Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards True Keys American D - First Impression

Macy, send me an e-mail at ryan at vilabsaudio.com. I'll get you in touch with one of our developers to make sure you're getting everything dialed in right. The Steinway we used doesn't have quite as much sustain as the Faziloi we used (that was a calculated decision), but it's certainly not nonexistent either so it might be a settings issue we can help with.

Renato,

I'm not sure lack of layers in the higher velocities is a bad thing. We just didn't need them there. We could have absolutely put them in there, but they just weren't necessary. Something tells me that reviewer wasn't really a player as they're really getting fixated on the technical details and not the playing experience.

Imagine sitting at a piano and visualizing even 15 different velocities for a key. The differences in many cases would be lost to rounding errors.

Macy, many thanks for your appraisal - fascinating reading.

Hmm. I know what you are saying but defending your product that way seems a bit upside down. I read the review and it didn't seem too nitpicky. Were it possible to test in the same way, an AP would stand up under the scrutiny of that testing. IMO, close scrutiny is warranted for DPs and sampled pianos because they tend to sound horrible and may be difficult to play when they get some of the details wrong.

I think VI is simply pointing out that it's the playing experience that counts, not the machine under the hood.

Pianists concern themselves with the former. Engineers get hung up on the latter.

Huh? I'm with Mac on this one. You have to play and listen first, then look at the specs or DPBSD if there is something you are not sure of. You just know what I think of SN which you love. Basically I think it sucks. Great dynamics and expressiveness, but unpleasant. IMHO, but not yours or others. The point is SN passes the technical questions but fails the musical ones.

I'm simply pointing out that there are devils in the details that can spoil the playing experience.

I don't believe I've ever said that I "love" SN. I think I've gone way out of my way to make it clear that, while it is technically superior to most DP sound generation options out there, it may not sonically be everyone's cup of tea. [EDIT] And, after owning it for a couple of years I'm still not sure what I think of SN sonically. What I really want a DP that's fully sampled.

I think there's a bit of a false dichotomy going on here. Something that tests well might or might not play well. Something that plays well is generally technically good. [EDIT] Unless of course you are talking about a DP that "sits well in the mix" where all bets are off.

As an engineer that also plays (albeit a better engineer than player) I get hung up on both. I'm running out of time this week but next week I'll try to capture a few measurements of the True Keys American and share them if they provide any insights into how it sounds or plays.

Well, if you don't know about it and you cannot HEAR it, what's to spoil ?

Macy,

I want your thought about before made a decision. I'm interested to buy the whole package, but I'm not sure if True Keys American will be a worth step up from Vintage D. About Italian and German, I don't have any concurrent, so it will worth.

I'm more engineer than player, that's because I want to clarify those things. If we take a look on the main competitors (Synthogy, Pianoteq ...) the main market is about the layers. Pianoteq promisses a incredible playability because it doesn't have samples, so 127 level in the standard MIDI.

From the reviews, I tend to believe that VILabs have managed the samples in a such way that the number of layers is no more such important than it was before.

From my quick peek the layers are blended.

Would I buy a fully sampled, 8 layer DP tomorrow? Most likely!

Are 8 or 9 layers all I would ever want in a DP or sampler? Hmm. I guess, all else being equal, the more the merrier. More samples for each note means chances are I'll hear something slightly different with each playing of that note, likely lowering long-term fatigue due to the unavoidable monotony of a sampled instrument.

IOW, gobs of velocity layers can pretty much get you around the need for the whole round-robin thing (because it's difficult to replay exactly the same velocity twice).

It seems to me that Truekeys is a Pianoteq killer hoping to maintain the illusion of Steinway D authenticity. In other words, no boxy sound of the pianoteq and "Look, there are a bunch of pristine steinway samples on your hard drive!" being inputted to the engine. However, my suspicion is that the slippery slope of playability to conquer Pianoteq comes at the price of authentic Steinway traits being somewhat less than authentic.

Its a marketing masterpiece though. The dopes at Roland shout from the tallest mountain: "Playability (Synthetic) is us!" and are now getting clobbered for it. The snooty snoot Steinway/Roland dealer in my city announced two weeks ago that its packing its bags and heading to some new location.

Dewster,

I'm with you. I think more samples is better. Computer power increasingly much faster than any other technology, so the computer power isn't the problem to handle more and more samples.

Maybe VILabs is keeping it to release a possible version 2 with more layers, since nowadays they are the best one or one of the best.

Great point!

I also believe it's important to judge a software piano/digital piano based primarily on how it sounds, and its technical specifications.

