My New Rhodes: Anyone Down for a Large Free Sampleset?

[Funkybot]

Hey hollowsun, I'll try and send you that initial test of the samples by midweek. My DSL line isn't up yet, but if anything I could always just send them out from a friend's place. I also think a smaller free version is probably a good idea.

[hollowsun]

Hey hollowsun, I'll try and send you that initial test of the samples by midweek. My DSL line isn't up yet, but if anything I could always just send them out from a friend's place.

Go for it.

If nothing else, just do one full-length note at a few different velocities for now (C3 for example) and email those to me so that I know we're in the right ballpark.

I also think a smaller free version is probably a good idea.

Definitely.

These fully-fledged, multi-layered, 24-bit velocity-switched-samples-per-note libraries with release samples, etc., are all well and good in an expensive payware product but for a freebie, let's see how we can compromise without affecting the overall quality unduly.

It can be done!


Steve

[Failed Muso]

I really can't understand some of the naysayers in this thread. Two guys have offered to create a free sample set, and all some people can do is flame, bitch and moan. Some have even tried to question Steve's 20 odd years of experience in the business. Trust me, what this guy doesn't know isn't worth knowing.

The phrase "Teaching Granny to suck eggs" springs to mind

I agree, there are some good VSTi's out there. Elektrik Piano is quite excellent, but as other VSTi's do, it's another drain on my CPU. I can load up Steve's CP70, and still get loads more instruments on my S6000 or Z8 with no reduction in the performance of my rig.

As for the S5/6000 being "out of date", I'd best call Robbie Williams and tell him his 3 S6000's that he runs his entire live set from are redundant. Best tell every other pro studio in the world that they need to get new equipment in !! Dear oh dear

Steve, as ever, you know my services are always at your disposal, just give me the nod.

For one, I am really looking forward to this sound set

[LIMITAPROACHINGINFINITY]

I would love this soundset more than anything~!

[hollowsun]

Steve's 20 odd years of experience in the business

And it's been 20 very odd years!


Steve

[egbert]

Some have even tried to question Steve's 20 odd years of experience in the business. Trust me, what this guy doesn't know isn't worth knowing.

Steve is that you?

The phrase "Teaching Granny to suck eggs" springs to mind

That's funny - the phrase that came to my mind was FOAD

[Failed Muso]

[egbert]

Well look - perhaps we all ought to get down on one knee and acknowledge that the pinnacle of the fine art of sampling was achieved 20 years ago. That 8-bit vintage quality - what those guys achieved fitting a Bosendorfer on a floppy - fantastic stuff. This is why the Fairlight and the Synclavier and the Emulator are the most popular samplers today. You just can't beat experience.

If anyone seriously doubts the benefits of recording at 24 bits on good gear there must be several hundred threads on the matter on this site and others.

[kevink]

Don't get me wrong - these modelled products are fine but some/many people would feel more comfortable playing samples from 'the real thing' and Funkybot (and I) are offering this for free. Can't see the problem myself!

And don't get me wrong, either, I think that's a great idea. I just took issue with the assertion that a soundfont is "the real thing", while a physical model is not. Go forth and samplify!

[hollowsun]

If anyone seriously doubts the benefits of recording at 24 bits on good gear there must be several hundred threads on the matter on this site and others.

I don't doubt or dispute the benefits of 24-bit recording but for individual, full level samples such as these, there is little practical or audible benefits to be had, especially for this instrument and I'd rather keep it 16-bit for any number of practical reasons, not least of which is to create an efficient and unbloated soundset.

Depending on the formats this ultimately ends up in (probably mostly s/w samplers), I would imagine it is better and more practical to be able to have more effects and virtual instruments running alongside it than (maybe) a marginal improvement of the final few milliseconds of a decaying note (that will largely be buried in the Rhodes' own limited signal-to-noise ratio and/or the mix anyway).

I am not 'anti-spec' or some Luddite old fart living in the past but my background and experience has taught me how best to make certain decisions to best optimise a set of samples for the job in hand and in my opinion, I think that recording a Rhodes at 44/16 is quite (if not more than) sufficient to effectively capture that sound. Nothing more or less.


