Playing a sample pitched down is essentially lowering its sample rate. A sample at 44.1 played back down an octave is really a 22khz sample with a nyquist freq below 11khz so essentially applying a gentle and continuously deeper low pass filter the further you pitch the sample down. Filter something enough and it begins to sound more 'sine' like, thus your observation that the pitched down samples have more pleasant sine-iness? But the actual sound of this effect would also depend on the algo your particular sampler is using to interpolate between the lower sample rate and the reproduction of analog sound at your 44.1(or higher) DAC.
you're definetly right here. Good point about the algorithm. I guess in general when you pitch down the algorithm recreates some of the high frequencies, but if you pitch up it doesn't recreate low frequencies, maybe that's why it sounds better to pitch down most of the times. Also some algorithms actually sound worse when you pitch down, because they don't recreate any harmonics.
So C1 is a note at a halved sample rate than C2 when both are recorded at 44.1? Your idea seems to be metaphorically mixing clock domains. Here the samples are at 44.1 and changing pitch doesn't change the sample rate, and talking about filtering is making assumptions about the unknown algorithm used. Maybe you're getting to the sample rate conversion method to change pitch? That's the simplest approach that nobody uses. In my day job we often use wavelets to attack this. Anyhow I don't follow Christian's pitch down argument. Historically in fact we know that people prefer pitching _up_. In the 1800's orchestras were tuning to not A440 but a few cents lower. Nowadays if you listen to 2nd or 3rd tier orchestras (or kids) you'll hear them sharpen up over the course of the concert. This seems (to me) to go along with the increase in tempi over the same time, which also came from the establishment of the great European conservatories. I think in the 1700's the tempi were much slower. But with the industrial revolution we seemed to get sharper and faster. Louder too, especially as concert halls got bigger. I think it's psychological - Christian is a mile-a-minute kind of guy and talks about the annoyance of the computer, so perhaps a fat, pitched down sound is deeply satisfying for you Christian? It doesn't for me honestly, the felt piano samples sounds too much like fat, warm Tribbles. Doesn't mean I don't like it, just not as the centerpiece of my meal. Music is all about movement and contrast, I could see using the Felt Piano in the right context, but I wouldn't use it as my daily piano sample. FWIW
@@DanMcLaughlin I'm not quite sure if the up moving frequency of the A440 is actually about the same thing that the down pitched felt piano is. If you look up the whole A432 discussion (which I don't want to get into too much) you'll find that many people actually prefer tuning down (note the word "tuning" not "pitching"). In my oppinion it all comes down to taste, but most of the times when you're talking about pitching, people tend to prefer pitching down. and yes you're right, there's a significant difference between a felt piano and a grand for example. one thing I noticed in that regard tho is "trend"
Dan: Yes, the sample rate (time domain) is the same (and would have to be -- though when listening to a sample recorded at 44.1k played back at a sample rate of 22k, it would be pretty close to an octave lower) -- I agree that some of this may be psychological as you say but perhaps also a result of the processing. As the audible frequency of a sample ascends, it can contain frequencies above the sample rate which causes aliasing, (aliased to the target frequency w artifacts, "harshness"). One has to actually apply a low pass filter while sampling and playing back to diminish this. At lower frequencies this is much less a factor and the sound may seem 'warmer...?'
@@DanMcLaughlinIf you record a sample at 44.1kHz, the highest frequency it can contain is 22kHz. That's what the Nyquist-Shannon Theorem states. To make sure there are no frequencies above 22kHz in the recording, the signal is filtered with a steep low pass filter that cuts everyhting above 22kHz. So there inherently is a filter at play. This filter is called the anti-aliasing filter, because frequencies above it would cause aliasing. If you now pitch this clip down, you are pitching every frequency component. So 22kHz might now be 11kHz. But the result won't contain anything at 15 kHz, because the original sample didn't contain any 30kHz. So you essesntially moved the anti aliasing filter downwards. The sample rate is basically halved. That doesn't mean that you can't save that file at 44.1kHz again … but you won't gain any additional Information. Another way to think about it: Pitching down means stretching. So if you take a one second audio clip recorded at 44.1kHz, it contains 44100 individual samples. Stretching these 44100 samples over a period of two seconds would mean you are left with 22050 samples per second. Sample rate halved. So the gist of this is: There is a difference between the file Format and the actual Information that file contains.
When you sample a minor second on the piano the overtones of both notes clash with eachother more heavily, because the strings vibrate sympathetically to the fundamental frequencies that are played. When you play a minor second with two seperately recorded samples the overtones of the samples still clash, however there are no strings amplifiying the clash of overtones.
My mother was an incredible piano player, she estudied with Nadia Boulanger and used to practice at home at least six or seven hours per day, so before I learned to walk, I used to crawl and seat under her concert piano while rehearsing. So for me, piano sounds come from above in splendid and spreading waves of harmonics. Pure magic. I've been following closely your diving into harmonics content in piano sounds and I totally agree with your efforts to find the Holy Grail of waves behaviour when sampling them. But there's something I still miss from your scientific approach, and it is the inmersive sound quality missing in most piano libraries. I own Spitfire's Hans Zimmer Piano and probably it`s the closest I've heard to these memories seating for more than seventy years in the Hypothalamus of my brain, so I want to encourage you to keep on researching for the ultimate experience of translating the soul of the piano to the binary codes.
