• Register
  • Help
Page 1 of 3 123 LastLast
Results 1 to 10 of 26

Topic: Getting the best pp layer with volume, little attack, no hiss?

  1. #1

    Getting the best pp layer with volume, little attack, no hiss?

    Two questions,really, about working with samples that are a little too quite in the pp layer, and recording the pp layer in general:

    1. I\'m working with some pp layer piano samples that were well-recorded, but that it looks like I\'ll have to compress them to get enough volume. (I\'ve tried substituting the p layer, and the hammer thump is just too strong.)

    Obviously, I immediately run into problems with increased hiss when I compress them. If I then use Cooledit to reduce the hiss, I\'m losing more of the timbre than I want.

    2.Can anyone suggest the best methods for getting the best pp layer while recording? How do you get the volume without geting too much attack, and at the same time give the wave form time to develop without getting too much room and\\or hiss? In sampling a piano, in other words, how do you get a good volume on the pp layer, give the wave time to develop, avoid too much attack, and not just turn up the mixer so that the noise floor gets raised? (Or do other people use such good equipment that they can turn up the mixer without getting the hiss I do?)

  2. #2
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Dallas, Texas

    Re: Getting the best pp layer with volume, little attack, no hiss?

    The soft layers are always harder.

    Basically, everything you said.

    The best hope you have is recording in a \"hear your heart beat\" quiet room. These are pretty rare. There are a couple here in Dallas that are that quiet, one is actually known for solo piano recording, the other is an ADR/voiceover room. Both are so quiet you can hear the blood moving in your body. Even at that, you\'ll get a layer of amplification noise, but it\'s easier to deal with. Along with that, you\'d want to bypass the studio\'s snake most likely, and cable discretely with top-shelf cabling, directly to a top-shelf preamp, and direct to top-shelf converters from that. Even then, you may end up dealing with noise, but on a far less invasive scale.

    Producers are split on whether you raise the mic gains on lower velocities. Most do, by a calibrated amount, then reduce the gain by the same calibrated amount in post-processing or mapping.

  3. #3

    Re: Getting the best pp layer with volume, little attack, no hiss?

    I realize that this is somewhat out of context, but it\'s an excellent opportunity for me to advocate my views on the topic of noise reduction and orchestral samples.

    I think people are too afraid of hiss in general. In the 1500++ scores I\'ve listened to in my lifetime I have yet to hear a single recording of strings that has little or no hiss in their PP relative parts. I find that if you noise reduce this dynamic you shoot yourself in the foot. With the noise reduction you kill the air and breath of the strings (I have yet to hear a successfully noise reduced PP violins sample and I have pretty much all commercially available orchestral libraries) and thus you diminish the whole purpose of writing in such a dynamic range. The best PP strings I have ever heard is in the old roland string library. It captures the flautando (it. for flute-like) effect, performed by touching the string with the bow so lightly that you are merely brushing it (close to the fingerboard), thus producing a dreamy, organic and sparkling sound. Any attempts to noise reduce such a timbre will result in a lifeless and sterile sound as you take away something that is perceived by the ear as actual depth and air in the recording. This effect is not very effective in a large concert hall because of the extremely poor projection (if orchestrated for violins alone), but is used very frequently in film scores, where it is almost impossible to capture the sound without compromising some hiss.

    Here\'s an example. An excerpt from E.T - John Williams.

    The first mp3 shows the original as recorded and presented on the 20th anniversary CD (which was digitally remastered, mind you):


    The second mp3 shows the same clip with X-noise reduction applied:


    The 3rd mp3 shows the frequencies that the noise reduction sacrificed.


    I don\'t think these clips need further explanation.
    Surely all of you agree that the noise adds depth, warmth and air to the overall sound of the PP passage. This is why I\'m a strong proponent of the gentle use of noise as air, and also why I welcome noise to a certain degree in libraries. You can always take it away, but you can\'t get it back.


  4. #4

    Re: Getting the best pp layer with volume, little attack, no hiss?

    Thomas, thanks for pointing that out (how\'d you seperate frequencies like that? [img]images/icons/smile.gif[/img] ) It is true, the air (hiss) that you hear does breathe life into soft soft string arrangements. I try to emulate that whispery effect by introducing brief bow noise before playing a sample (included with GOS updates), than descrescendo a slight bit to keep that \'air\' with the sustained note. This works to an extent, but will never really sound like a true ppp UNPROCESSED sustain - like Roland strings. [img]images/icons/smile.gif[/img]

    Out of all the string libs I\'ve tried, Rolands are the only ones with that sound. Hmmmmm...

