What am I missing out on with 44 khz?

[ambitran]

I noticed in different forums that people use 48 or 96 and speak like 44 khz has much lower sound quality. Is this the case? I've been using 44khz because of CPU usage and having an older computer. Is there really that much of a difference in sound quality?

[dtaki]

Nothing, really. You lose out on some higher frequencies that most systems can't reproduce or can't be heard by the human ear theoretically. 44.1 kHz is CD quality, which means that it is much higher resolution than what you hear from YouTube/Spotify/etc. Generally speaking, if you record your master at 48 kHz or 96 kHz, it will probably end up being either dithered down to 44.1 kHz or compressed to MP3 by the time someone listens to your song. I run at 44.1 kHz to make it easier on my CPU.

It's hard enough for the average person to tell the difference between 320 kbps MP3 and 44.1 kHz CD quality unless in a quiet room and/or with good headphones; I definitely don't think people can reliably tell the difference between 44.1 kHz and higher sample rates.

[ezelkow1]

^ you just miss out on some higher frequencies. Back in the day alot of sound cards were also natively 48khz, alot of games run at 48 as does ac3 audio for dolby digital and dts (I generally run all my audio including general WDM/DirectX at 48khz), so if you were running 44.1 audio it had to be resampled down and could be taxing on hardware. These days theres almost no performance hit to do it, but also theres not a huge hit in going up to 48 either, so its a good sweet spot. But if you really want that little extra bit of performance I wouldnt worry too much about dropping down

24bit on the other hand you gain alot of headroom, and diminishing returns at 32bit, so its fairly advantageous to run at 24bit instead of 16

[Djstarski]

44.1khz is perfectly fine to work with . 24bit is also good enough for music production .

[househoppin09]

You're getting decent advice in this thread, but not quite the whole story. It's true that you lose virtually nothing by running at 44.1 if your instruments and effects implement effective anti-aliasing measures. Unfortunately, you'll find that that's often not the case. Many stock Reason devices and even some of the better REs don't have any oversampling, and even when they do you often have to crank it way up to 4x or even 8x to come anywhere close to getting the job done, which may be prohibitively taxing on your CPU. The bottom line is that, assuming you use a fairly typical and diverse range of devices in Reason, you can expect to get some aliasing at 44.1, to the degree that it will indeed often noticeably impact the sound. Depending on the style of music, this can be good or bad, and there are even FX devices that deliberately introduce aliasing, of course. My suggestion is that you work at 44.1, but export at 192 (yes, all the way up at 192, no lower). You can then open up the 192 kHz export in a program like Audacity and downsample it properly to 44.1, in such a way that you won't introduce any aliasing. Compare and contrast the resulting downsampled 44.1 kHz recording with a direct 44.1 export out of Reason and use whichever one sounds better to you. Of course, they may not sound particularly different, especially if you're not using many devices that are prone to aliasing. In any case, it's worth doing the comparison from time to time if you really want to be sure. I've made it all sound very labor-intensive here, but in practice it's just an extra minute or two of your time, certainly worth it for any serious music you may be working on.

[Oquasec]

Not sure, I like 48 better

[Carly(Poohbear)]

24bit on the other hand you gain alot of headroom, and diminishing returns at 32bit, so its fairly advantageous to run at 24bit instead of 16

Might be my miss understanding of headroom however I see that as room left between your peak level and 0dbfs, if I had a track peaking at -6 dbfs, if I output it 16 bit or 24 bit it's still -6 dbfs, I still have the same headroom. (are you talking dynamic range? )

...

A lot of my tracks have what I call sparkle stuff going on (a lot of high freqs), I do notice a difference between 441. 16-bit and 48K 24-bit, I have done blind tests on my tracks which support that, however I will also say I also have some tracks where I can't tell the difference.

