I decided to make a separate thread for this topic which has been brought up in other threads.
My feeling on this is that the thinness is most noticeable in violins. It has never ocurred to me that basses sounded thin, or even, upon reflection, cellos. Maybe violas a little, but mostly the violins. And the phenomenon is mainly in fully scored passages, which in a live orchestra does not seem to happen. In live playing, the violins may simply disappear if overbalanced by the brass or percussion, but they never do this atrocious morph into a thin little whiny tone.
So why does this happen? Is it the fault of the samples or the mix? Is it a mysterious phenomenon peculiar to samples, as opposed to RECORDED live orchestras? Even in recordings of live performances, you don't seem to hear this, though it is definitely true that some recordings by expert engineers will sound much fuller than others.
What can be done? Wait for new samples as some people seem to think?
Some things that occur to me are the following:
Dynamic range This is a crucial aspect of the overall sound, since samples distort dynamic range in an extreme way. A solo piccolo can be louder than an entire trumpet section. And this pertains not only to the overall level, which may be simple to adjust in a mix, but to the variation of dynamics in a line and the variation of sound pressure throughout the range of the instrument. The piccolo, to use as an example again, simply disappears in its low register, but is audible over the entire orchestra in its top notes. All these elements must be represented in the performance/mix since the entire process of recording the samples deliberately nullifies the differences. In effect the midi performance must restore the differences between the instruments.
EQ one thing I constantly notice is that sampled strings always sound brighter, even grotesquely brighter, than live. When a live violin section is playing softly, they become extremely dark, almost like sine waves with just a little bit of rosiny bow edge audible. This never happens in samples. They always retain their brightness because the job of the sample recording engineer is to capture all of the frequencies. So I think that EQ is something that should be approached with a far more extreme approach than is usually assumed.
Ensemble Another aspect of the bad sampled string sound is the lost sense of many different instruments playing at once. Whatever can be done to restore this ensemble feeling is needed, whether through layering solos or different smaller sections (like VSL Chamber Strings) or adjusting the stereo size and overall level and not allowing its sense of space to be covered up. Also adjusting the timing of layers across tracks is something that is very important: in live ensembles you will hear players all over the place as compared to a perfect sampled unison line.
Sample Quality There is one sample library (not to be named here because I have already alienated enough people) that is absolutely awful in its basic sound quality, and always sounds thin, whiny and nasal. Even alone! There is another - Miroslav Vitous - which is old and far too limited to be a competitor for VSL (or any state of the art company) but which has a number of non-chromatic samples that are nevertheless really spacious and full sounding. This was a function of the original recording/playing quality. Some of this quality must be heard in a combination of texture and size, as well as basic things like intonation and tone quality. These elements are by no means automatic in a recording, but are the "art" of sample recording. Without this art fully present in the original sessions and the processing/post-production, it is useless to have thousands of notes and articulations.
It would be helpful to hear other ideas on this topic, especially practical approaches to overcoming the problem.
My feeling on this is that the thinness is most noticeable in violins. It has never ocurred to me that basses sounded thin, or even, upon reflection, cellos. Maybe violas a little, but mostly the violins. And the phenomenon is mainly in fully scored passages, which in a live orchestra does not seem to happen. In live playing, the violins may simply disappear if overbalanced by the brass or percussion, but they never do this atrocious morph into a thin little whiny tone.
So why does this happen? Is it the fault of the samples or the mix? Is it a mysterious phenomenon peculiar to samples, as opposed to RECORDED live orchestras? Even in recordings of live performances, you don't seem to hear this, though it is definitely true that some recordings by expert engineers will sound much fuller than others.
What can be done? Wait for new samples as some people seem to think?
Some things that occur to me are the following:
Dynamic range This is a crucial aspect of the overall sound, since samples distort dynamic range in an extreme way. A solo piccolo can be louder than an entire trumpet section. And this pertains not only to the overall level, which may be simple to adjust in a mix, but to the variation of dynamics in a line and the variation of sound pressure throughout the range of the instrument. The piccolo, to use as an example again, simply disappears in its low register, but is audible over the entire orchestra in its top notes. All these elements must be represented in the performance/mix since the entire process of recording the samples deliberately nullifies the differences. In effect the midi performance must restore the differences between the instruments.
EQ one thing I constantly notice is that sampled strings always sound brighter, even grotesquely brighter, than live. When a live violin section is playing softly, they become extremely dark, almost like sine waves with just a little bit of rosiny bow edge audible. This never happens in samples. They always retain their brightness because the job of the sample recording engineer is to capture all of the frequencies. So I think that EQ is something that should be approached with a far more extreme approach than is usually assumed.
Ensemble Another aspect of the bad sampled string sound is the lost sense of many different instruments playing at once. Whatever can be done to restore this ensemble feeling is needed, whether through layering solos or different smaller sections (like VSL Chamber Strings) or adjusting the stereo size and overall level and not allowing its sense of space to be covered up. Also adjusting the timing of layers across tracks is something that is very important: in live ensembles you will hear players all over the place as compared to a perfect sampled unison line.
Sample Quality There is one sample library (not to be named here because I have already alienated enough people) that is absolutely awful in its basic sound quality, and always sounds thin, whiny and nasal. Even alone! There is another - Miroslav Vitous - which is old and far too limited to be a competitor for VSL (or any state of the art company) but which has a number of non-chromatic samples that are nevertheless really spacious and full sounding. This was a function of the original recording/playing quality. Some of this quality must be heard in a combination of texture and size, as well as basic things like intonation and tone quality. These elements are by no means automatic in a recording, but are the "art" of sample recording. Without this art fully present in the original sessions and the processing/post-production, it is useless to have thousands of notes and articulations.
It would be helpful to hear other ideas on this topic, especially practical approaches to overcoming the problem.
