Best process for Render entire project?

[MRT]

I had always assumed that when I rendered an entire project that it was an exact copy of what I was hearing at the Master output. Now after doing critical listening comparisons, I've realized that selected render settings have an impact on the final file.

How do I optimize my setting for a perfect 1 to 1 file of the master outs?

I tend to do track recording at 16 bit 44.1.
Renders made with Master fader zeroed and loudness meter last in effects chain.
Renders made to 16 bit 44.1

I had always thought that an offline x1 render would be better then a fast as possible but it really seems to mess with my levels?

[hopi]

what other FX are used in the project?
are there envelopes of any kind?

and what are the render settings?

[MRT]

Yes the project has a ton of Fx loaded on the tracks and a couple on the master. I take that to mean that they could be the cause of the difference in the several renders I've done? I have the use project sample rate for mixing and Fx processing selected.
Yes I have a couple of volume and pan envelopes and a couple of Fx parameters automated.

So would I be better off rendering each track to a stem and mixing from there?

I've done a ton of renders (with various settings) and the ones that are closest to the original 16/44.1 are to 16/44.1 wav, Full speed off line, use project sample rate, resample mode good 192 and write BWF chunk.

[hopi]

well hard to say but I would suspect the differences are coming from the FX or some of them...

It would interesting to try this:

Do a Save as copy of the project to a whole other folder, with copy all media box checked

Open that project and go track by track and freeze each one
that will leave you just the wav file on each track
Play it and if it sounds Ok, render that project
By freezing each trk, you would still leave and FX on the Master not part of the freezes...

[MRT]

Good Idea, Had not thought of that. Should be able to tell by process of elimination where the problems are.

I was able to get a good render by setting the render to match the project setting as closely as I could,but I'm still curious ...shouldn't a x1 offline render in theory be better than a fast as possible?

[ramses]

Originally Posted by MRT

shouldn't a x1 offline render in theory be better than a fast as possible?

No. It's all math. The computer does not care, it simply calculates.

[MRT]

Thanks for the help... I think I'm sorted out fer now.

[serr]

You would be hearing the 24 bit, 44.1k output when you listen. Truncating the render to 16 bit might account for the difference. Did you null the render against the project to confirm you were really hearing a difference?

If there is a genuine difference, that's where it would come from.
I'd recommend tracking at 24 bit and rendering your final mixes to 24 bit. The small amount of drive space you'd save by reducing to 16 bit just isn't worth losing the 24 bit precision. Especially in tracking. You can re-render your mix at 24 bit but you can never get back the original 24 bit inputs if you printed them at 16 bit.

[Philbo King]

I'm actually a little overcautious on rendering - I render to 32 bit floating point, on the advice of Bob Katz (a mastering engineer). It has some distinct advantages if any further processing is going to be done.

If mastering is done inside the project, it is probably best to render direct from the project to the destination format.
So:
- 32 bit FP for later mastering
- 16 bit 44.1K with dithering for CD authoring
- MP3 or FLAC for streaming

[MRT]

Thanks for chiming in on this. I had a project I was working on and did a series of renders (12)at different render output setting, Saved those renders to a project sub folder. I loaded all 12 in a new project(no Fx) so I could A/B them. I was freaked out by how different they were.

I had always assumed that All digital renders would be the same and only and only the format of the output would change?

Now the project that i rendered was about 15 tracks audio and midi and about 15 to twenty inset effects. So as suggested earlier that how the effects deal with the render process are most likely the source of the difference, it begs the question...

Is it best to render each track to stems before attempting to render a final mix?

[Philbo King]

I usually only render stems when the client or mastering engineer requests it.
They are good for video/movie soundtracks where the producer might want to take the music down here & there to make the dialogue more distinct, or for separate channel files to be used for surround mixing.

There are some pitfalls in rendering stems. The main one is that if you run any compression or other nonlinear effects (saturation, distortion, clippers, expanders, etc.) using other tracks as a sidechain source, or have any of those in submix bus tracks or folders, they won't act or sound the same because a stem will not have those other source tracks active during a stem render.

It's better to get a stereo mix exactly how you want it to sound in one project.

[MRT]

Thanks for all the insight guys. You given me a lot to consider..