The Nord Piano Library is quite a good example of this. The 'XL'-sized samples typically require twice the memory of the 'L'-sized equivalents, however in my [admittedly limited] experience, they do not necessarily sound twice as good.

Cheers,
James
x

Clearly XL won't sound twice as good as L, but if they sound better at all, then your example supports the opposite of what you just said. Namely, that bigger size, more layers, etc. is better. And we can use those specifications to make judgments about whether one product is better than the other or not.

Unless you meant to say that in your opinion XL doesn't sound better than L at all. Is that what you meant?

I think too much is made of the number of layers as a single specification. The design, implementation, and condition (performance) of the piano sampled is more important.

The number of layers allow the designer to provide distinct timbre changes (breakpoints) across the MIDI 1-127 velocity range. The MIDI velocity values of the timbre break points vary for each key and the number of layers can be different for each key. So the idea is to have enough layers that the timbre changes gradually enough that the player doesn't notice any unexpected abrupt changes in timbre while playing. All sampled pianos interpolate (doesn't have to be linear interpolation) volume changes across the 1-127 MIDI velocity range so volume changes should be smooth (when done correctly) and have 127 levels. More layers only help make shaping the volume interpolation easier, although that could be done with key specific non-linear interpolation too. But not all sampled pianos use timbre interpolation, and that's when more layers can make a difference.

Regardless of how its done, the bottom line is whether or not the timbre (and volume) changes are natural and represent the sound of the original acoustic piano that was sampled. So there are both technical design issues (number of layers and interpolation methods if and when used) and implementation issues (all of the samples collected have to be assigned to different MIDI velocity ranges that can be different for each key and matched to the original acoustic piano). The implementation is a huge job regardless of how good the design may be. If volume or timbre variations are non-monotonic (i.e. volume decreases or timbre becomes more mellow when velocity increases) or there are large changes in volume or timbre at layer breakpoints, then individual notes will stick out at particular velocity levels and ruin even great designs. There are examples of bad implementation and excellent implementation.

So my point is that while design technique and number of layers is important, implementation is even more important. In my opinion, the Vintage D and Ivory II American D are examples of excellent design and implementation, which produces pianos that sound very authentic and are very playable. The Ivory II American D is interesting because it includes pre-defined sample sets ranging from 4 layers to 20 layers if you want to make comparisons based on number of layers (using its design techniques). The Vintage D has fewer layers (13 if I remember correctly) but uses different design techniques. I don't hear any "layer-related" sound or playability issues with either of them. I prefer the Vintage D because I prefer the harmonic richness of the really magnificent Steinway D from which it was sampled, and the broader timbral range over which it can be "voiced" (to use the acoustic terminology) while still sounding authentic. Whether that is a function of the Vintage D's software design, its sampling techniques and implementation, or the original acoustic piano itself I'm not certain (although I lean toward the latter reason), but note that it has a broader authentic timbral range even though it has less layers than the Ivory II.

My bottom line - the design technology and care in implementation (sample recording and processing) used by the Vintage D and Ivory II American D have reached a point that playability is no longer an issue provided you have the right computer resources to completely support their full capabilities - which is different between the two - the Vintage D is CPU intensive and the Ivory II is disk intensive (use an SSD). At that point, you choose by the sonic characteristics of which original acoustic piano you like better (both appear to have been exquisitely prepared prior to sampling).

I'm now trying to determine (in my opinion) if the TrueKeys American D falls into the same category as the Vintage D and Ivory II American D.

Might be tempted to make use of this offer, isn't the piano usually selling at double the reduced price? I actually like the piano voices of my current digital and I'm not sure a software piano would add anything noteworthy to them.

Can a good pianist maybe make even a faulty piano (software, digital or acoustic) sound good?

Yes, I think we can both appreciate this, however to someone less knowledgeable it is not such an illogical assumption.



Specifications are useful to a certain extent, however I would never recommend customers make a purchasing decision based solely on numbers alone.



No, that's not what I meant, nor what I said. However, now that you mention it, I do recall some Nord users preferring the 'L' version (over the 'XL' version) of the Fazioli grand when it was released last year.

Cheers,
James
x

Macy,

Vintage D uses more CPU because the samples are loss-less compressed, reducing the size by 50%. When trig a sample, first it need to be uncompressed. If the CPU power is the problem, we can batch-resave all the samples uncompressed. This reduce the CPU but double the size lib.

The discussion is getting very interested. I didn't find any other technical discussion about True Keys up to now.

It is lossLESS compression - no loss of fidelity at all.

Greg.