Like you all know there were too many instances of the snare in this mix, which were all out of sync. So I did a lot of shifting and nudging of tracks to align the snare, with quite a satisfactory result. Only less welcome artefact was that in doing so a doubling effect appeared on the vocal. Anyhow I had finished the mix already when I thought of using AI in the form of a stem splitter. I still had the weblink to an online stem splitter of Bandlab (https://www.bandlab.com/splitter) and I decided to use it on a separate mix. I tried to clean up the vocal, bass, guitar and violin. All except the guitar with great result. Like with any AI driven apps, sometimes it works and sometimes you get horrible results. On the vocal, violin and bass the results were good, so I swopped the original trackswith the treated ones. The guitar track still had the snare in it, but that may be due to the fact that I used the free version of the app, which doesn't have a separate splitter for the guitars, so I used the others option. I have been trying to implement this app on several other mixes that had a lot of bleed, and like I said sometimes it works but often you have to resort to pre AI methods because the outcome sucks. In the end I made a few more adjustments to the vocal track, and that's it.
That's a lot of processing but ultimately both mixes sound good to me. It's a totally different approach to mine, where I didn't do any time adjustments or alignment and didn't make any effort to increase the separation. The bass seems really big and heavily compressed to my ears. Maybe the bass is a bit too dominant at times e.g. violin solo? Great work though and really interesting to hear about your approach.
Ooft, Aart... I have heard some lovely mixes from you many times. but this is one heavy mix for a folk track. It has a good balance, aside from the kick and bass. I think maybe reduce some processing and it will come together nice.
Aart this is very intresting! I thinck such methods (with AI or without it, but with others progressiv techics) will take a significant part tomorow. but still open question is - how good it works today...
of course in case experiment, demo, traning - it good. or in case we need to remix song and multitrack of it not longer exist, and we have only not good mixdown... I think any result that will give this technic - is better then nothing... but working on sourses with a claim to quality (not simple this project was recorded in 88200) - we have more higher requairments to tools...
how do you think, did tool you chuse complite the task? (couse I mean 'comlite on high level')
Thank you all, @John Hope-Crossley , @Klaus Ferretti , @Nate Rae and @Vitalii Mizhenin for your interest and your comments. It seems that this week I had more eye (or ear) for detail than for the bigger picture. From the minute I started with this song it struck me that the snare sounded like it was hit twice. Apparently a case of delayed bleed. Most obvious in the bass track, which was considerably slow. I was so surprised that nobody else today even mentioned noticing it. I thought maybe something had gone wrong importing the tracks. But that is hardly likely. I am very happy with the sound of the bass and kick, and thought they deserved a prominent place, but I failed to hear they were a bit too prominent, to say it nicely. Thanks guys, for making me aware. Feedback is most wanted and needed.
Vitalii, I think AI is not good or bad, and in this case it's harmless, but I wouldn't trust it when it comes to political matters or things that concern our safety and wellbeing. But that aside, Like I already mentioned, the same app sometimes does a great job, and other times, depending on the material it sucks. The treated tracks came out quite good today. But with the bleed also the ambience was filtered out. That's why the second mix sounds much cleaner and dryer. I should have compensated for it by adding some extra reverb or so. And yes, this time it completed the task on a high level.