It's hard for me to comment on this subject, as I agree with your perspective. My philosophy of sound reproduction is that it just isn't the same thing as being in a room (good or bad room) with a 'real' instrument (including electric/electrified ones). A 'perfectly captured' violin coming out of an eight inch speaker is a representation of a violin, and not even close to being in a room with the same instrument in the same capable hands. The same holds true for an electric guitar and amp. A 'live' instrument resonates in all directions from many sources that sampling, compressing, equalizing and pumping out of a speaker just can't match. Having said that, I still think we need to approximate as accurately as possible, as well as 'modify' the sampled sounds, as thee are new and interesting. I guess what I mean is that we all need to be in rooms with musicians interacting with us (ensemble and audience) more frequently. Music is a human experience to be shared through both media. My fear is that, like the Internet, we become isolated in all our experiences.
I totally agree with@@tdubveedub's comment. Music is a bunch of air molecules moving in all directions and interacting with each others, so there's nothing like the real thing, no matter how much DSP power you can stock to convert phisics into digital emulations. But, there's always a "but", we are facing now a second generation of folks who trained their ears through computer speakers or headphones with low quality and distorted sound curves, so we, composers and sound designers have to relay heavily on the Art of Sampling and all the research and innovation that folks like Spitfire and others can assemble to provide us with dynamic snapshots of sound sources which mimic as best as posible the real thing.
@@teddybautista1805 I agree, and it is too bad that there are so many folks missing out on the ecstasy of chamber and orchestral music. Did you see the reaction of the attendees at a Star Wars convention when John Williams appeared with a complete orchestra and performed a medley? They went wild!
Very true. I've been listening to quite a lot of sampled pianos during the last year. Last week I spent in a house with a big old real piano. And boy oh boy did it 100% exceed those piano samples. The way each key uniquely has its own tone, and the way the whole thing resonates, the wood, the infinite variations in attack / transient response depending on how hard, soft, slow, fast you hit each note: totally totally different experience. It's somewhat like the difference between seeing a picture of a large monument, vs being there, seeing it IRL. Bruce Swedien, famed recording engineer (Michael Jackson, Quincy Jones, Count Basie, and many more) has said again and again that EVERYONE wanting to work with sound and music, should pursue seeing and hearing real instruments being playedz Orchestral music. African drumming. Live music with live drums, etc. I've taken that advice, and don't regret it for a second! I also listen to lots of vinyl from the pre-digital era. After having listened to lots of streaming media, it's often a shock to me, how huge difference there is.
I remember reading about Trent Reznor recording his guitar pieces and then slowing them down to the desired notes while getting that huge sound, back in the day. Makes sense to me.
Agree, that's wrong to say, what he did in the vid is add more bass content but kill midrange texture. More bass = louder = sounds better. More bass partly matches the equal loudness curve (low freqs). Then in a certain context when arranging/composing u may prefer one sound or another, typical pop song pad sound is soft and gentle without too many odd harmonics, lets u here the reverb more and usually sits better in a mix. There's Robert Rich and there's Tim Hecker...
Nice one. For sound design I record 196Khz and then you can pitch down to 5 octaves and still get 44.1Khz sampling rate. Record 1 minute and when pitched down you get 5 minutes soundscape. Add a massive reverb and you can do a ambient track using sounds around us.
Surely you can only pitch down two octaves and retain a 44K sample rate? Every time you go down an octave, you halve the sample rate. Which isn't to say that it doesn't still sound awesome...
@@charlesgaskell5899 You are correct. In order to pitch a sample down 5 octaves and end up with a 44,100 Hz sample rate, it would be necessary to start with a sample recorded at 1,411,200 Hz.
@@GlenBerry Plus, when you look up the specs of any decent microphone, they don't go anywhere near so you can record such high frequencies. Plus, even if you had a microphone to record those freq God knows what you record, a lot of hum anyway.
@@aileaiho768 I suspect you're confusing sample rate with pitch frequency. Sample rate is a feature of AD conversion, microphones don't have a sample rate unless they have an on-board AD converter (such as USB mics).
Just found this channel from my recommended. I’m a musician who has been playing for over a decade and I appreciate real informal videos about music &the science behind it. It’s very refreshing to take a break from modern mainstream bull and dive back into real music and what makes it. Especially as a pianist, I love informative vids. I saw all the gear and I knew I would enjoy this vid. I also liked the editing, how you moved from place to place and made clean transitions. You sir have earned a new fan and a subscriber, keep up the good work.
amazing job! so refreshing to hear a concept I had never heard of. So many videos on production are just the same and everything is being restated. Please make more interesting bits like this!
Christian - you are a unique voice in the world of sound and music, with a wonderful holistic grasp of all things from the technical to the emotional. This really resonates (and I use the word deliberately!) with me. Thanks so much for bringing us your insights!