It comes down to what you are doing with your tracks, my stuff is all personal, won't ever see the inside of a recording studio, if you are just going to convert it down to MP3, then 44.1K 16-bit would be fine... (BTW My portable player supports 48K 24-bit, hence why I also use that)

[Ostermilk]

I use 48 kHz all the time for one simple reason. I had an interface at one time that used to cause all sorts of stability issues @ 44,1 kHz and I've just stuck with it since then. Nothing to do with percieved sound quality or anything because in truth, I simply can't tell the difference between a file that's been produced with a 44.1 kHz signal chain throughout and one that's been produced with a 48 kHz setup.

Even if I could tell a 'difference' I probably couldn't state with any certainty which one was 'better'.

If you can that's great then just choose accordingly.

househoppin's caveat to do with certain devices 'aliasing' is a good one and something to be aware of but personally I'd normally deal with such things on an as needed basis rather than as a standard way of working.

Bit depth is important but when you're inside Reason everything is handled as 32 bit floating point anyway so the only time you have to worry about it is during export.

[bsp]

I noticed in different forums that people use 48 or 96 and speak like 44 khz has much lower sound quality. Is this the case? I've been using 44khz because of CPU usage and having an older computer. Is there really that much of a difference in sound quality?

Since no one mentioned it yet:

In case of Reason, the CV update rate is derived from the audio device sample rate.
CVs are updated every 64 sample frames. At 44.1kHz, this means that the update rate is 44100/64 = ~689Hz.
At 192kHz, the update rate is 3kHz.
This can make a considerable, audible difference when you are building modular setups / combinators. Envelope devices for example become a lot snappier.

I wish it were possible to freely configure the update rate (per combinator?).

My soundcard is usually set to 44.1kHz because the increased CPU and disk load at higher rates does not justify the minor gain in audio quality, in my honest opinion.

[ambitran]

I've been running 24 bit as well. Thanks everyone!


Sent from my iPhone using Tapatalk

[ezelkow1]

24bit on the other hand you gain alot of headroom, and diminishing returns at 32bit, so its fairly advantageous to run at 24bit instead of 16

Might be my miss understanding of headroom however I see that as room left between your peak level and 0dbfs, if I had a track peaking at -6 dbfs, if I output it 16 bit or 24 bit it's still -6 dbfs, I still have the same headroom. (are you talking dynamic range? )

Yea just meant it gives you more a floor/ceiling to work with, higher dynamic range. Meant headroom in the sense that you can start to pump up sections with more leeway before clipping or less worry about soft parts hitting the floor

[EdGrip]

My understanding was that Reason operates at 32-bit floating point internally, so you don't have to worry about headroom. So the 24 bit only applies to your interface settings, riiight? And your final export.

[Oquasec]

Reason does 64bitdepth, which is why I'm like wtf propellerhead.

[Data_Shrine]

I always run it at 82khz, 24-bit for audio. It's possible to hear the difference. I guess its subtle maybe at first.. Less aliasing, and generally better sounding when downsampled. I just want the best audo-quality possible for my tracks.

And Reason really needs an oversampling switch on many stock devices, especially the MClass & Subtractor. Even 2x would help.

[Data_Shrine]

(...) export at 192 (yes, all the way up at 192, no lower). You can then open up the 192 kHz export in a program like Audacity and downsample it properly to 44.1, in such a way that you won't introduce any aliasing. Compare and contrast the resulting downsampled 44.1 kHz recording with a direct 44.1 export out of Reason and use whichever one sounds better to you. Of course, they may not sound particularly different, especially if you're not using many devices that are prone to aliasing. In any case, it's worth doing the comparison from time to time if you really want to be sure. I've made it all sound very labor-intensive here, but in practice it's just an extra minute or two of your time, certainly worth it for any serious music you may be working on.

I'm curious. In the past years, I have always rewired Reason, so my downsampled file were exported from Ableton (which does the processing in Rewire mode). Are you saying that Reason doesn't downsample correctly on export ??