Hey aart, thats intresting.. a few things: Do you check if all the tracks were in line? i dint found any issue with the snare sounding delayed And i was very surprised of how good it sounds from the get go.. almost no phase issues with the bleed, the only thing out of phase was the kicks in and out, and thats probably the nature of the recording technique used. In second place i always do a lot of cleaning with band lab splitter, i found that you dont need to go 100%, you can just leave a bit of bleed for a more natural sound with less artifacts, just the reduction you need. Yeah sometimes is prety phasey and weird, but it is just a tool, and its incredibly powerful. The mix, like the guys already pointed, its sounding a bit over processed for me, the bass is prominent and distracting, i think is the compression on it what causes that spikey sound. And the kick is a bit heavy, to my taste, and maybe check that phase i mentioned on it. Just a few thing to consider, great work so far man
@Mariano Moncho , thanks for your comments. To be clear, the snare itself isn't delayed, but the bleed from the snare in the bass track is delayed compared to the snare track. I've heard it on several other mixes too. Second, if you want to leave some ambience in using the splitter, you'd have to work with more of the split tracks? Or do you mean blend the cleaned up track with the original one? That may be the best option. I always check the phase of the kick tracks and they were fine. It's just a little loud. Same with the bass, it's loud, but not extremely processed. The sound comes mainly from EQ-ing, making it more audible by introducing more mids and less lows. That in itself makes it sound louder and more prominent. Compression is not too much but in signals with such a long sustain you'll always hear the release of the compressor, no matter if it's fast or slow. Maybe parallel compression could cure that?
I forgot to mention, the bass in itself is late. If you listen to the bass you can hear, it's always behind the beat. That makes sense if the bleed in this track is late. It must have been pushed from its position when the tracks were exported from the original mix. (by accident or deliberately)
So I didn’t give the AI a rating) no good no bad ))) I talked about it as a compressor (for example). more precisely, about some plugin that is used everywhere today, but just 20 years ago it could not exist in principle. so does AI. Today some look askance - but tomorrow it will use everybody)))
my interest was that during the work process you heard individual tracks before and after processing and had a better opportunity to evaluate than we did, when we listen to the whole mix.
ps and as for the issues of safety and well-being - there really is no need for AI... let's say AI hands-off!!! people can handle creating problems for themselves
@Vitalii Mizhenin , it was also in guitar1, violin and vocal, but not all the same delay. In room track a little delay would be normal and even desired, because it's for creating ambience. :))
ambience it't course important thing in records such this. but sounds it not attractnve 😁 with what 'flame' from room track better then 'flame' from bass track?
To further complicate the matter, if you listen to the room in solo on headphones you hear that the placement of instruments in the stereofield is all wrong. The same goes for the bass track and a few other tracks that are full of bleed. This also negatively influences the clarity of the snare, kick, vocal etc. If you add up all those wrong stereo images, instances of let's say the snare come from various directions, contributing to the lack of focus. With a monomaker plugin you could center everything which is already better than what it is now, but more advanced spatial plugins like Schoeps Double MS can shift positions around until you find a more suitable stereo image setting.
@Aart dB For the splitter thing, sometimes, when i feel that it is not working like i want, i just open in the studio close to the download button, its like a daw online and you can tweak the volumes and other cool things, then just download the mix Hope that helps man!
@Aart dB I wasn't around over the weekend to submit my mix or comment on others but was intrigued by your attempts to deal with the snare sidestick. To be honest, when I was mixing this, I didn't notice the time problem but couldn't help but notice how damn loud that sidestick was in general and especially in the bass track. As I perceived this more as a volume problem than a temporal one, I just utilized a sidechained MB Compressor on the bass keyed to the snare mic to cut that nasty thing out which had little to no impact on the tone of the bass. I used the same tactic with the acoustic guitar track. I just went back to the project and bypassed these two compressors and, wow, I total hear the time delay now! The compression totally did the trick to make it disappear in my mix.
@Tyson Arp , thanks for sharing this. Sounds to me as a great option to solve the problem. I'm gonna go back to the project and make a version with the sidechained mb compressor, to get a grip on this method.
0 props
Space Description
It's a place where ones who want to improve their mixing skills can join their forces with an aim to make the process less boring and more productive
We use cookies to improve your experience. Essential cookies keep the site running. We also use optional cookies to enhance performance, analyze traffic, and personalize ads. By clicking “Accept”, you agree to the use of all cookies.
Sep 01, 2024
Sep 01, 2024
Sep 01, 2024
Sep 01, 2024
Sep 01, 2024
Sep 01, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 02, 2024
Sep 04, 2024
Sep 04, 2024
Sep 05, 2024
Sep 05, 2024
Sep 06, 2024