So after being given a trial premium month with some exportable .WAV files..(thanks!) i deceided to try it out on one of my recent tracks. I have posted one version mastered by myself and one by LANDR. see if you can tell which one was done by their software and which by my ears.
bear in mind im a hobby level producer and have very little mastering experience myself other than mad zachs schmastering 101 and some basic EQ knowledge.
Since you basically said you’re unfamiliar with mastering, it’s not really a fair comparison. I’d be much more interested in hearing comparisons between pro-level mastering and this auto-mastering shindig.
I tried Landr’s trial and I was not impressed. It absolutely sucked all the bass out, or conversely has pushed up everything from the middle frequencies upwards. Your track hasn’t got much low end anyway, but track B has been less squashed and sounds like the low end has been retained a bit more, which is why I think it’s not Landr.
Soon we can just have an algorithm software to make a song or dance track, and then what is the point?
When you have a track mastered, it shouldn’t sound much different from the original minus the loudness and a hint of dynamics. If you have a shit mix down, LANDR will give you a shit result. Hell, I’ve had good mixdowns and LANDR has given me a shit result. A mastering engineer should get personal with your track and really give it what it needs, not use the same template for every job.
for sure , they reccomend that you basically send it in with little if any work on the master track, so i killed my master effects before sending it in, then turned them on for my own mix down. the only eqing or compression i did during the creation of the song.
I’m actually kind of amazed both how well it did and how far off it was from what I’d be happy with. Frankly, I just don’t think landr is ever going to be good enough. It is cheaper than the %60/track that I’d spend, but…I’ve heard enough to trust the guy I’d send things to.
There’s also the fact that they call themselves a mastering service and will degrade themselves to deliver 192kbps mp3s. I think that alone should disqualify them from anyone even halfway serious. Just work on the mixdown until you’re happy and slam it into a limiter until you see GR peaking at like -3 if you’re too cheap to pay someone who actually cares.
The free version delivers 192kbps, the paid version delivers WAV..
In fairness its the first of its kind service and really handy for the average budding garage band or solo artist to throw a few MP3’s on Soundcloud or youtube without shelling out a fortune for mastering every single rough track individually.
However I gave it a whirl earlier on a completely raw wav that had zero post processing done and it made a total mess. The compression/Limiter sounded like they were whacked up to the max and there was serious clipping and distortion.
Someone who didn’t know better and doesn’t have a decent set of cans or speakers might upload it without a second thought and think “thats how its meant to sound” :-S However the creators have said its adaptive, so the more people that upload tracks the further it gets tweaked.
Ive been using Landr for testing tunes before sending off for mastering for the past couple weeks and havent had any issues. How much headroom are you guys leaving?
My Meters were at around -5db on the raw track, however interestingly it did a reasonable job on the same track when I added some compression and EQ prior to uploading.
this was the first thing i noticed and pointed out in the blog at which point they offered me a free month of premium service. i wasnt expecting miracles or anything, but again i just thought i would share the results and start the discussion since they were kind enough to give the freebie.
I usually leave about -6 to -8 db of room for the mastering engineer. Also, what effects do you use on the master channel deejaysnafu? I never put anything but an analyzer
Every mastering srvice Ive ever used requires at least 6 db of headroom.
I just do so out of practice now. Do that with Landr and you should notice a big difference. You should never have anything on your master channel. You can get a lot cleaner mixdow if you bounce out a multitrack and do your mixdown on a multitracka s you are using way less CPU and open up more CPU time to work on mastering each individual track.
You can get a lot more out of subtractive EQ’ing your master track than by running a compressor as well.
True. But, it still seems odd to me. I’d rather either do it myself or pay a real person. Besides, software can’t say “hey…this is cluttered” or “dude, this is amazing.”
True. But, it still seems odd to me. I’d rather either do it myself or pay a real person. Besides, software can’t say “hey…this is cluttered” or “dude, this is amazing.”
Most decent mastering houses process eq and compression in the analog domain and then maybe into digital eq/compression and finally dig limiter.
In that scenario ideally you would record your un-mastered mix with the very loudest transient peak reaching as close to digital 0dB without it clipping to get the best dynamic range. What mastering houses don’t want is digitally limited mixes with clipping. Obviously something like LANDR processes purely in the digital domain and the level you feed it will have varying results, so maybe you would leave 6dB digital head room.
I’m not convinced that’s accurate. First, there’s nothing inherently better about analog. There can be things inherently better about specific pieces of gear, but it’s not like there’s a magic silver bullet where running any signal through a specific box will make it sound better (a lot of gear snob techs like to describe a lot of things this way, but it’s not actually true).
You don’t need to peak anywhere near 0dBFS to run things through analog hardware. Analog hardware is made to work right at specific voltage references (+4dBu, usually), and they’re going to add/remove gain as necessary to make their hardware work right if they know what they’re doing. All peaking near 0dB does is make the mastering house turn it down before they can do their work, no matter what gear they’re using.
Whether this applies to LANDR is another question.
No, there’s nothing inherent to say it is factually better.
I worked in mastering for around 16 years and I still have friends who work at the best mastering houses in London. I don’t know a single engineer who will say their favourite eq is a digital one, or compressor (except digital limiters for loudness). There are some decent digital eq’s including plugins that sound pretty good too. Digital eq’s are great at notching out. Weiss do a very good digital outboard eq which has been pretty popular for mastering. Yes, there are occasions when it will be kept in the digital domain. Maybe some of the smaller mastering places are keeping it all in the box with plugins most of the time; but it seems many of the bigger mastering boys are using plugins more in conjunction with analog processing. No you don’t need it to be 0dBFS to work in analog; you have trim controls to bring it to the correct level.
