Mix does not obey Use Selection

[Starship Krupa]

I don't have time to work up a full bug report with reproducibility, etc., but I wanted to check to see if anyone else has seen this.

The song ends with the guitar doing a slow whammy drop and lift with a delay on it. The client and I want to fade the song during the final repeats.

When I make the pre-mixdown selection, I make it out to where I can just barely hear the repeats of the delay.

The final mixed down files, however, truncate the recording around the end of the clip itself, which sounds nasty and abrupt.

I came up with an emergency workaround, which is to park a dummy clip with the volume automated all the way down out where I want the mix to actually end.

I am of course using build 292, restarted Mixcraft, restarted the entire computer, and the issue persists.

After the project is finished, I'll work out a bug report, but I wanted to toss this out to the forum now. It's something to watch for.

[the pannacotta army]

I'm not sure this is a bug. Personally, it sounds like MC is behaving in the way I'd expect it to.
I would have thought what governs the "end point" of a rendered mix is the last event in the track view, and not any decaying residual sound resulting from any VST FX such as delay repeats or reverb tail.

Putting some "dummy event" in would therefore be the correct procedure to stop the premature end.

[Acoustica Greg]

Hi,

Yes, this is working as designed. The most common support request related to this kind of issue is: "how do I make the mixed-down file end at a precise point?" For some reason, lots of people like to chop off effects tails. Not sure why.

Greg

[Starship Krupa]

Yes, this is working as designed. The most common support request related to this kind of issue is: "how do I make the mixed-down file end at a precise point?" For some reason, lots of people like to chop off effects tails. Not sure why.

Not chopping off the effect tails is indeed what I was trying to accomplish.

So to clarify: no matter how far out to the right I make the selection, when mixing down, Mixcraft will chop it at the end of the last clip.

Thereby cutting off any reverb or delay tails.

[Juno]

When I make the pre-mixdown selection (...)

I probably don't understand the reason for this part, so maybe the following makes no sense. But i'd mix down the entire song to a wav without fades, import that wav and only then i'd do a fade out.

[Acoustica Greg]

Hi,

I tried it with Fusion Field and a short recording I made in Mixcraft. I see what you're saying! Did you file a bug report on it?

It looks like this is the way that Mixcraft 6 worked as well, but this is the first time I've heard anybody mention it, while dozens and dozens of people have asked how to just truncate the sound at a specific time.

Greg

[Mark Bliss]

It looks like this is the way that Mixcraft 6 worked as well,

It does and it did. I had my own (silly) logic as to why and simply used a workaround when this occurred.
Well except for one unrepaired project I still have posted on soundcloud, where I think this phenomenon truncated the ending.

[Starship Krupa]

When I make the pre-mixdown selection (...)

I probably don't understand the reason for this part, so maybe the following makes no sense. But i'd mix down the entire song to a wav without fades, import that wav and only then i'd do a fade out.

It's a feature that allows us to mix down only sections of a piece if we want. For instance, I'm using it on this project to cut off the count-in at the beginning of the song.

One problem I have with your workaround is that it's yet another trip through the CODECS and other stuff inside Mixcraft. Even with the use of uncompressed file formats, I suspect that the path is not 100% bit-perfect.

[Starship Krupa]

I tried it with Fusion Field and a short recording I made in Mixcraft. I see what you're saying! Did you file a bug report on it?

It looks like this is the way that Mixcraft 6 worked as well, but this is the first time I've heard anybody mention it, while dozens and dozens of people have asked how to just truncate the sound at a specific time.

I'm glad I managed to communicate what was going on.

I did not file a bug report because you and pannacotta said that it was supposed to work that way. I will now, though.

IMO, the way "use selection" should work is that the mixed down file should contain whatever is going on in the mix, be it tails or silence or whatever, within the selection points, and nothing outside the selection points.

[Acoustica Greg]

I concur.

[Juno]

It's a feature that allows us to mix down only sections of a piece if we want. For instance, I'm using it on this project to cut off the count-in at the beginning of the song.

One problem I have with your workaround is that it's yet another trip through the CODECS and other stuff inside Mixcraft. Even with the use of uncompressed file formats, I suspect that the path is not 100% bit-perfect.

For some reason i completely misunderstood the "pre-mixdown selection", yesterday. Now, i'm like: "how can i think that is not meant exactly like that?"

Thanks for explaining. My workaround makes more sense in slightly different circumstances; i've always figured it shouldn't matter in terms of bit perfection (since in the end, it's the same number of edits), but it's an interesting notion. Out of curiosity: is it just a suspection or can it really happen?

[Starship Krupa]

Thanks for explaining. My workaround makes more sense in slightly different circumstances; i've always figured it shouldn't matter in terms of bit perfection (since in the end, it's the same number of edits), but it's an interesting notion. Out of curiosity: is it just a suspection or can it really happen?

It's just a suspicion. Kind of a best practice thing with me: the fewer imports, conversions, exports, whatever, the better.

In music listening, I am a stickler for lossless formats and bit-perfect playback. Mixcraft 7.5's new ability to record and mix down using FLAC was a very positive development.

To my understanding, whatever one records using a DAW, at mixdown time, it's up to the program to combine everything and render it as a single file. The method and quality and accuracy of that process is up to the engine in the specific DAW, which is why it is possible for the same WAV files and collection of plug-ins to sound different when rendered on Cubase vs. Pro Tools vs. FL Studio, etc.

It would be an interesting thing to do some A-B listening tests of a similar project rendered in various DAW's. Reviewers never mention the quality of sound after mixdown; I would like them to have benchmark collections of tracks in WAV format, that they could then import and mix down using the same level and pan settings, with the same freeware VST's.

Might be an eye or ear-opener.

[Juno]

Found an interesting piece about it. Have a look:

https://www.image-line.com/support/FLHe ... _audio.htm

[Mark Bliss]

That article should be required reading for every user as they begin to learn about DAW production. Dead on.
And contained within are some good links to further great information on the subject. Like that video at the very end, (which is also posted in the Tips and Tricks forum here somewhere on some discussion on this topic.)

That said, I can understand Eriks point too. I personally have for example shunned using MP3 formats altogether, because A: I haven't needed to compress audio files for space in my uses, and B: I have a basic understanding of the math and what is occurring, and reject its use for similar reasons to what Erik describes. Even though I know full well that at sufficient rates, I personally could never hear the difference. I just refuse to do it. Its like some ridiculous moral stand and I know it.

And while the article is about digital audio quality in general, and more specifically debunking any DAW A vs DAW B silliness, Eriks point about repeated rendering and its possible effects is slightly different. And though I am certainly not an expert on the topic nor make any claim to having enough knowledge on it to be authoritative, I would venture that its another case of theory vs reality. As in there is grounds to argue that a change is made, however, in reality you would probably have to render a sound and reimport it perhaps a thousand times (or more) to begin to be able to actually discern the change.

So my suggestion to Erik is to keep being vigilant, and stick to your efforts to minimize the changes. Better safe than sloppy. Even if it doesn't really matter one little bit. (Pun intended.)

[Juno]

Eriks point about repeated rendering and its possible effects is slightly different.

Absolutely. I didn't post that article as some sort of counterargument, but rather as an interesting view that touches on the subject. In fact, i think that minimizing the number of renders is good practise.