The memory limitations of earlier multi-zone samplers, such as the SP-1200, led producers to record the samples pitched up (turntable varispeed, another sampler dedicated to transposition etc) and use the samplers transposition function to pitch the samples down and it did indeed give it a 'phat' sound. I know this was standard practice for the Beastie Boys, De La Soul, Bomb Squad/Public Enemy. I'm assuming Kraftwerk and Art of Noise were confronted with the same limitations. Brian Eno/Daniel Lanois' "Apollo" makes heavy use of the Lexicon Prime Time's sample rate pitch-shift down function and tape varispeed, funnily enough the standout track 'An Ending' is a bunch of CS-80 sine waves recorded hot to tape and you can hear the pleasing tones of "missing information" all over it. I noticed how more top-40 sample-based stuff start to sound very sterile once memory considerations were no longer a worry, but a lot of more alternative artists still used the degradation techniques. The most cheeky example I've ever seen of pitching down a sample and making it 'phatter' is 'Feel It All Around' by Washed Out which is just Gary Low's 'I Want You' pitched down a tritone. Had everyone wondering what crazy production techniques he was using until someone found he just typed in -6 on a rare italo disco song.
Whew! I am brand new to your TH-cam channel. What a great introduction for me. This video makes complete sense to me all the way through. I get it why closely-spaced notes sound differently when samples of them are played from when the original acoustic instrument is being played (e.g. C beating against C#). There's a real difference there that depends on the physics of harmonic vibration in those physical objects (metal strings and wood) versus digital facsimiles. This may be related to the reasons that some people prefer the sounds they can get from analog synthesizers over those produced by digital synths or VST emulations. The physics just isn't quite the same, particularly in polyphonic playing, even within purely electronic instruments. There are some people, also, who prefer "lo-fi" or "8-bit" audio over high-resolution digital sounds, at least for some musical applications. Those same folks may prefer slowed-down (slightly lower-res) samples over real-time hi-res notes for similar reasons. I have never really understood "fatness" (or "phatness") as applied to synthesizer voices. Now I can see what is meant, I think, a little better. I would go with selling some carefully curated Pianobook packages commercially through Spitfire while keeping unpackaged samples (or less extensively curated packages) available for free.
Christian, this channel just blew my mind. It's infinitely more intelligent and in-depth than any other music channel I find searching for solutions or inspiration for my music!
Keep Piano Book Free 100%. It's a godsend for composers who can't afford many great sounds. it gives rise to a great community beyond the market mechanisms which allow it to organically grow without an agenda. it also gets people sharing great instruments for the future, capturing the present for the future like a time capsule. yeah, please keep it free. great idea you had too with offering a greatest hits for purchase each year, perhaps also with a bit wider functionality, more velocity layers, etc... thank you for starting this!
I wonder if this what contributes to the preferred character of the sp1200 in hip hop. The limited amount of storage means records are played fast to record more and then stretched or pitched down..
Oh my god. You made an entire (incredibly detailed) video to answer my flippant comment on the last pianobook video. Bless you. For a bit of context, I have near zero experience sampling pianos (which is why I'm here, to learn), but some experience sampling synthesizers. I assumed, incorrectly, that sampling the cycle of the fifths as opposed to chromatically was to save time, in the notion of an MVP. Thank you for explaining the intent behind it. My initial reaction to pitching notes down (which spurred my aforementioned comment) comes from what I was taught in sampling synthesizers. Given that synths are more complex than sine waves at various speeds, pitching them down means reducing the resolution of the sample. It often sounds better to pitch up synth samples as this only increases resolution while decreasing loop length. Having now tried it for myself, I see how much better it is to pitch down piano samples than pitch them up. I'd warrant that this kind of 'cleans' the sample by reducing its resolution and reducing it's complexity. The lower notes when played on a real piano are very complex, while pitched down higher notes are simpler sine waves that have reduced complexity. As for your question, I wanted to ask what polls you're referencing? I stumbled onto this video series through TH-cam's recommended section, so I don't know if you're pointing to a Facebook, Twitter post or something? As for the final product and delivery, I wonder how difficult it would be to adapt the code for the Labs plugin? Perhaps in the same way Spitfire delivers Labs releases you could deliver pianos for the pianobook. I don't mean it as a download within the current Labs, I mean it as a copy of Labs. In the way Labs has many instruments, Pianobook as a plugin would have many pianos. The downside of this is the Labs plugin code is likely proprietary and reduces the community's ability to adapt it, etc. Maybe Spitfire should just open source the Labs plugin code any ways? I think you could/should charge a small amount for the best curated pianos- splitting the cost appropriately between the creator/sampler and Spitfire, if Spitfire is lending programing work.
this is a revolutionary concept in my opinion. i love everything about it. i like the rough suggestion you lay out at the end. looking forward to seeing how things progress.
Fantastic insights on sampling. Timely, too. I just finished sampling my electric guitar for a project, and noticed how much better notes played on the high B and E strings sounded when pitched way down to the lower octaves, vs the low low E string. There is something about it being stretched out that allows appreciation for some of that hidden detail, as if you zoomed in on a distant landscape.
Every sound in the video, from the SM7b voice recording, through all the piano and sampled pianos, all of it sounded...really good. Even the wind sounds into the outdoor rig sounded musical. And then I heard the words “Spitfire Audio” and realized why I simply must find more money. I have the free LABS...most of the things in there sound better than anything I have in NI Komplete. I just need the $$$ and some GAS counselling.
