What's new

A rant on the hiss that's in so many orchestral libraries.

Nick Batzdorf

Moderator
Moderator
It's also worth mentioning that room tone doesn't suddenly disappear in digital black like it does in case of a "roomy" sample (... which is why noise sticks out like a sore thumb). :)
Dietz, could you possibly explain that in different words? I don't quite follow. Are you just saying that room tone keeps going after the sample stops ringing?

In any case, I can honestly say that I haven't been bothered by samples since I used a Sequential Prophet 2002 (128K of memory, sounds loaded from one side of a floppy disc) - and then the studio's 2" tape noise drowned it out. :)
 

robgb

I was young once
I have not read all posts in this thread, so please excuse me if someone else has already suggested it.

When I use a problematic library, I bounce every midi-track into an audio track. Then I use a plugin like this in every single audio track

It's time consuming, but because every single audio track gets his own best noice reduction preset, the end result of the mix is light years better than if I did not do it all.
I recently bought Waves W43 when it was on sale (may still be) and it works like a charm on noise. Uses a different method than most of the other noise apps I've used, and I have yet to get any of the warbling effect you get with those other apps.
 

Dietz

Space Explorer
Dietz, could you possibly explain that in different words? I don't quite follow. Are you just saying that room tone keeps going after the sample stops ringing? [...].
Oh my - sorry, English isn't my first language. 8-)

Let me rephrase that: In Real Life, the room tone never stops*). A sample, OTOH, reaches its end at a certain point, and any room tone (and/or hiss) will fade away into total silence (or worse: it stops abruptly). This makes it even more noticeable, in addition to the inherent issue of cumulated noise per voice.

*) ... that's why there is a several minutes long loop of the original room tone available (in Ambisonics!) for every MIR Pro Venue. :)
 

Dietz

Space Explorer
And that’s why we add our own special-sauce reverb to library samples, even if there’s already space/room in from the original recording.
That's only fair! - We, OTOH, denoise our samples, so we don't have to drown them in reverb just to hide their shortcomings. 8-)
 
  • Like
Reactions: bun

Uncle Jesse

Member
I agree with you.

The Garritan CFX has terrible noise/hiss on very low velocity samples especially in the lower registers.. It really builds up when playing multiple voices at the lowest velocities and almost dominates the sound of the piano. It is driving me crazy at the moment.
 

Uncle Jesse

Member
I have not read all posts in this thread, so please excuse me if someone else has already suggested it.

When I use a problematic library, I bounce every midi-track into an audio track. Then I use a plugin like this in every single audio track:
It's time consuming, but because every single audio track gets his own best noice reduction preset, the end result of the mix is light years better than if I did not do it all.
Hey, how do you go about using x-noise if you don't have any isolated noise in the sample to get a noise profile for the learn function? The samples I'm having hiss issues with only have hiss present while the instrument is playing, so it's not possible to get an isolated noise profile.
 

Uncle Jesse

Member
I've even considered using Izotope RX to reduce the hiss on the actual samples for these libraries, but that's a huge project. There is no reason to be stuck with the bad parts of analog in these modern times -- you can record with great pre-amps and mics and not have nearly as much noise as many of these libraries have with softer dynamics -- I know because I did that all of the time when I was recording. Are they all going to tape? I know Spitfire does, not sure about other noisy ones.
Hey, Good post man! I completely agree.. How would you go about using Izotope RX if you don't have any isolated noise in the sample to get a noise profile for the learn function?
 

newman

Member
The Garritan CFX has terrible noise/hiss on very low velocity samples especially in the lower registers.. It really builds up when playing multiple voices at the lowest velocities and almost dominates the sound of the piano. It is driving me crazy at the moment.
That happened to me.
I think this is an excellent piano for playing but only under basically default settings:

- Default velocity curve
- Switch off the "limiter button" (broken)​
- No reverb hahaha
- Increase the "dynamic range" dial a lot YMMV
- Try the CyberGene pealing patch from the PianoWorld site for playability

I think the Full, Classic, Default piano is the best sounding mic mix by a long shot. The is no noisier than the other mic "pairs" as illustrated by the karvala graphs referenced below:

http://forum.pianoworld.com/ubbthreads.php/topics/2695381/re-a-bit-disappointed-in-garritan-cfx-lite.html#Post2695381
 

germancomponist

Senior Member
Hey, how do you go about using x-noise if you don't have any isolated noise in the sample to get a noise profile for the learn function? The samples I'm having hiss issues with only have hiss present while the instrument is playing, so it's not possible to get an isolated noise profile.
I always use the staccato recordings for this ....
 

