One-Synth-Challenge: General discussion thread

[Guenon]

Hmm, any word on the stem / freezing / rendering thing? My system is starting to choke as predicted ...

[Guenon]

To reiterate, would it be okay to:

1) mute some tracks, say, everything but drums (etc)
2) render the project, in one go and full length
3) load the render as wav on an audio track, doing absolutely no further editing on it
4) remove the drum instances, archiving the project with the instances still in place
5) continue composing with more CPU

[z.prime]

In the spirit of the competition to have a level playing field, I would say this could be allowed as long as you can provide all the track projects individually. This is somewhat in line with the restriction to only use free plugins in that regard. However, there are times when the synths are very high CPU, which offers a challenge in and of itself to make something with a limited number of instances. My vote on this is it's OK but strongly discouraged.

[Guenon]

I would say this could be allowed as long as you can provide all the track projects individually.

Naturally, the idea would be to archive the project before any instances are removed.

However, there are times when the synths are very high CPU, which offers a challenge in and of itself to make something with a limited number of instances. My vote on this is it's OK but strongly discouraged.

But people are using freeze freely, which circumvents the high CPU challenge anyway? I'm asking this because the actual freeze function doesn't work as intended on my system, and I generally avoid using it, running everything in realtime. Tyrell just seems to be killing my CPU, on the type of material I'm currently working on, big time.

[Guenon]

And thanks for discussing this, btw. I think it's a good discussion to have, as it specifically says in the rules, "for convenience or CPU saving, synth bouncing are allowed but only where the EXACT same result could have been achieved using the original synth", and also "Audio must not be touched at all once it leaves the DAW (aka after bouncing a track or a stem)"

So, in other words, it currently states bouncing would be okay even for convenience, without any technical concerns like high CPU load (bouncing "for convenience or CPU saving" ), and also specifically mentions using stems, as long as they aren't touched afterwards.

If it reads explicitly like this in the rules, in my opinion it shouldn't be frowned upon or strongly discouraged? I'll do my very best to keep the Tyrell track rolling in one single project without bouncing, though. Just want to be absolutely sure it's not a negative thing if I have to bounce it at some point.

[L-EctroBit]

Hi Guenon. Recently this topic was consulted by doctorbob in the Helm contest.

viewtopic.php?f=1&t=467516&start=105#p6543231

Hi all

Just started on my track this month. Going to need 30 or so instances of Helm, but I find I run out of cpu (patch dependent!). So, I am going to split out the project and render down small subsets to wav files and then mix. Each stem is a full length track! Hope this is OK.

dB

I don't know if your case is similar or different.

Edit: the purpose of my comment is not to create controversy, but rather to clarify the subject, because I'm also interested in this method of work here in OSC.

[Guenon]

I don't know if your case is similar or different.

Edit: the purpose of my comment is not to create controversy, but rather to clarify the subject, because I'm also interested in this method of work here in OSC.

Thank you, it's very similar indeed. Splitting into full-length subset bounces is exactly what I was considering for mixing the Tyrell submission, if I can't make it all run at the same time. And yeah, definitely, the purpose of my questions is to discuss this in good spirits , no controversy. I'm very interested in this, too, as it's brought up in a neutral fashion in the rules, like quoted above (bouncing as a valid method of managing a project), but in practice there's a somewhat unclear "forbidden path" vibe surrounding it, heh.

[Frostline]

To stay within the rules how can one end up with a good sounding final product if any stems or bouncing is used?
The whole "can not be touched again" in regards to stems/bounces seems to mean the stem can not then be compressed, have any reverb added to it or have any EQ or other enhancement done to either the track containing the stem or the Master track of the project. Because that would be touching/changing the original sonic character of the stem right?

So if you can't put any processing on the Master or stem how do you get it all to blend together and at a consistent sound level/quality?

Everything I have ever read in regards to mixing/mastering shows mixing down to a stereo wave @ about -6dB(on average, some people like -12dB and some like using a hotter mix) and then doing a Mastering pass over the stereo wave to do final clipping, compression etc and raising the final max level to closer to 0dB.
To me the rules for this contest indicate I can't work in this method. I have to do everything in the mixing phase. Which I honestly can't seem to get to work.

It just seems to me like the whole not using stems issue is directly proportional to the amount of PC processing power you have. If you have the PC to never need stems then it is not an issue. But if your PC is a bit weak,or the current VSTi is a bit greedy then having production techniques forbidden that would help overcome PC deficiencies is quite a handicap.

Just seems ironic in a competition where the instrument and effects all have to be free that the person who can spend more on a PC has more freedom in exploring creatively what the free instrument can do vs the person who is barred/strongly discouraged from using free production techniques that don't depend on brute force processing power.

[Guenon]

To stay within the rules how can one end up with a good sounding final product if any stems or bouncing is used?