Hey Christian, loved this one and been mulling it over this weekend. I think maybe our preference for pitch shifting downwards rather than up comes down simply to our love of resonant sound the physical vibrational aspect of music. I feel that when we hear the sound of a piano, we hear the physical body of the instrument in resonance. This resonance is a quality of the tone that we are drawn to and as we shift that sound downwards it as though the pianos size and vibrational depth is increased. The tonal weight below and around the vibrating string is heightened. Of course the pleasing aspect of this shifting is limited as we soon notice the sound is being artificially altered which fast becomes a negative trait. I notice the importance of this physical resonance to the sound of the piano when moving to the upper ranges. When we play a high note on the piano, the instrument still has this large resonant body to give a feeling of depth and weight to that note. If we were to achieve this same note by speeding up a sample of a lower note, we would be in effect shrinking the size of this resonant chamber which I feel we inherently interpret as a negative change in the sound. I do think as well this changes from instrument to instrument. Sometimes I prefer the sound a guitar slightly sped up, kick drum or snare (I guess we tighten up attack of the transients which for some sounds would be preferable). Synthesisers as well seem more immune to pitch shifting than other instruments. Perhaps this is because we as listeners are not perceiving the sound as a resonant physical body, and therefore no perceived reality is being contradicted when we artificiality change the sound. Regarding your sampled piano in particular, I think that when considering why the sampled version sounds more pleasing and more suitable than the real deal, that there are considerations of genre and taste here to think about, rather than a purely objective comparison. I think we can agree that the defining characteristic of the modern zeitgeist piano style is a desired reduction in harmonic content. Whether by playing whisper soft or placing muting felts over the strings, essentially the goal is usually to reduce that hi-frequency resonance that a piano naturally has. When pitching sound downwards, the tone is dulled and the transient attack reduced, helping us greater achieve this lovely soft ghostly sound. Whilst dulling a tone through pitch shifting really lends itself to Satie, I’m not sure it would apply as well to say, Rachmaninov. Secondly, I feel that what we love to hear in sound and instruments is consistency across their tonal ranges. No lumps and bumps. If we take one note and spread that tonally through pitch shifting, each of these new notes share the same tonal fingerprint and I feel its this consistency that our ears love.
Your delivery in this reminded me very much of James Burke's connections! (not sure if you are old enough for that to have been an influence - very much meant as a compliment) :-)
I think giving everyone the opportunity to use Pianobooks for free would be nice as it's been a community effort. Some people in the comments have talked about open-source software models, which I think is a pretty good idea. If Pianobooks also get a curated, paid version, it would make sense for the funds to be reinvested in the project. On the topic of file formats, I don't know what's the standard at the moment, but it would be helpful to have formats compatible with Ableton - it recognizes EXS with some annoying steps, and it can also recognize Ableton Packs seamlessly. Sadly they're a bit trickier to put together - they can only be put together by companies, not individuals. Maybe I'm weird, but I really hate having to use plugins, as Live Suite and Max for Live effects and instruments cover pretty much everything I need in a more elegant way.
I've been thinking along the same lines. Currently I'm leaning towards something more 'universal' like .sfz soundfont files. Most daw/samplers will recognize them at least on some level.
Sharing everything - open source and free would be really great for the musical community and i'm totally down with once its set and curated to offer a paid version of it as long as its not expensive, id think 60$ or less would be a fair. And as to what to do with the money - i think half should be used to keep the project going open source because it still takes money keep it going - its going to need a community website and forum and plenty of bandwith for downloading - its a reinvestment into the project and the other half should be used to get musical instruments into schools in both the UK and the USA that do not have them so that the next generation can learn and keep the project going for decades to come.
This is absolutely fascinating. Thank you so much for sharing your work and ideas! That final piano pad sound that you landed on was superb and I wish I had it!
For an idea as to how to distribute to Pianobook, see the Blender Foundation's model. The Blender Cloud is a paid service that distributes training and various other things, and they frequently crowdfund things and take donations, but the main final product is packaged, and completely free and open source. Anyone can contribute anything, start their own branches, and go into detail, completely changing the product if they need to, but the blender cloud offers training made by qualified professionals. It may not seem immediately relevant, but the models followed by this and other FOSS projects are definitely worth looking into, and it may be worth asking questions of those involved in them to see what they do/think, to get a good background of undeniably similar things to what you're doing onto which you can build your own model.
I think the Piano Book should be the sort of "open source" form as it has been from the start until today. There are just sooo many commercialised pianos out there that I don't really see the point of adding more to the pile. So keeping it within the community and giving it away freely could be the start of building a bigger community for the next instrument that's going to be sampled. And I really like where you're going with all this. It is a fantastic project from the very start.
I am so glad I stumbled onto this video. At one point my face was literal inches away from the screen, felt like a revelation of sorts. The appeal of the whole lo-fi community, DJ Screw's influence, hell, even bass-oriented music in general, started making a lot more sense all of a sudden. I always wondered why I enjoyed the sound of pitched-down samples so much more, and always gravitated towards more "sine-like" instrumentation in my productions. Maybe the lack of "information" that's contained within those sounds makes it more pleasant to perceive? Maybe because the most pure of tones are essentially sine waves? Either way, amazing video, it's an instant sub from me.
What your talking about for piano book is how the FOSS (Free Open Source Software) world works. Everything is out there free to download and use, but there are companies that offer packaged version, support, etc for a cost. A lot of companies who use FOSS software as a way to contribute will let their programmers spend some time working a FOSS project as their way to contribute back. So I would say look at the FOSS model for ideas on piano book.
Hi Christian, the fact that AS you talked about a charitable endeavour with piano book you got a rainbow says it all😀. A charity based around music would be great, piano lessons for blind kids, that kind of thing would be lovely. I think you got it right as a community project that's free and then the best of 2019/2020 etc etc on Spitfire with the money going to charity for the polished version.