Dietz

Space Explorer
Hey, how do you go about using x-noise if you don't have any isolated noise in the sample to get a noise profile for the learn function?
You can achieve remarkable results with X-Noise using nothing but the default "flat" noise profile and the high-shelf EQ, especially in case of tape hiss and mic/pre-amp noise-floor issues.
 

Uncle Jesse

Member
You can achieve remarkable results with X-Noise using nothing but the default "flat" noise profile and the high-shelf EQ, especially in case of tape hiss and mic/pre-amp noise-floor issues.
Cheers for this! So you would advise I go with x-noise over RX or z-noise for this? Really appreciate the help. I will have a play around with the high shelf eq in x-noise.
 
OP
H

Headlands

Active Member
In my case, it was in the quiet section of a cue, where it was a single string patch playing. With the CS2 patch I was using, the hiss (or room noise) was just too annoying, especially when the sample cuts off. Switched to a Hollywood Strings patch...problem gone. Some libraries are worse than others, and of course it's natural to have certain amounts of "hiss" but in this case it was sloppy engineering on the developer's part IMO. Hollywood Strings isn't perfect either...there'a few string patches where I can hear stuff like coughing, whispering, and even chairs moving. Quite amusing actually!
Yelp, CSS is bad in many cases. I recently was encountering immense amounts of hiss on the basses that cut off when the samples cut off, as you mentioned. In this case (yesterday) my client actually mentioned it to me because the sus basses were on their own in one section, so I ended up using Izotope RX. Not ideal because you can't sample the noise in these samples, but it helped a fairly decent amount. That kind of hiss just doesn't need to be there in 2019. Period.

Everyone take note: I'm NOT talking about room tone. I'm talking about hiss, as in tape hiss, electronic noise kind of hiss. I've spend years recording and know that the kind of hiss I hear in some of these libraries does not need to be there. Does it keep me from making music? Duh, of course not. :) It doesn't mean it's something that can't be improved to make things better.
 
Last edited:

Uncle Jesse

Member
That happened to me.
I think this is an excellent piano for playing but only under basically default settings:

- Default velocity curve
- Switch off the "limiter button" (broken)​
- No reverb hahaha
- Increase the "dynamic range" dial a lot YMMV
- Try the CyberGene pealing patch from the PianoWorld site for playability

I think the Full, Classic, Default piano is the best sounding mic mix by a long shot. The is no noisier than the other mic "pairs" as illustrated by the karvala graphs referenced below:

http://forum.pianoworld.com/ubbthreads.php/topics/2695381/re-a-bit-disappointed-in-garritan-cfx-lite.html#Post2695381
Wow that thread is so interesting. I had no idea others were having the same issue with the CFX. It is driving me insane!

Thanks for your advice! Increasing the dynamic range won't help my issue cause the piece I'm producing is played entirely on very low velocities. You can imagine the hiss issues I'm having with the CFX.

Where is the the CyberGene pealing patch? I couldn't seem to find it.
 

Dietz

Space Explorer
Cheers for this! So you would advise I go with x-noise over RX or z-noise for this?
You're welcome! Keep in mind that the EQ shapes the _noise_, not the signal itself. Therefore, increasing the treble will also increase the effect of the denoiser, which means a darker output signal (or just more artifacts).

RX De-Noise and Z-Noise are quite different animals. Z-Noise actually the Waves-version of the legendary Algorithmix Denoiser (AFAIK) and depends completely on a good noise-print (but why not give it a try, too! :) ...). RX is even more capable, but it eats CPU for breakfast.

... the beauty of X-Noise is its simplicity and efficiency, which lends this plug-in a lot to the quick fix.
 
Last edited:
  • Like
Reactions: EgM
Top Bottom