[househoppin09]

Reason doesn't really downsample at all on export, per se. My understanding is (and someone please do correct me if I'm mistaken here!)-- when you export at your chosen export sample rate, it's just as though you had gone into Reason's audio preferences and set the active sample rate to whatever export sample rate you chose, then pressed play. The only real difference should be that, of course, during export you don't have to worry about whether your system has the juice to keep up with everything in real time. So if you export from Reason at 44.1, you get a 44.1 audio file, but it's not actually downsampled from something higher. Reason is just doing everything at 44.1 in the first place, so downsampling doesn't even come into the picture to begin with. And, of course, that means that each device in the rack will also be doing all of its work at 44.1, unless it has an oversampling feature. So when you export at 44.1, there's no higher-sample-rate version to downsample from, because it's never getting generated in the first place at any point during the whole process, except temporarily within the internals of the devices that do provide oversampling. Does that make sense?

As far as your use of Reason as a ReWire slave within Ableton, I have no relevant experience there so I'm not sure how that works, but if I had to take a stab at it I'd guess that Ableton works much the same way as Reason in this regard. The question you need to be asking is, does Ableton provide any sort of global, DAW-wide oversampling? I'm not sure, but I rather doubt it. So, if you're exporting a 44.1 file from Ableton, you're probably going to have the same issues that you would when you export a 44.1 file directly from Reason: the higher-sample-rate content will never get generated in the first place, and therefore aliasing artifacts of varying severity will sometimes be generated within various devices of any type (stock Reason devices, Rack Extensions, stock Ableton devices, VSTs, whatever) that don't provide oversampling. All of that can generally be avoided quite nicely (with caveats, see below) by simply exporting at 192, if you're inclined to put in a minute or two of extra effort to do so.

Think of it this way: when you export at 192, it's as though you've found a "global oversampling" control that affects your entire DAW at once, and you've turned it up to 4x (relative to 48) or ~4.35x (relative to 44.1). And on top of that you get the added bonus that, unlike device-specific oversampling, you don't even need to worry about the effect on your CPU, because at export things don't need to be done in real time. Then, the last step is to downsample the resulting 192 kHz audio file to 44.1 (or 48, or whatever you want). When you do that "properly" in a program like Audacity, sophisticated filtering will be applied to avoid the generation of aliasing artifacts in the conversion process, and you then end up with audio at your desired lower sample rate, but none of the aliasing that you'd have gotten if you had generated the audio at that lower sample rate in the first place. Devices that have built-in oversampling simply carry out the entire process that I just described, but they do it transparently and on-the-fly, without you having to do anything other than twiddle the oversampling knob. It's more or less the same basic process, and of course you pay for it in CPU strain just as you'd imagine. So, just think of exporting at 192 and then manually downsampling as being a sort of "hack" to add global 4x or 4.35x oversampling to your entire DAW. Of course, this assumes that all of the devices you're using work properly at 192! Some unfortunately don't--the U-He Uhbik-A reverb Rack Extension comes to mind. I don't have it, but people consistently seem to say that it doesn't function as expected at sample rates above 44.1, and I think there are a few other REs that behave similarly. (There's always got to be some extra annoying complication that throws a spanner in the works just when you think you've got it all figured out, right? )

If anything is still unclear, just ask and I'll try to clarify further. And if anyone has spotted any mistakes in what I've said, please correct away, of course!

[Data_Shrine]

So if you export from Reason at 44.1, you get a 44.1 audio file, but it's not actually downsampled from something higher. Reason is just doing everything at 44.1 in the first place, so downsampling doesn't even come into the picture to begin with. And, of course, that means that each device in the rack will also be doing all of its work at 44.1, unless it has an oversampling feature. So when you export at 44.1, there's no higher-sample-rate version to downsample from, because it's never getting generated in the first place at any point during the whole process, except temporarily within the internals of the devices that do provide oversampling. Does that make sense?

As far as your use of Reason as a ReWire slave within Ableton, I have no relevant experience there so I'm not sure how that works, but if I had to take a stab at it I'd guess that Ableton works much the same way as Reason in this regard.

