I have two issues with the Cubase 8 algorithm

I’m not sure if this just happens to me, but I think the algorithm adds an unwanted artifact to the samples.

This is the original drum loop. The kick and snare sound clean at the original speed. ↓

https://soundcloud.com/van-papeer/drum-loop-original/s-AvtaG


But when I shorten it with the elastique Pro - Time (or anything else for that matter), it adds a tonal “thwak” to the sample, and it just sounds bad ↓

https://soundcloud.com/van-papeer/drum-loop-shortened-with-elastique-pro-time/s-gqDls


My second issue is that when my algorithm is on the elastique Pro - Time mode on a selected sample, the transpose won’t do anything to the pitch unless I’m in the Pitch mode.

Is this a Cubase 8 issue?

And for the second issue, I believe that is how it is supposed to function so that you can change the speed without affecting the pitch and vice versa.

Well, I went over to an acquaintance of mine who also owns Cubase 8 Pro.
The pitch changed without having to change the algorithm at all. I selected the wav file, moved the mouse wheel to correct the pitch on the ‘transpose’ section and it worked despite it was on ‘time mode’.

Also, I heard that elastique Pro is not a native Cubase algorithm, but a Pro Tools algorithm and was outsourced.
Which means I shouldn’t get any artifacts like the “thwak” you heard in the audio files I uploaded above.

So, what’s going on really?

Cubase paid for the elastique algorithm to have it in their system. It’s identical to the Pro Tools one.
It doesn’t make any sense.
Even FL Studios has a better result with time stretching.

Cubase 8.5 is currently using one the most advanced time- and pitch-shifting algorithms out there: zplane’s Elastique Pro v3. Same is true for Samplitude, Reaper, Studio One, Fl Studio and others.
Ableton Live 9 still uses Elastique Pro v2. So in terms of “stretching” Cubase has by far less artifacts when shrinking / stretching. So I don’t understand Headlands comment at all.
However Ableton Live 9 default algorithm is beats, what actually would be “slicing” the file (hit points) and quantize it in Cubase (cp. rex file). In this case there aren’t any tonal changes of course.

That said “stretching” always generates some sort of artifacts in every software. This is actually normal!

Anyway, the following solutions don’t generate that sort of artifacts:
a) You could use slice at hit points and then change the tempo
b) You could use the Elastique Pro Tape algorithm in Cubase for a more classic turntable like speed up or slow down (no artifacts, but like a turntable control pitching is happening)

Heard elastique was one of the best algorithms out there, so Headlands comment didn’t make sense to me either.


Anyway, while I fully understand there are “some sort of artifact” after manipulation and are totally normal, you don’t hear that tonal “thwak” on FL Studio’s algorithms, or even on Logic’s.

I tried every single algorithm for time stretching and it gave me the exact same result.
Although, when I tried it on an acquaintance’s computer, it worked just fine. No “thwak” artifact that you hear from my uploaded samples on soundcloud.
I was beginning to think if this was a bug of some sort.

It still does it. Is there a different setting for the time stretches that I’m not aware about?

I am pretty sure you don’t get these artifacts with the slicing at hit points method, neither with the Tape algorithm.
From which tempo to which tempo are you stretching / shrinking this loop? Maybe you could upload this loop here or provide a download in some sort so we can check.

FL Studios, Logic, Ableton or any other DAW for that matter don’t require you to use the slicing at hit points method.

I don’t understand why I would have to resort to that method when it works perfectly on other DAWs, and on the same Cubase 8 Pro software on a different computer.

FYI, the tempo makes no difference IMO because I’ve only reduced the sample loop by less than 10%. It would still add that tonal thwack no matter the percentage.

This is a legitimate issue I’m having and it seems like no real Cubase engineers are paying attention to it…

As I said you have to differentiate between “slicing” and “time-stretching” (in its classical sense).
Cubase doesn’t slice per se, just stretches the audio material. Stretching always generates artifacts (depending on the algorithm quality). But as soon as you start using Elastique Pro in its standard equivalent setting there is no difference between DAWs using it (actually the stretching usually null).

However be aware that Elastique Pro v3 was implemented in Cubase 8.5 first and not Cubase 8. If you are comparing stretching between 8 and 8.5 there are audible differences.

Logic incorporates “slicing” in it’s auto flex time mode. So if you are flexing a drum loop in Logic it usually chooses automatically the “slicing” algorithm (it’s actually called “Slicing”). As said in Ableton Live the standard mode is “Beats” which is “slicing” (and not stretching).

In other words you can’t compare stretching to slicing quality wise in view of drums. As Cubase doesn’t use “slicing” in it’s musical mode (automatically stretching or just changing the loop with the time stretching arrow tool) you have to watch out what you compare it too.

