Question about 44.1 sample libraries and 48 KHz DP sessions

[FMiguelez]

This is almost embarrassing to ask because it seems such a basic question, but I'm wondering...

If all my sample libraries were recorded at 44.1 KHz, and I do my DP sessions at 48, WHO makes the sample rate conversion, and WHEN does it happen?

I ask because I seem to notice an increase in CPU usage when setting my template like that, but I don't know if I'm being totally unbiased and objective...

I use VE Pro for the slaves. So DP, Track 16 and my Tascam mixer are all set to 48. Then the signal travels from the slaves to the master as streams of 1s and 0s representing a 44.1 sound. Then what??? Is DP making the conversion in real time, thus my increased CPU usage?

What about the quality? I haven't noticed anything peculiar yet, but I'm just at the beginning stages of a new project, so I don't know if there will be cumulative artefacts due to the sample rate conversion in EACH track???

This makes me wonder, then, WHAT's the point of working at 48???

Even if 99.99% of what I do is for broadcast, isn't it simpler and better to work at the native sampling rate of the sample libraries and just do the upsampling to 48 at the mastering stage (or letting the post houses deal with it), as I've been doing for years?

Or am I misunderstanding everything and worrying about a straw man?

Thank you for any insight.

[billf]

Here's one opinion on the topic:

If you work purely with your own recorded material, then the choice is again quite easy: go with 48000 for everything your record, and stay in the 48000 world all the way through to your final stereo mix. If your destination format is video-based, you're good to go, and if your destination format is music-based (CDs, MP3s, etc.) then you can always dither down to 44100 with a good audio-editing program (or let your mastering house take care of it).

If, however, you plan to incorporate a lot of purchased (or downloaded) samples and loops into your project, then you will very much be constrained by the sample rate used in your loop libraries.

Here's the essential problem: if your project contains mixed audio files of differing sample rates (some are 48000, some are 44100, some are 32000, some are 96000, etc.), then your sequencing software must perform on-the-fly sample rate conversion of some of those audio files to ensure that all the internal material being processed matches the sample rate defined for your environment/project.

And the sad fact is that all DAW software has generally crappy built-in sample rate conversion. All popular DAWs are in the same boat—they all suck at this. To perform really accurate sample rate conversion in real-time would eat way too many system resources and would limit the number of tracks/clips/loops/samples you could work with without bogging down your computer.

http://karma-lab.wikidot.com/misc:choos ... -frequency

[monkey man]

Dang.

[mikehalloran]

Is anyone using Izotope RX for converting sample libraries?

[billf]

Is anyone using Izotope RX for converting sample libraries?

I use it to do conversions, but only on an as needed basis.

[FMiguelez]

I was half-hoping someone would tell me this wasn't an issue and more of a misunderstanding on my part, but according to your answers, it is!

So, if I have say, 30 connections via VE Pro to an orchestral setup, and VSL library is at 44.1, then DP is doing 30 sampling rate conversions on the fly????

If that's correct, then I wonder what's the "least worst" approach... Stay at 44.1 and upsampleto 48 in the mastering stage, or have a little audio degradation (due to the on-the-fly conversions) and wasted CPU, and stay at 48 all the way...

According to billf's quote, I should do the former (produce at 44.1 and upsample the masters afterwards).

Do you all agree? Is that the consensus then?

I know some people, like MLC, who work all the way at 48 and use lots of sample libraries that are at 44.1. I'd be interested in what they have to say about this...

So then, this also means that working at 96 or above, as many people do, is meaningless and a waste of space and CPU if one is mostly working with sample instrument libraries... Those higher rates only mean something when the material us recorded at that rate from the start, correct?

[mikehalloran]

The perceived differences are far greater between 44.1 and 48 than between 48k and higher. This is due to the effects of the anti-aliasing filters that still operate in the audible range at 44.1. The effect tends to be most obvious on the high hat and with cymbals sounding a touch less real.

That it's a lot less obvious nowadays over 30 years later doesn't mean that some can't hear the differences still.

The best rule of thumb to my mind comes from George Martin who wrote, "All You Need Is Ears".

If you have to deliver in 48k and upsampling on the fly sounds good then you might want to let it be.

Otherwise,

billf » Wed May 27, 2015 10:21 am

mikehalloran wrote:
Is anyone using Izotope RX for converting sample libraries?


I use it to do conversions, but only on an as needed basis.

[MIDI Life Crisis]

I
I know some people, like MLC, who work all the way at 48 and use lots of sample libraries that are at 44.1. I'd be interested in what they have to say about this...

I never give it second thought. The sessions are at 48 and the libraries load and sound fine to me.

[Shooshie]

And the sad fact is that all DAW software has generally crappy built-in sample rate conversion. All popular DAWs are in the same boat—they all suck at this. To perform really accurate sample rate conversion in real-time would eat way too many system resources and would limit the number of tracks/clips/loops/samples you could work with without bogging down your computer.

I agree with most of what this post said, but... this takes it a little far. Most DAWs don't 'suck' at sample rate conversion. About 5 years ago there was a comparison chart showing the differences in anti-aliasing of audio at the Nyquist frequencies, and granted, some came out far worse than others, but DP was right up near the top. We're talking theoretical when we're up there, anyway, because nobody knows FOR SURE what the effects are of audio beyond our hearing range, though if your pets start looking quizzically at your monitors when you're mixing, you can bet that the anti-aliasing is going badly. There is a difference, for sure, and sometimes I think I'm aware of it, but I can't actually hear it. (I CAN hear a difference between certain audio recorded at 44.1K and 48K. Beyond that, it's mostly a feeling.) Anyway, DAWs do a pretty good job of sample-rate conversion, as has been proven again and again with flipped polarity/cancellation tests many of us have performed and reported on here.