3:59 anyone know what that controllers left to the computer keyboard are? I faintly remember something like this being announced a few years ago, individual control bits you can stick together to do form a custom controller, but I cannot remember the product or company name.
first of all this is an amazing video dealing with some many interesting aspects of sampling, instrument playing, technology and psychoacoustics. so thanks ever so much! regarding painobook, something about your openness and creating this community makes me want to keep it a community thing. i.e concentrate efforts to keeping this knowledge, open source in a way, open to endless manipulations and upgrades done by the members of the community. if going for the "best of boxset" I can see some kind of model of 50% towards keeping the initiative a live and well maintained (investment in a future interface or someone to upload and take care of the site for example) and 50% going to charity (one for example that makes music lessons more accessible for kids). I know I'll be happy to contribute to such a cause.
The word Genius comes to mind. You Sir, Are what the sample world needs. I use drum and percussion samples. And tuning them down does make them sound better. But you know why! Wow! Just Wow!
Could someone please explain what he is doing from 6:25 to 6:50??? I'm using Ableton I don't know how I could achieve this sound, I think it would be with the Sampler device? Thank you
He just sampled a piano note with a sustain pedal, so in ableton you can use the loop option in the simpler to loop the tail of the sample so it never stops, with medium release, then add some plate reverb and cut off a bit the lows u should get something similiar
Hi Christian. I'm still working on my own piano samples to contribute. I guess my first thought would be just to make it all open, but that's from my perspective as someone who has a little to give, but much more to gain. It's easy to give something that's isn't as good as something else that's better. Perhaps a minimum entry "fee" of one piano gains one access to others, but perhaps not all? Or, roll it into Labs and make a pianobook instrument with options for other pianos to be added through that system.
OMG! LOL! I watched this video twice before the lightbulb went on and I realised that we worked together a while back when you were working with Marcus (can't remember his surname!)! Nice to see you doing well Christian. Great video!
Wrt your questions about reconciling community contribution with commercialisation: have you explored the use of Creative Commons licenses? A possible application is where contributors submit samples either as a CC-BY-NC (non-commercial) or CC-BY (commercial allowed). Others can utilise the samples as per the license. You're also free to curate the samples that allow commercialisation. Over at a site ccmixter.org, they have adopted a variation called CCPlus, where it's like a royalty payment arrangement if your non-commercial samples are requested for commercial use. Note: you'd still have to deal with a potential issue of a copyright dispute wrt a contributed sample, but if you're going for commercialisation then I suppose treat the rights clearances process as no different from any commissioned project.
That was a really nice take on the topic, thank you! I feel kind of bad about it, but i have to take one thing away from you: we don't hear the low end before the top end. The sensory cells that perceive high frequencies are located at the front end of the cochlea, while lower frequencies are percieved further inside the cochlea. This means, we hear the top end a couple of microseconds before the low end. Not that anyone could actually hear that. The difference is way too small.
On the topic of free vs paid product for Pianobook, this certainly brings to mind open source software. It seems that Pianobook has a few key values: - Sharing/community: Contributors can share what they have and use what others share. - Education: Learning sampling, preparation and production techniques. - Creativity: Sparking ideas with sounds to which you might not have access, otherwise. These are all values that open source software includes. As long as the community is engaged, this could be self-sustaining. Why not filter decisions through these values? This could be turned into a commercial product later with simplicity being what people are paying for, as you suggest, Christian. I've seen a few (software) projects become sour and go off the rails by considering commercial avenues too soon. It tends to narrow the thinking and prevent the values above from being the main drivers. As always, great stuff.
Warren Harrison sure thanks Warren the initial motivation was me making a commercial library and showing how I did it, and then I thought I could involve the community so I’m rolling back the original idea as the commercial aspect doesn’t fit for me anymore. I think your advice is sound though.
2:12 I prefer the live piano sound. The sampled version sounds much warmer but as if it's missing the full range of freq. especially in the high end. The resonance on the live piano has been greatly stripped on your sampled sound, initially. Then when you 'phatten' the sound it's starting to get better. That's how my ears/brain interpret. Great video btw.
Love the experiment. Noticed the same when I pitched a piano down a few octaves. I'd like to see the experiment done with higher resolution recordings tho. My theory is that when you record at 96kHz and pitch it down you'll get a brighter sound than with a 44.1kHz recording treated the same way, since you are basically capturing more high end frequencies.
It depends on the ability of the microphone to capture high-range frequencies - the highest pitch that can reliably captured is half the sample rate, and response curves are normally given up to about 22kHz. If the microphone can't respond to frequencies higher than that, then it doesn't matter what sample rate you've used, those frequencies won't be present
@@charlesgaskell5899 that's true aswell. of course you should take that into consideration. Still I find it interesting to see how big of a difference there is comparing the end results side by side
I like the idea of Piano Book being a commercial product. On the one hand I can see how if successful it could open a new direction for Spitfire as a company in regards to its products and services; and on the other seeing and maybe even being involved in the creation of an instrument library is of huge educational value as a composer. How you would marry that all up with rewarding the contributors on a commercial venture I have no idea, but I think its worth exploring further.
