Cubase 9 Hitpoint detection still broken

Every year i pay for my upgrade, every year this fundamental feature still lets me down, and SLOWS me down as i manually adjust each and every hitpoint. Why does Cubase put the hitpoint at the end of the sample instead of the start?? Am i doing something wrong? (see attached pic)… :frowning:
hitpoints c9.JPG

Hi Rexxyboy,

could you please upload or pm me this small segment of the audio file from the picture ?

Andreas

[quote=“Rexxyboy”]Every year i pay for my upgrade, every year this fundamental feature still lets me down, and SLOWS me down as i manually adjust each and every hitpoint. Why does Cubase put the hitpoint at the end of the sample instead of the start?? Am i doing something wrong? (see attached pic)… :frowning:

+1,000.

In 10+ years I have never had this work reliably. Ever.

A few years ago, Tempo Detection finally started to work correctly so I figured they’d get this sorted out as well. Not for me. If there’s a secret, I sure wanna know it.

You’re complaining so much that you’re missing the fact that a developer is here to help with the problem.

Hi Andreas I just wanna to chime in and say that you don’t need that very segment of the audio, just go to MediaBay Rock Pop Toolbox and pick up 29 Min Jupiter-Bass-DI-Verse B for example and you’ll get the same result. (at least in C 8.5)

Regards

Hi Winter Rat,

thanks for the hint.

A quick analysis showed me, that the Hitpoint Detection often sets its marker when the attack noise of the next tone appears. Or maybe it is kind of a note stop noise. Easy to hear, when you enlarge the sections manually and listen to it slice by slice. The markers move more towards the biggest transinet position when the Minimum Length parameter is increased.

Regards,
Andreas

+1
Very random detection, would love to see some improvement.

Given your fairly recent Join Date, what you may be failing to understand is -exhaustion-. Many people (such as myself) have painstakingly documented such issues for (literally) years. It’s great that Andreas shows up now. But there have been other years where a dev chimes in, but is then never heard from again. It’s like the girl who never calls you back. One does get tired of various bugs or design flaws never getting addressed. Cubase is great, but after enough time, such disappointments might make even Mother Theresa a bit stroppy.

+1

This has been a hit or miss for the most part.

When it works, it’s great! When it’s doesn’t. well… I just step away and get a coffee refill.

You’re not forced to use this DAW. Maybe it’s time to move on if it take such a toll on you.

And the “I’m here since the Atari days!” reponse doesn’t impress me. My first version was 4, but it doesn’t make anything I say more or less valid.

Hmmm… Nope. Not here anyway…

Using that same file quoted by @Winter Rat, I get a lot of ‘good’ hits at obvious note starts, but in addition, several towards the end of notes as well. Increasing the Minimum Length parameter strangely removes the hitpoints at the start of some notes, and leaves the trailing hitpoint towards the end of the previous note, intact…!

This facility can be made to work, with a good dose of manual intervention - but it hasn’t been anything to write home about, despite a couple of updates along the way (IIRC…?)

Its not hard to replicate how easily the Hitpoint algorithm misses/ignores/can’t detect, some real and visible main (high) peaks - no matter how much you adjust the Threshold, or Minimum Length parameter value.

I don’t know… are my expectations too high…? I’d hope there’s agreement from Steinberg that improvements can be found/added…

Cheers,
Bob

PS:- I tried my tests with ‘Auto Hitpoint Detection’ enabled in prefs.

Your expectations are -not- too high. I had been using Cubase for maybe 5 years before a guy showed me Beat Detective in Pro Tools. And my jaw dropped. He looked at me like I was an eejit. It was like showing native people a lightbulb—literally magical. I could not, for the life of me, understand how P/T had something so much -better- than Steinberg—given how SB had -always- been ‘first’ in various technologies. I just -assumed- everyone else was as mediocre.

And then I got a freebie copy of Ableton Live/Lite in 2011. And again… it just -worked-. Pretty much every time unless you throw something like Frank Zappa material at it.

Without sounding -too- snarky, I cannot for the life of me understand why other vendors make this look so easy and yet SB struggles so mightily. I mean… again… Cubase is (still) -so- much better than other products in so many other ways.

Strange. Seeing good hitpoint results here for that file both in C8.5 and CPro9, using the default hitpoint settings.

What are your hitpoint settings when you get this result ?

Coming to Cubase from Pro Tools, I completely agree with this. I miss it dearly from Pro Tools…unfortunately Cubase is not in that league yet as far as hit point detection. It’s very frustrating in Cubase by way of comparison. I hope they bring it up to par very soon!

Hi Andreas, of course. Thanks for checking this out. I’ve actually just recorded a video to demonstrate it on my system. Below is a link to the video as well as the file that i recorded in that video. Hoping you can shed some light on it for us.
Make sure your Cubase is set to 24bit 48k, and the tempo is 107bpm, and load bass file at Zero (hard left!).

The video is a medium bitrate to keep the size down to 50 mb, Many thanks!

Bass File (Jazz bass through Neve pre-amp)

https://dl.dropboxusercontent.com/u/1674896/Cubase/Hitpoints/BASS%20AMP_26%20107bpm.wav

Here is a video of me making that file

https://dl.dropboxusercontent.com/u/1674896/Cubase/Hitpoints/Hitpoint%20Demonstration.flv

(First post, long time user 15+ years) Looking for an update to this as well.

It is weird. It works great on some material, but I almost always have to do it manually in the end. Very time consuming (I am sitting right now and doing this on one minute of materials that has about 1/16th notes that are not exactly right on…).

In my opinion, that’s a difficult file for the hitpoint algorithm due to the clicks at the ends of the notes. But I agree, Cubase should do better with a file like this.

Thanks for the files. They really point out what you mean. I think your recorded bass file demonstrates, what I mentioned in my first comment. The noise, hard attack, click sound or whatever you call it, is interpreted as the attack for the next note. So we might need a parameter or two to adapt the algorithm for bass or similar instruments and make it ignore such kind of noises. It’s not that the hitpoint detection is not working, it’s more that the percussive component should be ignored in some cases.
It’s on our list for further development of hitpoint detection.

Regards,
Andreas

That would be great Andreas. The option for it to just choose the tallest transient would be ideal for bass (and straight up snare drums as well). Maybe that can be a slider as well (like the threshold slider).