Can't get enough of 8DIO's Legacy 1928!

Does it have resonance? I could have saved more money for it, but I decided to purchase the Galaxy vintage D, which I think is good anyway.

yes it does

Further to my previous comments, the above are some recent pianos that I feel have offered something unique.

The Imperfect Samples pianos capture piano ambient noises in a way that I don't think I've heard in any other sample library - eg not just the sound of the string release, but of the key being hit by fingers and of the key being released in addition to the pedal sounds, sympathetic resonance or round robbin. These sounds really add to the realism of the performance. I'm not sure I entirely agree with the Imperfect Samples philosophy in other aspects (eg the choice to play with fingers rather than using machines to get consistent velocity layering) - I think imperfection should come from the original instrument itself and from the end players performance, not from the way the samples are recorded. None-the-less to me this is still the most realistic sounding piano sample library.

The 8Dio Legacy also tries something new in using extensive round robbin notes (eg 6 staccato, 3 sustained for each note and velocity layer). Once again I'm not sure I agree with the concept - theoretically a note played with the same intensity should give very close to the same sound every time, and as such the main variable would be the intensity of the key strike. As such, I think multiple velocity layers should be a higher priority than round robbin. Once again though, it sounds gorgeous. It also includes morphing with the sound of other instruments (eg a hang drum), which does create some beautiful new sounds.

The Cinesamples Piano in Blue has recognised that sample libraries aren't necessarily used to try to sound like a real piano, but rather to sound like a 'recording' of real piano (which is what they are). They've sampled it from different perspectives, with a close perspective recorded using the same microphones as were used on Miles Davis' Kind of Blue recording. The mixing console and tape etc are a little more modern then that recording, but they're on to something there in creating piano sounds that sound like vintage recordings.

All that said, I should add that I'm neither a piano player yet, nor a sampling expert, so please take everything I say as potentially completely and utterly incorrect.

Thanks for the welcome back. And it's nice to hear there's still life in the old ns_kit7!

About the piano - all very interesting thoughts, thanks. It is definitely important that people can *play* the piano. You'll maybe remember this was one of the main goals of ns_kit7, and it will be no different with the piano I'm planning. There are many piano libraries which sound good, but they don't play that well; certain keys jump out, the pedal responds unrealistically etc.

Also, with regards to using it on lower spec'd systems - I think the solution to that is to make different versions. So, you can use a "lite" version for performance and mixing then switch to the "heavy" version for mixdowns.

Yep, nice sounding piano and certainly a lot of history in the studio and the instrument! Some of my all time favorite pianists have performed on that piano!

Very interesting thoughts. I hear what you're saying about round-robin on the 8Dio piano. I'm currently experimenting with round-robin and trying to decide if it is necessary. I think it is necessary for some thing (fast repetitions of an active string, for example), but I'd agree increased dynamics are more important than increased round-robin.

Yes, from the demos, I'd say that Imperfect Samples efforts sound the most realistic to me. And, as opposed to you, I think that a great deal of that is precisely because I feel like I'm hearing a real person play it and - according to them - it is a real person. But there are problems with this - when the noise of the finger hitting the key is part of the sample, it make performing the piano live very difficult.

Cinesamples sounds great, definitely. They've done a fabulous job.

I hear you with regards to Pianoteq. Despite what the the fans of Pianoteq will try to tell you, it just doesn't sound like a real piano. It has a very even response and in many ways, I'm sure it would be a better piano to use for practicing or even live work, but it just doesn't sound as good as the best sample libraries to my ears.

Yes, I like the rough character of AcousticSamples Old Black Grand. However, I hear in the demos that cropped front end of the sample, which somehow seems to plague so many sample libraries.

Thanks to everyone who has commented on this thread. It's really appreciated and given me a lot of things to think about. Please feel free to continue the discussion here or over at http://bluemorfo.com - the new name for my Sample Library endeavours!

Cheers!

What, exactly is the 'noise of the finger hitting the key' and how would the audience in a real situation hear such a thing?

I think the sampling world is going insane in a misguided and unrealistic effort to be more realistic.

It's like adding finger squeak noises to guitar libraries.
In real playing these noises usually only happen when the player moves position.
When the noises are added to a sample library they just happen randomly, or worse, every time the note that the squeak is baked into is played.

This is not realistic.
It is increasing the phoniness.

I agree that what makes the Imperfect Samples pianos sound the most authentic is that they sound like real instruments played by a real person. For example they have sampled the sounds of fingers hitting keys and keys returning to position etc. However, these are actually separate samples that can be adjusted in volume and frequency of occurrence, and are separate to the actual samples of the hammers hitting the strings (although undoubtedly the microphones would pick up some finger artifacts for those samples as well).

The reason I felt that a machine would do a better job than fingers for the sound of the hammer striking the string? When you're working with as many as 25 velocity layers, I think that it would be very difficult to accurately physically play a key at 25 different velocities with human touch. I think it would be even more difficult to accurately replicate those same 25 velocities for each key across the entire key range. At this many velocity layers I think a machine would give a much more detailed and balanced translation. Imperfections would then come from the response of the different keys on the original instrument and from the imperfections in the touch of the pianist playing the midi controller, rather than having imperfect velocities built into the samples that do not correspond with what the end pianist actually played. Other sounds such as fingers hitting ebony and ivory could then be included as separate sampled sounds to increase the end product realism.

All that said, this is just my theoretical debate about the best way to record samples. In practice the Imperfect samples pianos sound incredibly stunning and IMHO are the best sampled pianos by far (likewise my theoretical thoughts on the value of round robbin do not change the fact that I also think the 8Dio Legacy sounds awesome).
Last edited by Lode_Runner on Sat Jun 16, 2012 5:01 pm; edited 1 time in total

If you are sitting in the audience in a concert hall listening to a piano performance, you wouldn't hear the finger touching the key. If you are sitting in front of a piano playing it yourself, you would hear these sounds.

For a piano sample library designed for live performance, it might be best to leave the sound of the finger hitting the key to the actual sound of the keyboardist touching the keys on their actual keyboard (assuming a wooden keys keyboard - this wouldn't sound the same on plastic keys), and not include it in the samples. However, for a sample library that is designed for recording, it can be good to be able to replicate these artifacts where these would be picked up by microphones that are placed close the player's perspective.

Imperfect Samples allows you to mix those sounds in or leave them out as you please, so it gives the best of both worlds.

I agree with you about the addition of random finger squeak noises to sampled guitars. In all honesty, speaking as a guitarist, I'd usually be doing all I could to not have those sounds in my recordings anyway.

Are you talking about striking the key hard so that it bottoms out strongly?

Otherwise, there is no sound of the finger touching the key.
What kind of pianos do you guys play that you hear your fingers touching the keys?

A finger tapping wood or other material makes sound - just like tapping your finger on a desk, or the sound made by a typing keyboard when someone is typing. It would be drowned out by the sound of the instrument when you're half a room away, or playing loudly, but I can hear it when I'm sitting in front and pressing keys softly on my parent's upright, and I can hear it when I'm playing my keyboard (especially when my guitar nails hit the key). Then again, maybe I'm just ham fisted, and a good piano player would have their fingers touching the keys lightly before pressing down?

No, not the sound of the key bottoming out. What I'm talking about is the very real and very present sound that the tip of a finger makes when it strikes a key from a distance, especially when playing at forte or above. There's a big difference in sound between this type of approach to the key as compared to playing the depressing a key when you're already in contact with the keyboard.