How should I upsample files to mix?

You’re correct. Probably where you’d hear the difference (although subtle) would be working with various plugins at that rate. But I’d think you’d lose a lot of that as you save to any format other than lossless.

1 Like

What does it exactly mean?

@wagz have you considered 88.2 instead? It’s exactly twice of the nyquist.

I’m not necessarily trying to “improve the sound” through upsampling. It’s more like a buffer zone or loss prevention, IMO. I’m doing it because of how I plan to manipulate the audio. I want to keep everything in floating point, with the ability to manipulate files a lot and bounce them without going back and forth between 24 bit and 32 bit floating. Also, I run my N4 libraries @96khz. I think that some plugins sound better @ 96khz. I should mention that I already have a process, but I was actually wondering how others would handle this process. I haven’t really made any music in a long time. Sorry if the question is weird.

1 Like

[quote=“Arber87, post:7, topic:446”]
@wagz have you considered 88.2 instead? It’s exactly twice of the nyquist.
[/quote]Yeah. I used to record in 88.2, but I prefer 96khz on many instruments, typically.

1 Like

I personally wouldn’t bother up-sampling from if you’re given 48k, but that’s just my taste. I had an excellent prof in college that covered a lot of the sampling and bit-depth material, dude was a brain with a penchant for mastering classical music.

@holster Try this Sample Rate Conversion website for seeing the different up/down sampling logs? It’s a bit dated because PT only goes to version 10, but I looked through and it seems that DAWs don’t update that particular algorithm too often.

I think the main thing I would draw from the research is that I wouldn’t use Cubase for sample conversion? But PT stock doesn’t seem to be top notch either. I’ve been usually going with Izotope if I’m in PT, but I’ll also use the nice Pyramix SRC if I’m in mastering mode.

1 Like

Your daw will convert to 32 or 64 bit floating point as you work with the file(process), you do not need to “upsample” yourself.

Example , with the mix contest files, they were recorded at 24/48. After I downloaded them into my daw, studio one, which has 64 bit processing. I did not have to change anything to get the “benefit” from that. Whatever the processing power 32 or 64 (floating point) your daw has, you do not need to do anything about it.

1 Like
I remember i once used this to convert a bunch of mp3 samples into wav. You might as well give it a try: 

Dang! That is amazing! VERY comprehensive and a great find. Thank you!!!

1 Like

Not a problem! My mastering prof loved data (he is regional head for AES in Nashville), so we went through and compared DAWs and discovered some really nice surprises (some rude awakenings, too).

@Arber87 Switch is also on that website I posted, check it out.

Yeah. On that SRC site, you can see that Cubase doesn’t exactly handle conversion in the most transparent way.

1 Like

not many that I though would, actually do. Which was a bit surprising!

I tend to use SOX - converting through foobar2000. For a free converter, I think it is much better than many of the others.

2 Likes

FooBar SSRC looks like an incredible sample converter! I’mma have to get that one!

@Taomine i have been aware of that study since 2014. I suggested that for two reasons:

  1. It’s free
  2. It can convert a batch of files while you are sleeping at night and still a snare will be a snare in the morning.

I know it’s not the cleanest of converters, but he is converting samples which are probably taken from another record.

Sincerely,

Arber

Looking at the last couple responses…
keep in mind that this was put in the “beginners” category.

We all start somewhere. There are plenty of questions to ask along the way and you won’t get any flack for asking them here.

Let’s just keep that in mind when we’re responding to questions in this category :wink:
There is certainly room for debate on here and there is already a good mix of abilities represented (and growing), so the replies will vary.

Just know that us ‘mods’ will be fiercely protecting the beginner category as a safe place to ask questions without starting a debate :beerbang:

I am sorry, i thought OP was looking for speed instead of quality conversion. As far as free or paid conversion tools go on windows, r8brain free is the best option. You should give it a try!

1 Like

Thanks for the clarification!!

Yeah, I’ve used r8brain happily in the past.

re; bit depth, I absolutely would leave the files at 24 bit and let whichever DAW you’re using calculate your processes at whatever bit depth it operates at (unless it’s very old, it’s going to be at least 32 bit float). converting the source files isn’t going to net you anything other than a bit of wasted time and hard drive space, because 32 bit float = 24 bit depth with an extra 8 bits which slide the 24 bits up and down for very low and very high level signals. Just converting a 24 bit recording to a 32 bit float file will result in literally the same data with 8 extra 0’s at the end which do nothing!

As for the sample rate conversion, if you convert up you’ll probably have to convert back down at the end, so I’d only upsample if you were absolutely sure that the benefits are going to outweigh two rounds of conversion.

2 Likes

When you import to a project that has a specific sample rate/bit rate, if the conversion is up sampling then the audio will be unchanged, you’ve just encoded it using a different scale, but the data shouldn’t change. I don’t see what you’ve found wrong in Cubase.