Best Regards,


Steve

[Funkybot]

Hey Hollowsun, the test files should reach your mailbox in a few minutes now (twenty or so at most). And your absolutely right about the signal to noise ratio especially in regards to the decay of the note. It was hard to tell at some points where the note decay ended underneath the noise (which was probably flooring at around the -50db range). Anyway, check the test files I sent you out and let me know whats up. You can respond from the e-mail address I sent them to you from, but if you could also CC me to that same name/my same KVR name at gmail.com, I'd appreciate it as I slowly try and transition to gmail as my main e-mail client.

[Sascha Franck]

Again, I can only agree with Steve here. Very often the advances of 24bit sampling simply won't pay of as much as one would wish.

And then, judging from the sonic quality of his very own Hollowsun samples, there's nothing else to be said but: Excellent!

Allright, now bring on the samples

[hollowsun]

Hey Hollowsun, the test files should reach your mailbox in a few minutes now (twenty or so at most).

Yep... got 'em. Can't look/listen right now but will do so later.

And your absolutely right about the signal to noise ratio especially in regards to the decay of the note. It was hard to tell at some points where the note decay ended underneath the noise (which was probably flooring at around the -50db range).

Ha!

Anyway, check the test files I sent you out and let me know whats up. You can respond from the e-mail address I sent them to you from, but if you could also CC me to that same name/my same KVR name at gmail.com, I'd appreciate it as I slowly try and transition to gmail as my main e-mail client.

Ok... will do. More later.

Thanks


Steve

[meeks]

I'm very excited about these!!!

[but I also learned from that little flame up ]

ohh, the tension....I hope it lasts. -WW

[egbert]

I don't doubt or dispute the benefits of 24-bit recording but for individual, full level samples such as these, there is little practical or audible benefits to be had, especially for this instrument and I'd rather keep it 16-bit for any number of practical reasons, not least of which is to create an efficient and unbloated soundset.

Steve - you and I don't have a quarrel. I am mindful that this discussion is on KVR where most of us use software samplers which are typically 24 bit and have been since the late 90s. Cost is no impediment to using 24 bit quality - SFZ is a free top quality sample player which admits 24 bit samples and SFZed is a free instument editor which allows quick instument building from multisample sets.

As far as bloating goes - long decay 24 bit samples typically compress further than 16 bit samples so downloads are not necessarily that much bigger. Many soft samplers are most CPU efficient when they load samples in 32 bit float mode in any case so there is no RAM saving when comparing 16 bit and 24 bit linear samples.

As far as perceivable quality in a mix - sure most if not all listeners could not tell the difference. The same probably applies to a million other production choices - to offer some analog analogies - you could record at 7 1/2"/ sec instead of 15"/sec and save tape or use a cheaper preamp etc etc. Most people opt to preserve quality when they have the chance. I don't think this is a controversial idea.

The Rhodes is an instrument which has a timbre which varies greatly with velocity. Leafjeffs excellent soundfont has three velocity levels and 16 bit samples and it is already up there for anyone to use.

The initial impression I formed was that this project was going to go for more character/quality. I therefore fully expected the sampling would involve multiple velocity levels.

While it is possible to reset your equipment for recording each velocity level and try to record as close as possible to fullscale for your pp samples, it seems that the most common approach taken by most pro instrument samplers these days is to record all samples in a piano set at the same gain setting.

With this approach the quiet samples never get anywhere near full scale. They are therefore recorded at much lower resolution if recorded at 16 bit as I said previously.

One corrective could be to record everything at 24 bit, then normalise and dither to 16 bit for the sample set. The lower levels of the quiet samples can simply be programmed in most samplers - eg pp sample set is scaled down to 20 or so dB below unity gain.

I don't think it is inappropriate to have a discussion of the best approach to sampling instruments - I have no reason to think that you disagree. People are on this forum to exchange information. In any case, neither you nor I will be doing the sampling - we are just reponding to Funkybot's requests for assistance.