If you want slicing like in Logic or Ableton Live you “have to” use the slicing at hit points method. You can easily create a shortcut for it or even create a macro for your special needs.
This is not an issue, it is more of handling error then :wink:.

However if you still have different outcomes between different computers with the same Cubase version - there must be some error of course - but obviously in your settings then.

I am happy to test it with your drum loop on my system. I am still confident that you can at least get the same (usually better) results with Cubase than with Logic or Ableton Live. You could activate the download option and soundcloud or deactivate hidden mode.

Hi vanpapeer,

Please forgive me for my honesty in this long reply but you’re stressing over nothing and I’ll tell you why and will tell you how to resolve your issue!

I’ve listened very carefully to your two samples and it may shock you to know that a basic 1 bar drum loop does not make a hit record.
Once you’ve produced, mixed and mastered the full production I can assure you that no one in their right mind will hear a ‘tonal thwack’ on the drums and if it’s an incredible song I can also assure you it won’t be rejected by radio stations, etc because they are not listening out for a tonal thwack but how the entire song blows them away!

I’m old skool and have been a very successful professional keyboards player/record producer/composer for nearly 4 decades and if I come across a production issue I find a workaround and I recommend you take these steps:

Slice up the original loop and drag it ( on to the bottom three blocks) in Groove Agent. You will then have each drum sound over the keyboard so you can simply play the rhythm into Cubase yourself. (It’s such a simple 1 bar phrase - a monkey could play it with his hands tied behind his back!) You can then create new and interesting rhythm variations for different sections to take the listener on a journey like all good songs should! You will also have complete control over each individual drum sound which can be enhanced in hundreds of ways.

If you can’t be bothered to do all that simply find a new (and possibly better) loop at your songs correct tempo so you don’t need to time stretch it or simply choose preset kits from Groove Agent (or any other drum machine VST’s you may have) or use individual 1 shot hit samples and construct your own mega kit!

I hope you can now see that you’re concentrating on a speck of dust rather than seeing the bigger picture?

It took me quite a few years into my career to realise that the general listener on the street doesn’t hear what we hear. I was beating myself up unnecessarily because I couldn’t get my productions to sound the way I wanted them to but it doesn’t matter because the truth is the general public are either touched by the whole picture or not and don’t forget that most people listen to a 128kbps mp3 file on laptop speakers so definitely won’t hear your tonal thwack!

You may be saying to yourself who does this guy think he is trying to convince me not to be angry with Steinberg for including an algorithm that is not perfect in every way - well forgive me for blowing my own trumpet but I’ve had 18 #1’s in 37 charts all around the world, have won many awards, was signed to Sony Records and have had my own production company since 1989 composing/performing and producing music for singers/songwriters.

Most importantly you may be interested to know that all the music to those 18 #1’s were composed, performed, produced, mixed and mastered by me in my studio using either Cubase 5 (or a lower version) as I’ve only recently upgraded to Cubase 8.5. I used Cubase 5’s inferior timestretching and pitch shifting algorithm all the time and it didn’t stop me from getting so many #1’s and my music played daily all around the world!

Sorry for the length of my reply but I hope it has given you and everyone else food for thought?

Have a nice day!

Kind regards

James Colah

http://www.twitter.com/jamescolah

Okay, first of all, a massive thanks for your sincere and detailed opinion/reply.

But unfortunately, you’re missing the point of my post here.

I think you didn’t care to read the other replies that I’ve made on my own post.

I tried time stretching on an acquaintance’s computer, on the same Cubase Pro 8, which didn’t create any weird artifacts like I mentioned. Also, the ‘transpose’ function worked without having to select ‘elastique Pro - Pitch’ mode, meaning it the pitch of the wav file was affected when I put it to say, ‘-1’ while the mode was on ‘elastique Pro - Time’.

Something that works perfectly well on someone else’s computer does not work on mine. That is the biggest issue here.

And one more thing… you mentioned that “no one in their right mind will hear a tonal thwack on the drums”, but trust me the tonal thwack bleeds through the song and it changes the feel of the song for sure. If that kind of detail doesn’t bother you… well, that’s just the way I work I guess. There are tones in snare drums, kicks and even rides for that matter. Sue me for being a perfectionist.

But seriously, I appreciate your long and honest opinions on workarounds and such :slight_smile:

Thankf for your replies. I’ve just enabled the download for the two wav files and I also uploaded a different drum loop sample as well.

If you hear the stretched version of the samples, you will definitely hear the tonal “thwack” artifact added to the result. Which is very strange and should not be there.