Love what you're doing here! I was only dimly aware of it until now. Pianos are amazing instruments. By the way, I'm going to remember that reply about rainbows, ready for when I move north in a few weeks (going to Aberdeenshire).
Fascinating. I would hypothesize that the upper harmonics of low-frequency notes interfere with the fundamentals of higher notes, so cutting off the high end of your low notes makes them blend better.
When the Casio Sk-1 sampler came out back in the 80's, it had 2.5 seconds of sample time. So I would sample records at 45 rpm or 78 rpm and play back the sample on the lower keys to recover the original tempo. The sound was wider and thicker and crunchier since the sample rate was 8 bit. But you are right...anything sampled then played back pitched down is phat to death. Although some like that scratchy alias effect, I try to avoid it at all cost. I dig dirty samples but aliasing turns a good sample into trash.
Even as you were talking, especially about the different contributors, I couldn't help thinking of the way that a lot of open source software is managed. Might be worth looking at some of their models.
I'll tell you what I thought it would be: a site called Piano Book with a Kontakt Library for download and an "About" page. That would already be more than a minimal viable product. You've already gone way the hell beyond that and into space past Voyager 1! I first have to digest this awesome, eye-opening video of yours, but here are some immediate thoughts: a) We could change the name of the community project so that if we go forwards with future projects (please let there be future projects) that maybe do not include pianos, there is a sense of continuity. Music Books, maybe? b) Great explanation for why samples sound bad when you pitch them higher. Another reason for only transposing downwards, is that if you pitch samples upwards, you will hear the shift from one sample to the next much better c) If the project is to be sold through Spitfire Audio, which I am strangely uncomfortable with (I should be stoked; I'm a fanboy), then it shouldn't cost much at all, maybe the price of a double album, and the money should go towards future projects and compensation for your work and your teaching. d) there needs to be some defined parameters for wiggle room, so that people (hopefully there are some women invested in this as well?) can work around any problems their instrument may present, or so that they are able to include elements that they think will make it unique. So, if the particular keys chosen by you, (I forget which ones you chose) contain that one key where there is always a weird buzz that you can't get rid of, you could decide that no, in this particular library it's not Bb and F, but B and F. And if people want to include a particularly interesting IR file, or a sample that otherwise triggers with every note, there is room for that. Does that last one even make sense?
Yep we like the Interpolation of the downpitching more than the original mess, cleaner + kind of more powerful / solid in a way. Relating that to our origin in the uterus is interesting, good reflections. If you want to see how powerful can be that effect get, instead of kontakt, use a emu ultra, a ASR or one of the few first akais.... now THAT is amazing.
the ultimat answer is what feeling you want to achieve in the soul of the listener ... raw sounds of pianos round the world could be free ... what takes time and skill is to process the data the result is always a synthetic sound (you could buy); a piano sounds is different in different rooms even in a analog situation (chamber concerto) the sound will be different depending on the place you sitt. ... its like so many things all is about the context
Interesting discussion! I will always try pitching a sample up and down, usually do pitch down if anything. Sounds like a great project, what about other instruments? or non-instruments...
I am interested in contributing a Wurlitzer grand that I love. But I don't use EXS or Kontakt. I use Grace Sampler, ReaSamplomatic, and occasionally Ableton Sampler. Raw WAV samples is what I prefer, so I can compile them myself. Do you think there would be any interest in compiling for Grace Sampler, or simply providing the Raw WAV samples?
Playing a sample pitched down is essentially lowering its sample rate. A sample at 44.1 played back down an octave is really a 22khz sample with a nyquist freq below 11khz so essentially applying a gentle and continuously deeper low pass filter the further you pitch the sample down. Filter something enough and it begins to sound more 'sine' like, thus your observation that the pitched down samples have more pleasant sine-iness? But the actual sound of this effect would also depend on the algo your particular sampler is using to interpolate between the lower sample rate and the reproduction of analog sound at your 44.1(or higher) DAC.
you're definetly right here. Good point about the algorithm. I guess in general when you pitch down the algorithm recreates some of the high frequencies, but if you pitch up it doesn't recreate low frequencies, maybe that's why it sounds better to pitch down most of the times.
Also some algorithms actually sound worse when you pitch down, because they don't recreate any harmonics.
So C1 is a note at a halved sample rate than C2 when both are recorded at 44.1? Your idea seems to be metaphorically mixing clock domains. Here the samples are at 44.1 and changing pitch doesn't change the sample rate, and talking about filtering is making assumptions about the unknown algorithm used. Maybe you're getting to the sample rate conversion method to change pitch? That's the simplest approach that nobody uses. In my day job we often use wavelets to attack this.
Anyhow I don't follow Christian's pitch down argument. Historically in fact we know that people prefer pitching _up_. In the 1800's orchestras were tuning to not A440 but a few cents lower. Nowadays if you listen to 2nd or 3rd tier orchestras (or kids) you'll hear them sharpen up over the course of the concert. This seems (to me) to go along with the increase in tempi over the same time, which also came from the establishment of the great European conservatories. I think in the 1700's the tempi were much slower. But with the industrial revolution we seemed to get sharper and faster. Louder too, especially as concert halls got bigger.
I think it's psychological - Christian is a mile-a-minute kind of guy and talks about the annoyance of the computer, so perhaps a fat, pitched down sound is deeply satisfying for you Christian? It doesn't for me honestly, the felt piano samples sounds too much like fat, warm Tribbles. Doesn't mean I don't like it, just not as the centerpiece of my meal. Music is all about movement and contrast, I could see using the Felt Piano in the right context, but I wouldn't use it as my daily piano sample.