  5. #5
    Senior Member
    Join Date
    May 2003
    Stuttgart, Germany

    Re: Getting the best pp layer with volume, little attack, no hiss?

    Thomas, thanks for sharing. Incredible post IMO. I haven´t your expierience but I tried to implement noise reduction on strings once and never again fot the reason you mentioned. So I completely agree with you.


  6. #6

    Re: Getting the best pp layer with volume, little attack, no hiss?

    It\'s too bad that there\'s doesn\'t seem to be noise reduction software that doesn\'t take out some of the sample with the hiss. Or at least that\'s my very limited experience with Samplitude and Cooledit. Even if I capture a noise sample from an \"empty\" part of a sample in Cooledit and then try to take that out of the rest of the sample, it tends to take out a little of sound that I want.

    Do all the noise reduction methods rely on the pitch of the sound you\'re trying to cut, so that it takes out the same pitch in the data you want? My impression is that they take out a range of frequencies, no matter how I shape the removal. Would some kind of microscopically capable multi-band noise reduction, one that could isolate very narrow bands in the spectrum, be possible to create, so we\'d have a better chance at getting rid of the hiss, but still keeping the air and bow scrapes? I suspect, in other words, that while we may not be able to get rid of hiss that\'s at the exact pitch of each bow scrape, we might be better able to get rid of hiss that\'s on each side of the spectrum a little better, maybe by subdividing the spectrum so we could work on units of frequencies of, say, 521.543 hz instead of just 521 hz.

  7. #7

    Re: Getting the best pp layer with volume, little attack, no hiss?

    Hi! Very nice thread!

    If you remember, the first noise reduction systems offered by Dolby, where based in pre-enphasis, that is, enhacing the highs before recording on tape, and applying the inverse EQ when playing. This had the effect of reducing the natural hiss on playback. This was simple and effective at the same time.

    Today, things are different. The best noise reduction algorithms rely on statistic analysis of the input signal to diferentiate between usefull signal and noise. That\'s why they usually ask for a short segment of \"sample noise\". They analyze this short segment, and derivate the frequency footprint and statistical behaviour of the noise. And this works pretty well in most cases, for example when you are trying to reduce background noise from a voice recording.

    But, (and back to the topic [img]images/icons/wink.gif[/img] ) in the particular case of pp strings, you are trying to diferenciate between one noise and another noise... whow! This can only be accomplished at SETI!

    I hope this brief explanation help you understand why the noise reduction system eliminated usefull portions of sound in the examples of E.T. The system can not understand that you are interested in that noise [img]images/icons/confused.gif[/img]

  8. #8

    Re: Getting the best pp layer with volume, little attack, no hiss?

    Is this one of the commercially available programs that lets you control the attack of the noise reduction?

  9. #9

    Re: Getting the best pp layer with volume, little attack, no hiss?

    This is what I did for piano samples.

    - Set up the samples to play back as soundfiles in ProTools. (or other DAW).
    - Then add the Noise Reduction Plug-In in the mixer. Base the noise reduction on the hiss only (not the tone of the piano). You need a sample of the isolated hiss for this. (Ask the developer for it if you can’t find such an isolated piece of hiss on your disk.)
    - Edit the noise reduction amount on the effect parameters section and (here it comes)
    - set up the noise reduction amount fader for automation.
    - With automation active fade the noise reduction in slowly when you hear the hiss becoming apparent.
    - Make sure you have no noise reduction during the attack and early decay where the signal/noise ratio is still acceptable.
    - Repeat the adjusting of the NR for each sample individually.
    - Confirm/consolidate soundfiles (make sure names are not changed).
    - replace noise samples by NR-processed samples.

    You can use about any good noise reduction program that supports automation. WAVES and CEDAR are very good.
    If you really want the best results you need to check out SonicSolutions No-Noise.

  10. #10
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Dallas, Texas

    Re: Getting the best pp layer with volume, little attack, no hiss?

    You cannot compare noise-reducing a tutti recording, where the total noise will be 1x the room/gear noise, with sampling, where the total noise will be room/gear noise x polyphony. In a sampled piano you could easily end up with 80x the room noise in a pedaled passage. No way that is acceptable.

    We discussed this to no end in the hall vs. soundstage topics. No recording process is free of compromise. I agree noise reduction can hurt the space in a recording. But there are a lot of factors in play, mainly the polyphony. You have to compromise somewhere, or you have to record in an essentially noise-free environment. You can\'t have it all.

Go Back to forum

Tags for this Thread


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts