• Register
  • Help
Results 1 to 10 of 10

Topic: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

Share/Bookmark
  1. #1

    Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    When I play ff, GigaPiano peaks and distorts, but if I reduce the amplitude so I can really press down hard without distortion, pp and mp notes aren\'t loud enough and sound thin.

    I don\'t know if this is a problem with the samples, or more probably, my own ignorance about controlling amplitude in a digital environment. I get this same result regardless of whether I reduce the volume on my mixer, in the Attenuation box on the Sample tab in the Giga Mixer, or the Master tab on the DSP page. Still too soft and thin in the pp and mp layers or still distorting in the ff layer.

    It\'s occurred to me that I could increase the attentuation of all the ff samples, but I wonder if this is really necessary, since I\'ve never seen anyone discuss this problem on these forums. It seems more likely that I\'m doing something very basic wrong.

    Does anyone have suggestions?

    (According to Paul White in the Sound on Sound books on recording, digital signals need to reach a certain level or they lack accuracy--apparently a lower amplitude means a lower bit resolution. Is this accurate, and why Gigapiano sounds thinner at lower volumes?)

  2. #2
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Location
    Dallas, Texas
    Posts
    5,755

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Hi Jake,

    There is nothing you can solve here. The Gigapiano, unfortunately, is a very BIG SOUNDING instrument due to its engineering design. Ironically, in a a mix, sometimes the biggest sounding instruments can sound small.

    Here\'s a lesson in the basics of engineering, and in the fallacy of \"bigger is better.\"

    Basic issue: The Gigapiano is recorded with the wrong microphones at the wrong distance for your current application.

    Larry Seyer used mainly Audio Technica 4033s on that piano, at a point blank distance. The 4033 is a large-diaphragm, FET amplified condenser--a choice normally used for vocals and solo instruments an engineer wishes to push \"to the front\" of a mix. In addition, a Soundfield (expensive, multi-large diaphragm single point mic system) was used, although the level of its contribution doesn\'t seem particularly noticeable.

    One of the very basic mistakes of any sort of engineering is \"falling in love\" with the large diaphragm microphone\'s HUGE presence and track-filling size. The issue is this: for the main vocal, or solo instrument, this phenomenon helps that track stay on top when other tracks are competing for space. So, if EVERY track (or tracks you need to recede) is recorded with this hugely exaggerated perspective, there is simply not room in the virtual soundscape for all those signals. They crowd each other out--and to get to your point--the consequence is that the tracks are either \"too loud, too soft, but never just right.\"

    Pianos ARE close-mic\'ed in studios every day to good result, but generally not with a microphone that accentuates closeness. The AT 4033 is a VERY unusual choice for a piano. Take Kip\'s Bosendorfer sample as a perfect example of the more classic choice. Although it is a close-perspective recording, it can either take a prominent position, or recede, but you\'re always going to hear it because a very SMOOTH response, small-diaphragm mic plot was used (in this case, B & Ks...very fine mics).

    So...back to the GigaPiano.

    What you were actually doing that gave you success when the outputs distorted was COMPRESSING the instrument (although using digital \"overs\" to accomplish it isn\'t the best way to do it). Even though you were suffering the overs, you were getting a better mix because you were REDUCING that extreme dynamic range and intimacy. Now that you don\'t have the compression resulting from those overs, when you balance your track so that the loud parts sound right, the medium and soft parts fall under the overall level you need in the mix and virtually disappear.

    I wasn\'t in the studio when the GigaPiano was recorded, so I can\'t comment on specifics, but my guess is that someone sat and played the piano while those mics were being placed, and that the engineers got seduced by how much bigger and bigger and closer and better it kept sounding--and they screwed themselves and you right up the pooper by not putting that piano into a bunch of different mixes and making sure that they had produced a usable instrument. And indeed, the GigaPiano was the \"proof of concept\" for GigaSampler, and I\'m certain a design goal was to produce a piano that people would sit and play and immediately freak out over. Probably the desire to get this quality overshadowed common sense.

    Should Larry Seyer should have known better? He\'s a Grammy awarded engineer, after all. That\'s a hard call. That mic plot may have worked in a particular production. But it doesn\'t work for the great majority of productions. This is the ART of the sample developer and engineer...that person needs so much experience with not only recording, but with the unique aspects of the end use, that it may just have been outside Larry\'s exprerience as an engineer and producer of mainly bands. Something that worked for him, and maybe even sounded amazing in the studio at the time, was really not a very flexible choice--hence the common complaint that the GigaPiano sounds unnatural, and that it doesn\'t mix easily. This gets blamed on the sample stretching and even on the release triggers, but it is really just about the microphone choice and placement. You won\'t find very many settings where it will be the ideal choice...one may come up here and there (it does, really), but just not often.

    SOLUTIONS:

    Hard to fix this one. The mic plot here is the audio equivalent of standing next to an elephant and looking at him through binoculars. The perspective is so huge that it\'s hard to squeeze it down. You can reduce the phenomenon some, but you\'ll never completely change it--the perspective combined with the mic plot is just so extreme the sonic imprint is always going to be there.

    Try compression on the rendered track. This will at least reduce the dynamic range, allowing you to raise the volume of the track and therefore raise the level of the quiet sections. A multiband compressor may help more, by reducing the amount of sustain in some registers...and a substitute for this would be to put an EQ in front of the compressor, so that you\'re reducing problem frequencies before hitting the compressor--then the compressor will act more on the desirable frequencies than the undesirable ones.

    But your ultimate solution, and the reason that multiple piano libraries are a good idea, is to obtain libraries with differing engineering schemes...so you get the sampling equivalent of a CHOICE of mic placements.

    This is also why I tend to jump into threads where people are complaining about a particular piano\'s sound, because what they are really complaining about is that they don\'t like a particular mic perspective they\'re hearing. BUT...your GigaPiano problem illustrates just why that\'s not really a valid indicator of quality. That piano that sounds just brilliant to you playing solo, or in a certain setting, can sound like in another setting--where the piano you thought sounded like crap is suddenly the perfect choice.

    If you were running the session from scratch, you\'d move the mics or change mics to fix this in tracking--getting your sound almost completely dialed in before mixing even started. Unfortunately with samples, you just can\'t do that--even with mic modelers, convolvers, and the like--the basic imprint is still there, and you just cannot change it once it\'s committed.

  3. #3

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Bruce

    You seem to know the history of the sampling of this piano (a Yamaha 7 footer I believe). Can you comment on why the middle register is what I call \"muddy\", though others claim its the wear and tear on those registers from players using that region so much, that causes it to sound so dull. Thanks

    Larry

  4. #4
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Location
    Dallas, Texas
    Posts
    5,755

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Originally posted by Larry Negro:
    Bruce

    You seem to know the history of the sampling of this piano (a Yamaha 7 footer I believe). Can you comment on why the middle register is what I call \"muddy\", though others claim its the wear and tear on those registers from players using that region so much, that causes it to sound so dull. Thanks

    Larry
    <font size=\"2\" face=\"Verdana, Arial\">I doubt that\'s why it\'s muddy. People tend to keep their pianos in good shape in a studio setting, and any piano tech worth his salt would be keeping the hammers maintained when needed. I really think it\'s more to do with the super close/large diaphragm perspective, combined with the frequency curve of the 4033s. That microphone is extremely honky in the middle register (which is why people tend to use it on darker male vocals)...so it\'s supporting the very frequencies in that very region which would tend to accentuate \"muddiness\" over \"clarity.\"

    With pianos, mic distance is a very simple proposition. The closer you get, the more sustain you get, and the less impact you get. The more distant you get, the less sustain you get, the more \"wood\" or percussive content you\'ll notice. A large diaphragm mic sounds \"closer\" at a given distance than a small diaphragm mic. A cardioid mic sounds closer at a given distance than an omni. Finding the best location for a given purpose involves adjusting all of these factors in balance with one another. Finding the best location for a sampled instrument depends upon whether you want to make a piano that\'s ideal for certain tasks, or to find a somewhat happy medium that is adaptable to many tasks.

    There is probably some very minor contribution to muddiness in the mapping, where 31 discrete pitches were used and stretched rather than the full range. This too, will tend to add to the washy quality if you happen to be triggering multiple notes in the same sampled region. Another unpleasant side effect is hearing the overtones and instrument resonance \"ride up\" with the chromatic pitch, instead of remaining constant against changing pitch in the case of resonance, and displaying unique characteristics per note in the case of overtones.

    So, it\'s probably a combination of \"all of the above.\" But with piano, as with any studio-mic\'ed instrument, the biggest contributor to the final tone other than the instrument itself will be the choices of microphones and placement. In a tracking situation, where you\'re setting up a whole ensemble or rhythm section, moving mics around by inches here and there, and switching from one brand/type to another, are things that engineers spend the first hours to days accomplishing. It is the initial and most important step in defining the overall recorded perspective of a project.

    By the way, all of the technical information along with some limited photographs of the mic plot exist in the documentation. I didn\'t have any sort of other information on the session...this information is delivered with the instrument itself. If you want inside info on this one, Dave Govett is the man to ask. While I don\'t know if he sat in on the actual tracking sessions, I believe that he did the final mapping.

  5. #5

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Thanks Bruce. I do want to make clear, however, that my problem is not getting Giga piano to sit right in a mix. It\'s playing solo that gives me the problem: ff clips\\distorts when I really bang on the thang, and if I lower the amplitude so ff sounds right, pp and mp sound thin.

    I understand what you\'re saying about the elephant through the telescope perspective, but I guess I can\'t understand how the instrument could have been recorded in such a way that you can\'t play both loud and soft without changing settings. Assuming the engineers left the mics in the same place for all velocities, wouldn\'t they have heard the same clipping I do when they sampled, telling them to reduce the gain for the ff samples?

    Regardless, thanks again.

  6. #6

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Jake,

    There have actually been a few threads on improving the sound of Gigapiano. They even included altering filter settings.

    I\'d seriously have a go at highlighting just the ff layer of samples and dropping their attenuation a bit. If you use the blue balls you can maintain relative differences between different regions in the ff layer, but move them all down by the same amount.

    If you screw up, you can always copy the instrument over from the CD again.

  7. #7
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Location
    Dallas, Texas
    Posts
    5,755

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Originally posted by Jake Johnson:
    Thanks Bruce. I do want to make clear, however, that my problem is not getting Giga piano to sit right in a mix. It\'s playing solo that gives me the problem: ff clips\\distorts when I really bang on the thang, and if I lower the amplitude so ff sounds right, pp and mp sound thin.

    I understand what you\'re saying about the elephant through the telescope perspective, but I guess I can\'t understand how the instrument could have been recorded in such a way that you can\'t play both loud and soft without changing settings. Assuming the engineers left the mics in the same place for all velocities, wouldn\'t they have heard the same clipping I do when they sampled, telling them to reduce the gain for the ff samples?

    Regardless, thanks again.
    <font size=\"2\" face=\"Verdana, Arial\">Hi Jake,

    Your assumption points out the fallacies that are widely believed about recording in general, and specifically recording for samplers.

    Just putting two mics at the right place and recording doesn\'t necessarily guarantee you anything. Microphones are not ears. They may have less sensitivity or less warmth at one amplitude vs. another. Speakers are not instruments. So even given that \"perfect placement\" and a perfect mapping, your speakers may well not be putting pianissimo in the room as efficiently as it\'s putting fortissimo in the room. The room itself can be contributing uneven absorption or loading. There is so much alchemy involved in this process that one cannot just reduce it to a simple equation. That\'s why people work all their lives to become great engineers or producers. There is a LOT to it.

    But in the case of the GigaPiano, you have microphones which very specifically \"enhance\" closeness, that are placed at such an additionally close perspective, that the dynamic range will necessarily be skewed. Imagine putting your head inside the piano lid, and playing to get a solo-piano perspective from that. Another complicating factor, which brings about any other number of intervening techniques to produce the instrument you know as GigaPiano.

    Even that is OK in terms of engineering. I guarantee you that while I don\'t agree that Larry Seyer chose the best plot for this particular application, he didn\'t just pull it out of his butt. Somewhere, at sometime, that plot produced exactly the result he needed. But Larry is not producing solo piano recordings in his line of work, so that is an aspect you take into consideration, and realize that the GigaPiano was probably never conceived to do what you want it to do.

    The side-effect of getting clipping at fortissimos is easily solved...reduce your total output from the GigaStudio by a few dB. This is the correct way to approach the problem. You may wonder why it wouldn\'t make more sense to just have a quieter fortissimo, but that practice would drive the level of all samples down, and would therefore expose quieter levels to more dithering and quantization noise.

    When I say \"in a mix,\" a solo-piano track is a \"mix,\" too. Trust me, even a recording of a real solo piano for a CD gets mastered and gets post processing to sound its best. Most often, what\'s called \"bottom up\" compression, to bring up the low dynamics a bit without reducing the range or scope of the high dynamics. You do this by combining compressed and lifted signals with uncompressed signals in the analog world, or with the Waves C4 or Linear Multiband in the plugin world. But whatever the case...you cannot expect the sound to just \"spit out\" of the sampler and sound like a finished product. And you DON\'T want that!! Otherwise, you\'d never be able to mix!! These things have to happen in a logical stepwise fashion for good finished results, and they\'re all a part of the overall production chain. Shortcutting it at the sampler end will just screw you down the line.

    That said, if you want a playing experience with a little less dynamic range (to warm up those quieter moments), do exactly as Chadwick has instructed you.

    And ALL that said, the GigaPiano is in NO WAY mic\'ed in a fashion to deliver results as a solo-played instrument. Its mic\'ing scheme, in fact, is more close and hyper-huge than most pop recordings can hold due to all the factors I described--so you\'re never going to get a sound from that perspective that sounds like a typical solo-piano recording. Don\'t beat yourself up trying...you need to find a source instrument that is closer to the sound you want your finished product to have, and you\'ll spend more time playing and less time scratching your head wondering why it ain\'t going your way!!

  8. #8

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Bruce: Thanks again. Got it this time.

    But I have to admit that I\'m still confused (still!!) about why the pp and ff levels are incompatible. Seems like either of the possible mistakes, regardless of the mic, would have been caught before or during mastering: If the engineers sat at the mixer while someone played all the velocity levels for a single note, they would have heard the clipping. If they instead recorded a velocity level for each note, and then combined them into a set of samples, wouldn\'t they have also heard the clipping once they combined the samples into the multiset and played it?) I am not attacking the engineers. (OK, maybe I am. Don\'t think so. ) I\'m just wondering if this problem of having to choose between clipping and thin pp\\mp (or increasing the attenuation of low notes) comes from bad settings on my end. The only reason I\'m concerned is that no one else on this forum seems to have had this problem, and I\'m by no means one of the more experienced people.

  9. #9
    Senior Member Bruce A. Richardson's Avatar
    Join Date
    Sep 1999
    Location
    Dallas, Texas
    Posts
    5,755

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Originally posted by Jake Johnson:
    Bruce: Thanks again. Got it this time.

    But I have to admit that I\'m still confused (still!!) about why the pp and ff levels are incompatible. Seems like either of the possible mistakes, regardless of the mic, would have been caught before or during mastering: If the engineers sat at the mixer while someone played all the velocity levels for a single note, they would have heard the clipping. If they instead recorded a velocity level for each note, and then combined them into a set of samples, wouldn\'t they have also heard the clipping once they combined the samples into the multiset and played it?) I am not attacking the engineers. (OK, maybe I am. Don\'t think so. ) I\'m just wondering if this problem of having to choose between clipping and thin pp\\mp (or increasing the attenuation of low notes) comes from bad settings on my end. The only reason I\'m concerned is that no one else on this forum seems to have had this problem, and I\'m by no means one of the more experienced people.
    <font size=\"2\" face=\"Verdana, Arial\">Hi Jake,

    There is no clipping in the FF samples.

    If people were to engineer instruments so they wouldn\'t \"clip\" GigaStudio, then what that would mean is that they\'d have to sum up all the polyphony on all the FF samples across the keyboard, then reduce the amplitude of those samples to a level so \"safe\" a user could never drive the instrument into clipping no matter what he did.

    but THEN...

    The levels of the PP samples would be SO low, and so far down into the 16-bit bandwidth that they\'d basically be covered in quantization and dithering noise.

    So, no one would do that, and it\'s a good thing. Recording the samples, then going about the business of cutting up the raw material, prepping, noise reduction, cataloging, mapping, etc...those are all huge tasks, and it doesn\'t really go in a fashion such that they \"load up\" the samples and play them as they\'re capturing.

    Just load the GigaPiano, go into the SETTINGS | SAMPLING tab, and lower your Master Attenuation tab until the clipping stops.

    Or, turn down the GigaPiano\'s levels in the DSP Station.

    The phenomenon of why the pianissimo and fortissimo are a \"mismatch\" however, relates to the extended post that I made here. You are trying to get a solo piano sound out of a particular instrument which was recorded with a mic plot that is more typically a rhythm-section \"comping\" style perspective.

    That is why it\'s thinning out on the pianissimos. If you want to redesign the instrument, lower the fortissimo and forte levels slightly as Chadwick pointed out, and you will have what you\'re wanting.

    The engineers made a decision. If it doesn\'t please you, you can change it. That\'s the beauty of it.

    And again...I don\'t know that you really paid attentionto what I said in the post because I explained this: If you were to record a real live piano, you wouldn\'t just call that recording finished on your first playback, either. You would probably use some mastering compression, for the VERY purpose of \"thickening\" the pianissimo register and \"warming up\" lower amplitudes. It is not necessarily a \"defect\" that this piano was designed as it was (although I myself don\'t choose the GigaPiano for very many tasks).

    Good luck. Re-read the stuff I wrote about engineering, and it will help you understand the roots of what you\'re hearing in this instrument and in others. And my suggestion would be for you to listen to some of the piano demos of other libraries and to see if one of the third-party solutions might make you happier. Most people do not ultimately end up using the GigaPiano...it was a proof-of-concept instrument for the original Gigasampler, and a lot of study and thought has gone into piano sampling methodologies since its release.

    However it comes out, good luck.

  10. #10

    Re: Amplitude and GigaPiano: too soft or too loud? What am I doing wrong?

    Thanks again. Have to ask a question I don\'t like-
    Given all that\'s been said about the choice of mics and their placement, if I tried to choose among the Mag, the latest Bardstown Bos, and the Post Gradndioso, would the choice be entirely a matter of preference in timbre, or are there recording issues that I should be aware of?

    Sorry if this puts anyone on the spot. I want a piano that sounds good once it\'s recorded but also sounds very good as I\'m playing it. The latter requirement is the most essential. My playing is strongly influenced by what I hear as I play. But I want it to sound good as I play and sound exactly the same when I listen to the recording of it.

Go Back to forum

Tags for this Thread

Bookmarks

Posting Permissions

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