FWIW
@@DanMcLaughlin I'm not quite sure if the up moving frequency of the A440 is actually about the same thing that the down pitched felt piano is. If you look up the whole A432 discussion (which I don't want to get into too much) you'll find that many people actually prefer tuning down (note the word "tuning" not "pitching"). In my oppinion it all comes down to taste, but most of the times when you're talking about pitching, people tend to prefer pitching down.
and yes you're right, there's a significant difference between a felt piano and a grand for example. one thing I noticed in that regard tho is "trend"
Dan: Yes, the sample rate (time domain) is the same (and would have to be -- though when listening to a sample recorded at 44.1k played back at a sample rate of 22k, it would be pretty close to an octave lower) -- I agree that some of this may be psychological as you say but perhaps also a result of the processing. As the audible frequency of a sample ascends, it can contain frequencies above the sample rate which causes aliasing, (aliased to the target frequency w artifacts, "harshness"). One has to actually apply a low pass filter while sampling and playing back to diminish this. At lower frequencies this is much less a factor and the sound may seem 'warmer...?'
@@DanMcLaughlinIf you record a sample at 44.1kHz, the highest frequency it can contain is 22kHz. That's what the Nyquist-Shannon Theorem states. To make sure there are no frequencies above 22kHz in the recording, the signal is filtered with a steep low pass filter that cuts everyhting above 22kHz. So there inherently is a filter at play. This filter is called the anti-aliasing filter, because frequencies above it would cause aliasing.
If you now pitch this clip down, you are pitching every frequency component. So 22kHz might now be 11kHz. But the result won't contain anything at 15 kHz, because the original sample didn't contain any 30kHz. So you essesntially moved the anti aliasing filter downwards. The sample rate is basically halved. That doesn't mean that you can't save that file at 44.1kHz again … but you won't gain any additional Information.
Another way to think about it: Pitching down means stretching. So if you take a one second audio clip recorded at 44.1kHz, it contains 44100 individual samples. Stretching these 44100 samples over a period of two seconds would mean you are left with 22050 samples per second. Sample rate halved.
So the gist of this is: There is a difference between the file Format and the actual Information that file contains.
yo this is like a bbc documentary but with someone with a personality. great job
hahahahaha he has the perfect voice for BBC
I know! There's nothing like an English sound engineer. They know WTF they're doing.
Refreshing to hear something about sampling I haven't heard a million times before.. 👏
When you sample a minor second on the piano the overtones of both notes clash with eachother more heavily, because the strings vibrate sympathetically to the fundamental frequencies that are played. When you play a minor second with two seperately recorded samples the overtones of the samples still clash, however there are no strings amplifiying the clash of overtones.
My mother was an incredible piano player, she estudied with Nadia Boulanger and used to practice at home at least six or seven hours per day, so before I learned to walk, I used to crawl and seat under her concert piano while rehearsing. So for me, piano sounds come from above in splendid and spreading waves of harmonics. Pure magic. I've been following closely your diving into harmonics content in piano sounds and I totally agree with your efforts to find the Holy Grail of waves behaviour when sampling them. But there's something I still miss from your scientific approach, and it is the inmersive sound quality missing in most piano libraries. I own Spitfire's Hans Zimmer Piano and probably it`s the closest I've heard to these memories seating for more than seventy years in the Hypothalamus of my brain, so I want to encourage you to keep on researching for the ultimate experience of translating the soul of the piano to the binary codes.
It's hard for me to comment on this subject, as I agree with your perspective. My philosophy of sound reproduction is that it just isn't the same thing as being in a room (good or bad room) with a 'real' instrument (including electric/electrified ones). A 'perfectly captured' violin coming out of an eight inch speaker is a representation of a violin, and not even close to being in a room with the same instrument in the same capable hands. The same holds true for an electric guitar and amp. A 'live' instrument resonates in all directions from many sources that sampling, compressing, equalizing and pumping out of a speaker just can't match. Having said that, I still think we need to approximate as accurately as possible, as well as 'modify' the sampled sounds, as thee are new and interesting. I guess what I mean is that we all need to be in rooms with musicians interacting with us (ensemble and audience) more frequently. Music is a human experience to be shared through both media. My fear is that, like the Internet, we become isolated in all our experiences.
I totally agree with@@tdubveedub's comment. Music is a bunch of air molecules moving in all directions and interacting with each others, so there's nothing like the real thing, no matter how much DSP power you can stock to convert phisics into digital emulations. But, there's always a "but", we are facing now a second generation of folks who trained their ears through computer speakers or headphones with low quality and distorted sound curves, so we, composers and sound designers have to relay heavily on the Art of Sampling and all the research and innovation that folks like Spitfire and others can assemble to provide us with dynamic snapshots of sound sources which mimic as best as posible the real thing.
@@teddybautista1805 I agree, and it is too bad that there are so many folks missing out on the ecstasy of chamber and orchestral music. Did you see the reaction of the attendees at a Star Wars convention when John Williams appeared with a complete orchestra and performed a medley? They went wild!