It does indeed seems like Reason does not downsample as you mentionned. I've looked into the help file and there is no mention of downsampling when exporting audio.

Ableton on the other hand, does downsample the file. I don't know which website to use to upload a screenshot, so I'll just describe it. In the Ableton export window, if you run the project at 88,2khz in the preferences, and choose "44,1k" for export, there is an info text that will state : "Project will be rendered at 88.2k and downsampled to 44.1khz"

I'm disappointed that Reason does not offer this, and now I understand why I feel my songs don't sound as smooth when exported from Reason itself. But oh well it's good to know, so now I'll just downsample the file in Ableton. I'll keep mastering in Ableton I suppose !

[Rason]

There is a hypothesis, that higher sample rates actually do make a difference (there must be high sample rate used both at recording or the in plugin/DAW to begin with and at export as well). Humans do not hear the stuff above 20, however, the high tones interfere with the audible ones and change them..somehow. It hasnt been stated if it makes the sound better or worse. But anyway the difference is so slight that the "audiophile" must have a high end listening environment and then still the blind- fold test is not clearly proving anything. Some of the hi-fi manufacturers have recently started to advertise HD audio in high sample rates, but that really seems to be just marketing move to attract people who dont know that much about it.

[househoppin09]

So if you export from Reason at 44.1, you get a 44.1 audio file, but it's not actually downsampled from something higher. Reason is just doing everything at 44.1 in the first place, so downsampling doesn't even come into the picture to begin with. And, of course, that means that each device in the rack will also be doing all of its work at 44.1, unless it has an oversampling feature. So when you export at 44.1, there's no higher-sample-rate version to downsample from, because it's never getting generated in the first place at any point during the whole process, except temporarily within the internals of the devices that do provide oversampling. Does that make sense?

As far as your use of Reason as a ReWire slave within Ableton, I have no relevant experience there so I'm not sure how that works, but if I had to take a stab at it I'd guess that Ableton works much the same way as Reason in this regard.

It does indeed seems like Reason does not downsample as you mentionned. I've looked into the help file and there is no mention of downsampling when exporting audio.

Ableton on the other hand, does downsample the file. I don't know which website to use to upload a screenshot, so I'll just describe it. In the Ableton export window, if you run the project at 88,2khz in the preferences, and choose "44,1k" for export, there is an info text that will state : "Project will be rendered at 88.2k and downsampled to 44.1khz"

I'm disappointed that Reason does not offer this, and now I understand why I feel my songs don't sound as smooth when exported from Reason itself. But oh well it's good to know, so now I'll just downsample the file in Ableton. I'll keep mastering in Ableton I suppose !

Ah, interesting. Shows you how much I know about Ableton!

So, that's a nifty feature that Ableton offers, but all it's really doing is saving you the minute or so of time it would take you to downsample the resulting audio file in another program like Audacity. This can just be chalked up as yet another nice audio-processing feature that Reason lacks, though in this case the workaround is easy enough. As you've found, a direct 44.1 export from Reason doesn't give you the same results as an 88.2->44.1 export from Ableton, which makes sense. If you ever do want to export from Reason and get the same results that Ableton is delivering, you'll need to tell Reason to export an 88.2 audio file, then open that file in any good audio editor and downsample it to 44.1 there.

It's hard to say whether the added "smoothness" you think you're hearing is real, as that's exactly the kind of perception that often goes away in blind A/B tests. However, it's certainly plausible that you may be hearing the lessening of aliasing artifacts in your Ableton exports, that results from rendering the project at 88.2, which is equivalent to applying 2x global oversampling across the entire DAW. In that case, you definitely might want to try exporting at 192 like I suggested upthread, as the "smoothness" that's appealing to you will be likely to increase even further.

[Kategra]

It also affects the stock reason delay, DDL1. I discovered this when automating the number of steps:


The difference was obvious when I exported at a different sample rate then the one I produced the song.
I also noticed that as I went from 96kHz to 44.1kHz, the high frequency response changed, being brighter at 44.1Khz. I did not know at the time that I was hearing the extra aliasing distortions of Reason devices.