I agree it could be worded better and clearer, but the thing is, there is no point in rendering a bounce/stem in the first place if you aren't allowed to load it back into your daw for mixing and rendering the final track. Exactly as you say, yes. So, the actual point of these rules (and please anyone correct me if I'm wrong!) is to make sure nothing is done to the audio that wouldn't be possible if you were using the realtime plugin instances instead.

As in: if you render out stems and load them back in, you treat those waves as if they were the actual output coming from the synth in realtime. Again, they are what comes out of the synth, period. You don't do edits on them, touch the waveform itself in any manner, and so on. Setting levels, eq, etc, for the purpose of rendering out the final mix, is of course fine (as rendering the stems would be pointless if mixing them wouldn't be allowed , just as you point out, and this would be done in exactly the same manner if you were operating on the output of the realtime synth instances, and would sound exactly the same, in practice.)

Everything I have ever read in regards to mixing/mastering shows mixing down to a stereo wave @ about -6dB(on average, some people like -12dB and some like using a hotter mix) and then doing a Mastering pass over the stereo wave to do final clipping, compression etc and raising the final max level to closer to 0dB.
To me the rules for this contest indicate I can't work in this method. I have to do everything in the mixing phase. Which I honestly can't seem to get to work.

For the final stage subtleties, if you're used to working with a stereo wave: render out your mix like you describe. Load it into your daw, do a mastering pass as you would normally do - for reference, not as your actual submission! Have a device chain for subtle compression, eq, maybe a bit of volume automation, limiting and so on, like you're used to. Just don't edit the waveform itself. When you're done, toss the render and store the device chain as a preset in your daw. Load the actual realtime project again, and load that chain on the master channel (or make a dedicated "mastering bus" and route everything through it). As it's the final touch and the last processing stage of the whole project, the chain will operate on an identical signal you just processed separately, but now it's all within your main project/mix.

[rghvdberg]

So Frostline,

Are you trying to tell me life is not fair?

[Frostline]

So Frostline,

Are you trying to tell me life is not fair?

LOL, I needed that!
Maybe I am just a bit of a complainer when it comes to rules and regulations that I don't understand the principle behind.

But it is not limited to just the use of stems.

I still don't 'get' the kick drum example in regards to using samples either. The VSTi made the sound, if it inconsistent(maybe due to the CPU choking) then I would want to use the good one in a sample and move on.

Or my offence of redoing my freeze multiple times because the delay in Helm was giving me issues. But had I just used a 3rd party delay at the beginning and ignored the built in delay of the VSTi I would have been fine. I just fail to see the logic behind one being bad and one being acceptable if the resulting sound is the same. A more logical rule to me would be if the VSTi has an effect you can not use a 3rd party plugin of same type of effect. And then I can see where redoing a freeze might be more of an issue.

And I now see how much easier the whole process can be with better hardware given the rules. I switched PCs and have not had to freeze anything yet(knock on wood) and am @30+ tracks. The better PC allows me to do things that I could not do within the rules on the inferior PC of last month. So maybe that is a cost vs technique thing that has a whiff of unfairness to it? I already know how to spend money(if I had some to spend), I'm here to learn techniques to avoid spending money.

@Guenon A big thank you for the mastering technique that is rules compliant.
It is quite possible that it will not result in any better sounds from me, but I am definitely going to give your suggested method a chance.

[rghvdberg]

With osc #79 (Lokomotive) I was having trouble with the kick drum fading away. Very strange behavior. Sometimes it would sound perfect, couple of beats later it seams like some phase thingy would almost cancel the sound. Now I wanted to sample one 'perfect' kick and use that throughout the track.
Then I re-read the rules ...

Stick to the rules cause before you know it ..
Fire and brimstone coming down from the skies! Rivers and seas boiling!
Forty years of darkness! Earthquakes, volcanoes...
The dead rising from the grave!
Human sacrifice, dogs and cats living together... mass hysteria!

[rghvdberg]

One more thing:
There will never be a level playing field, and that's ok because such is life.

Some people use sophisticated DAW's like Bitwig or Live which cost a fair amount of money but enabling the user to do things faster/better than on a simpler DAW.
I for myself use Linux with Ardour or Qtractor. Not the most powerful tools but for me it's ok. I choose this because I'm unwilling to pay lots of money for software I mainly use as a hobby. There' s no return in investment.
For other people it might be too expensive (like better hardware).

Well, I can' t forbid others to use the DAW of their choice or to bend the rules to somehow make up lack of features my DAW has?

And I bet ya .. no matter what, the same people will always be at the top of the list.

[z.prime]

I don't think the DAW differences are too big. Sure, some are faster at certain tasks but mostly the speed comes with just learning a workflow that works for you. There was a month I was forced to use MuLab not too long ago and ended up winning even though it wasn't my normal DAW. It was definitely not as fun though, since i was so much slower with it. Not really knocking it... point is that the particular DAW doesn't seem to be a deciding factor.

[rghvdberg]

That's what I meant with my last sentence.
Talent will overcome all obstacles. ;l