If you're running a global positioning satellite, where even the effects of relativity must be calculated for it to be of any use, THEN you might want to use something better than DP for sample rate conversion, but for music audio, DP's output is hardly "crappy."

Just had to say that; I think we let possibilities vs. reality get out of proportion and without re-calibrating our perspective, conversation becomes troubled with meaningless superlatives.

Shooshie

[Shooshie]

So then, this also means that working at 96 or above, as many people do, is meaningless and a waste of space and CPU if one is mostly working with sample instrument libraries... Those higher rates only mean something when the material was recorded at that rate from the start, correct?

Yes and no. Practically speaking, the material recorded at 44.1k cannot be made to sound better by upsampling it to 96K. But the interactions between those instruments, the harmonics produced when they play together, may get some breathing room at 96K. But I'm one who makes most of my projects at 48K, which IMO is enough for the breathing room and everything else, lifting the Nyquist frequency interaction far enough about the hearing threshold as to be inconsequential in the kinds of things we do.

Shooshie

[David Polich]

Converting sample library data from 44.1 to 48khz is out of the question. That would take
weeks.

Personally, I'd go with staying at 44.1khz. The difference between that and 48khz is not even an
octave. Upsample your mastered audio stems to 48khz if you must.

I've always stayed at 24bit/44.1khz. For everything. I don't deliver audio for video, true. But the
48khz standard for audio for video is arbitrary and not based on any major difference in sound
"quality".

[monkey man]

Call me a luddite, but I needed to hear that, David. Thank you. I suspect Fernando will welcome the sentiment as well.

It's been 24/44.1 since day dot for me too, and seeing as the AVB system I'll be upgrading to will be USB2 based, the minimum input-channel count I can live with dictates that I can't go past 48kHz anyway. So it's 24/44.1 for the foreseeable future too.

FWIMBW, I always figured that less SR conversion would be required overall, all things considered (like sample libraries, MP3 and CD formats etc.), if one stuck with 24/44.1.

All this said, I can't hear a peep above, oh I don't know now as it gets worse every year, somewhere between 8 and 10kHz? It therefore matters not to me. I leave what I can't hear alone in the mix, figuring that if I don't touch it, I can't stuff it up. It's gonna sound more natural the less it's messed with anyway, and the mastering guy can give it a shelving lift if need be.

Just my 2 bob's worth. Bottom line, 'Nandito? Don't worry 'bout it, brah!

[FMiguelez]

Call me a luddite, but I needed to hear that, David. Thank you. I suspect Fernando will welcome the sentiment as well.

Oh, I totally do.

After this and some further reading, I think I can make an informed decision and just forget about this once and for all.

So after all this, I am convinced that, for my situation and the way I work and what I do, my best option is to stay at 24/44.1 and upsample to 48 at mastering, or simply let the production houses deal with it, like I've been doing for years.

Thank you everyone who participated!

[Shooshie]

I've always stayed at 24bit/44.1khz. For everything. I don't deliver audio for video, true. But the 48khz standard for audio for video is arbitrary and not based on any major difference in sound "quality".

I know we've been through it before, but I have to say it again. There is an audible difference between 44.1k and 48k. The difference is more pronounced (to me) than the difference between any other pair of sample rates, though the same difference can be heard when comparing 44.1K to any higher rate. In other words, I don't hear a similar difference between 48K and 96K, but I do hear the difference between 44.1k and 96K.

Yes, it's just an octave difference, but apparently it's THE octave they should have kept in the standard. 48K should have been the CD standard.

I know this how? I can only give personal anecdotes, but to me they were definitive. In the year 2000, I was working in Las Vegas, recording a piano album. The artist and I experimented with microphones, mic placement, room treatments, and such, and one day we thought we'd found the sound we were looking for. He had to go to work, playing at the Bellagio Hotel, but when he came home I played some things I recorded. He said "what happened to the sound?" I was kind of sheepish about it, because I'd been thinking the same thing. But I thought I was just tired of hearing the same thing repeated all day. When he said it, I reviewed everything in my mind and remembered that I had switched to 44.1K from 48K, thinking it would save disk space (80 MB was a big hard drive, then), and also thinking that you could not hear that high, anyway, so it wouldn't make any difference in the sound.

I switched back to 48K and recorded a few minutes. Then we compared the two. While we could not do A/B testing, because of the time it took to reset the sample rate, we didn't really need to. The difference was obvious. While neither of us could finger the exact thing we were hearing that was different, the 48K just had a spacious feel. The sounds were luxuriant. It was more like having your head in the piano. The other was more harsh. Brash. Listening through paper walls or something. Any words I use to describe it will sound exaggerated, as it was a subtle thing, but noticeable.

I've recorded projects at 44.1K without realizing it, so I know it's not something that always jumps out at you, but certain kinds of projects — the piano project for example — seem to benefit more from 48K.

All I can tell you is what we experienced. Granted, this artist had amazing ears, and the same could probably be said for me, 15 years ago. But if we could hear it, I'm sure most people could hear it.

Shooshie

[bayswater]

I hadn't thought about this. I did some stuff at 48 a while ago and thought it sounded better. I've been changing the templates and settings to 48 and hadn't considered rate conversion overhead. And now I think about it, tracks from VIs recorded at 44.1 can't sound better -- same at best and possibly worse.