to see if the author actually gets to some sort of point that they're trying to make. like, yeah, it isn't bus processing, i get it, move on already!

luckily, the article does spend the latter half of the article explaining what should happen besides bus processing, but even that is way too long. like, they spend an entire paragraph describing the different types of plosive sounds and noises that automated mastering can't fix or even detect (which, okay, makes sense), but this could've been said in one sentence. the author seems to get lost in their own article too, and goes on about noises, pops and clicks, and then somehow in the same section they talk about track sequencing and DDP. like, okay, i thought we were talking about plosives and iZotope RX? the sections sizes are wildly different and make no sense as well, there's no overarching "flow" to the article.

overall, i get the point, but the article needs to be rewritten and re-edited by someone who actually understands good writing. i disagree that it's a "great read". it's informative, no doubt, but it could've been much better written.

[Scotty]

A human mastering engineer can also give you feedback on your mix. Although a good mastering engineer will also sonically beat an AI service, there is the added benefit of having someone check if the track is ready for release. Imagine a track with a vocal drowning in reverb. A real mastering engineer might tell you that, but not the case with Landr.

Landr will tell you that, you have to listen to its output and compare it to your vision. At the bare minimum you'll learn what it can't do and you'll get better at mixing and putting in the colouration that you are wanting by hand. I have been mixing into landr for a few years now. An average mastering engineer can't do better than Landr with my material in so far as I have used it to learn. I give my Landr masters to the mastering engineer as a reference and ask them to listen to it and tell me what they would improve. If they can't tell me what they are hearing specifically then why bother. Anyone can add a bass bump or some top end air. My material is already sonically balanced the way I want it to be. I don't want a mastering engineer to paint a canvas on top of my material. It needs to be subtle and effective. They need to hear some little detail I've overlooked, find a way to make the vocals a bit more clear without losing too many other details. They need to know what problems they are trying to solve before getting to work. Just my two bits.