[selig]

So if you export from Reason at 44.1, you get a 44.1 audio file, but it's not actually downsampled from something higher. Reason is just doing everything at 44.1 in the first place, so downsampling doesn't even come into the picture to begin with. And, of course, that means that each device in the rack will also be doing all of its work at 44.1, unless it has an oversampling feature. So when you export at 44.1, there's no higher-sample-rate version to downsample from, because it's never getting generated in the first place at any point during the whole process, except temporarily within the internals of the devices that do provide oversampling. Does that make sense?

As far as your use of Reason as a ReWire slave within Ableton, I have no relevant experience there so I'm not sure how that works, but if I had to take a stab at it I'd guess that Ableton works much the same way as Reason in this regard.

It does indeed seems like Reason does not downsample as you mentionned. I've looked into the help file and there is no mention of downsampling when exporting audio.

Ableton on the other hand, does downsample the file. I don't know which website to use to upload a screenshot, so I'll just describe it. In the Ableton export window, if you run the project at 88,2khz in the preferences, and choose "44,1k" for export, there is an info text that will state : "Project will be rendered at 88.2k and downsampled to 44.1khz"

I'm disappointed that Reason does not offer this, and now I understand why I feel my songs don't sound as smooth when exported from Reason itself. But oh well it's good to know, so now I'll just downsample the file in Ableton. I'll keep mastering in Ableton I suppose !

Reason definitely DOES down-sample (and up-sample too!) - if not you'd be stuck at one sample rate!

Reason just doesn't tell you it's downsampling (for better or for worse). If you, for example, record at 96 kHz, then switch sample rate to 44.1 kHz, you are now down-sampling! IF you then export at 44.1, you are still downsampling in order to take the 96 kHz audio and export it at 44.1 kHz.

You can also just change the sample rate in the "export" dialog, leaving the playback at 96 kHz but "printing" a copy at the down sampled rate of 44.1 kHz.

Also, in comparison tests Reason's sample rate conversion stands up well against the others (didn't always compare well though, at some point they greatly improved the SRC algorithm).

[househoppin09]

Yeah, I probably should have been more clear--it's not that Reason never downsamples or upsamples at all. Obviously, since recorded audio is stored at the sample rate at which it was recorded, each audio track will itself have to be downsampled or upsampled as necessary. And, yes, the sample rate conversion used for that is quite good. (BTW, selig--do you happen to have some rough idea of when it was improved?)

What I meant, of course, was that Reason doesn't do any sample rate conversion on the entire rendered audio stream as a whole. To my knowledge, if you do what selig describes and leave the playback sample rate at 96 but export/"print" a copy at 44.1, Reason doesn't make that happen by rendering your song at 96 and then downsampling it. Instead, it just behaves exactly as though you had temporarily changed the playback sample rate to 44.1 for the duration of the export process. Of course if you have audio tracks recorded at 96, Reason will be downsampling those, but it will be doing so for each track individually in the process of rendering the overall 44.1 audio stream. Make sense?

This can be tested quite easily: simply record a melody (MIDI, not audio) on an instrument that aliases much more noticeably at 44.1 than at 96, set Reason's playback sample rate to 96, and then export at 44.1. Last I checked, the resulting audio file in that case will have all of the aliasing that's present when playing back at 44.1 within Reason, proving that the audio isn't actually being rendered at 96 and then downsampled to 44.1. If anyone wants to try this out, a good instrument to use is a SubTractor set to Init Patch with the main filter cutoff lowered to 48, playing notes above C5 or so. Switch Reason's playback sample rate back and forth between 44.1 and 96 and you should hear very obvious and audible differences in the amount of aliasing. Once you get used to what SubTractor sounds like at each sample rate, you can export under different conditions and listen to the resulting audio file to figure out what Reason must be doing internally.