Original Sample 2:
https://soundcloud.com/van-papeer/original-drum-loop-2/s-CzVFv

10% stretch reduction with elastique Pro - Time:
https://soundcloud.com/van-papeer/drum-loop2-shortened-with-elastique-pro-time/s-SEeJU


Here’s the same Sample 2 reduction stretched with FL Studios.

https://soundcloud.com/van-papeer/same-sample-2_stretched-in-fl-studio/s-wh18j

No artifacts or tonal “thwack”. Clean.

Hi again vanpapeer,

Firstly thank you for your kind words.

I did read all your comments and all the replies very carefully so am not missing everything.

I feel your pain about the artefacts but you’re missing the point and bigger picture when I pointed out that when I create music professionally and come across an issue I don’t let it stop me in my tracks but find a different way of getting my end result and if that means using a different drum loop/sound then so be it.

Please understand that 50% of the general public may say the original drum loop sounds great but another 50% may say it’s crap as one mans meat is another mans poison so please don’t sweat over it or you will be stuck in a downward spiral of despair and anger! :wink:

Do you not also realise that no two Cubase computer systems will function exactly the same so am not at all surprised that you’re having this issue but your friend is not and this also tells me that you can’t blame Steinberg as the problem obviously lies within your computer configuration as I have tested the timestretching algorithm on my Cubase 8.5 and it sounds absolutely fine and I’ve been cursed with my ears as I have bats ears and can hear everything!

So once again (for your own sanity) please let it go and concentrate on producing a hit record any way you can!

Kind regards

James Colah

http://www.twitter.com/jamescolah

Well, first of all you have acknowledged that there’s a problem “within my computer configuration”. So that is the issue I’m trying to tackle here. My reasoning is that Cubase 8 Pro should not have this issue compared to Cubase 8.5 simply because of the unwanted and needless tonal artifact that the DAW adds to the result. It’s nonsensical.

And two, I will not let go of this issue until I’ve found out why it’s doing this to my samples. A “hit record” to me is about fixing these issues on my downtime and figuring out how to make cleaner samples that I want to add to my songs.

Thanks again :slight_smile:

Thanks for the download option. These files are hard to compare. Especially the original drum loop is quite smashed. In true peaks this loop is actually way over 0dB. Not a good start for stretching or shifting.
Then there are differences in loudness up to 3dB and overall this looks like other things than stretching could have been made.
Could it be that you have used another more cleaner original drum loop in FL Studio?

However I used your original quite smashed drum loop and stretched it from 85 to 94.082 (not possible to match your files to the third decimal!?) with three different methods / algorithms in Cubase Pro 8.5.
I can’t hear that amount of “thwack” in my results even with the standard Elastique Pro v3 Time algorithm.
And overall my result are different sounding.

As you are still talking about Cubase Pro 8, I assume that you have overseen my note, that Cubase Pro 8 doesn’t use Elastique v3 - Cubase Pro 8 still uses the old Elastique v2(!) which do sound a lot less clean (more artifacts).
Depending on the FL Studio version used - the newer versions also use v3(!).
In this case you would compare apples and oranges.

Please download the files as the previews of zippyshare are played at a lower resolution.

Really appreciate your help :smiley:
I checked out your samples and it’s very clean compared to my samples.

I just want to ask straight-up, can you hear the tonal thwack added on my stretched samples? Tell me I’m not going crazy :confused:

I’ve just tried using the FL Studio e2 (assuming it’s elastique v2) algorithm on the same sample, but there are no artifacts added like in Cuabse 8.

And for the smashed-sounding sample… it’s because all my stereo tracks are routed to a master bus which is a mastering chain (pretty sure the RMS is around -9). But, even without squashing the living hell out of it, the artifact (audible, tonal “thwack”) is still there.

I’m thinking the v2 should not add that weird tonal thwack… no matter how inferior it is to the v3 on the Cubase 8.5.

And if it means anything, I’m using a Focal Solo 6 Be’s for my studio monitors with a Babyface Pro audio interface.

Sorry, I actually realized that the FL Studio file is the only wav file. So this of course explains some of the differences (overshoots, losses, …) of the MP3 files.

There is definitely quite a difference. You can see it also in the wave file comparison (attached) that your file has notably smeared / stretch transients which lead to more tonal content of course.

I would recommend upgrading to 8.5 :smiley: then or even trying the old Standard algorithms in Cubase (Standard > Standard - Drums) maybe you get better results.

Great find on the shape of the wave. Thanks for the analysis. Much appreciated.

But… I mean, I’ve tried this on my acquaintance’s computer like I mentioned above and he had no problems at all.
So I was thinking maybe there is a way to fix this on my DAW as well… :cry:

I guess the only real way to fix this is to upgrade to 8.5.

Please, legit Cubase engineers, help me out.

Am I the only one thinking this is a serious issue?

The algorithm adds unwanted artifacts to the result… I’d love to hear something from some legit Cubase engineers.

So… still nothing?