Very true. I've been listening to quite a lot of sampled pianos during the last year. Last week I spent in a house with a big old real piano. And boy oh boy did it 100% exceed those piano samples. The way each key uniquely has its own tone, and the way the whole thing resonates, the wood, the infinite variations in attack / transient response depending on how hard, soft, slow, fast you hit each note: totally totally different experience.
It's somewhat like the difference between seeing a picture of a large monument, vs being there, seeing it IRL.
Bruce Swedien, famed recording engineer (Michael Jackson, Quincy Jones, Count Basie, and many more) has said again and again that EVERYONE wanting to work with sound and music, should pursue seeing and hearing real instruments being playedz Orchestral music. African drumming. Live music with live drums, etc.
I've taken that advice, and don't regret it for a second!
I also listen to lots of vinyl from the pre-digital era. After having listened to lots of streaming media, it's often a shock to me, how huge difference there is.
Eureka!! Oh my God!!
That is the *smartest* and most *efficient* method I have *ever* heard for keeping chopsticks clean at a restaurant!
I remember reading about Trent Reznor recording his guitar pieces and then slowing them down to the desired notes while getting that huge sound, back in the day. Makes sense to me.
Sampling is sharing, and sharing is caring.
I just discovered your channel and i'm already forever thankfull to your illuminating 15 minutes.
" and we like sine waves because sine waves are phat."
-Genius 2019
Agree, that's wrong to say, what he did in the vid is add more bass content but kill midrange texture. More bass = louder = sounds better. More bass partly matches the equal loudness curve (low freqs).
Then in a certain context when arranging/composing u may prefer one sound or another, typical pop song pad sound is soft and gentle without too many odd harmonics, lets u here the reverb more and usually sits better in a mix.
There's Robert Rich and there's Tim Hecker...
made me laugh
Is this some form of mental illness?
Nice one.
For sound design I record 196Khz and then you can pitch down to 5 octaves and still get 44.1Khz sampling rate.
Record 1 minute and when pitched down you get 5 minutes soundscape. Add a massive reverb and you can do a ambient track using sounds around us.
Surely you can only pitch down two octaves and retain a 44K sample rate? Every time you go down an octave, you halve the sample rate.
Which isn't to say that it doesn't still sound awesome...
Interesting concept.. what recordings are best, lively street sounds?
@@charlesgaskell5899 You are correct. In order to pitch a sample down 5 octaves and end up with a 44,100 Hz sample rate, it would be necessary to start with a sample recorded at 1,411,200 Hz.
@@GlenBerry Plus, when you look up the specs of any decent microphone, they don't go anywhere near so you can record such high frequencies. Plus, even if you had a microphone to record those freq God knows what you record, a lot of hum anyway.
@@aileaiho768 I suspect you're confusing sample rate with pitch frequency. Sample rate is a feature of AD conversion, microphones don't have a sample rate unless they have an on-board AD converter (such as USB mics).
"What sampling something cannot do is recreate the infinite chaos of an instrument."
Just found this channel from my recommended. I’m a musician who has been playing for over a decade and I appreciate real informal videos about music &the science behind it. It’s very refreshing to take a break from modern mainstream bull and dive back into real music and what makes it. Especially as a pianist, I love informative vids. I saw all the gear and I knew I would enjoy this vid. I also liked the editing, how you moved from place to place and made clean transitions. You sir have earned a new fan and a subscriber, keep up the good work.
I thoroughly enjoyed this, dope find
This is the most engaging, interesting video I've seen in a long time. Big ups for maintaining my small attention span!
this is the most interesting and educating video I ever seen on youtube 😀
TH-cam recommendations actually spot on for once, this is bloody amazing! Got a few new things to try out now from this! :)
amazing job! so refreshing to hear a concept I had never heard of. So many videos on production are just the same and everything is being restated. Please make more interesting bits like this!
Christian - you are a unique voice in the world of sound and music, with a wonderful holistic grasp of all things from the technical to the emotional. This really resonates (and I use the word deliberately!) with me. Thanks so much for bringing us your insights!
This video had me floored. Instantly subscribed, Christian this is amazing content
The memory limitations of earlier multi-zone samplers, such as the SP-1200, led producers to record the samples pitched up (turntable varispeed, another sampler dedicated to transposition etc) and use the samplers transposition function to pitch the samples down and it did indeed give it a 'phat' sound. I know this was standard practice for the Beastie Boys, De La Soul, Bomb Squad/Public Enemy. I'm assuming Kraftwerk and Art of Noise were confronted with the same limitations. Brian Eno/Daniel Lanois' "Apollo" makes heavy use of the Lexicon Prime Time's sample rate pitch-shift down function and tape varispeed, funnily enough the standout track 'An Ending' is a bunch of CS-80 sine waves recorded hot to tape and you can hear the pleasing tones of "missing information" all over it.
I noticed how more top-40 sample-based stuff start to sound very sterile once memory considerations were no longer a worry, but a lot of more alternative artists still used the degradation techniques.
The most cheeky example I've ever seen of pitching down a sample and making it 'phatter' is 'Feel It All Around' by Washed Out which is just Gary Low's 'I Want You' pitched down a tritone. Had everyone wondering what crazy production techniques he was using until someone found he just typed in -6 on a rare italo disco song.
Whew! I am brand new to your TH-cam channel. What a great introduction for me.
This video makes complete sense to me all the way through. I get it why closely-spaced notes sound differently when samples of them are played from when the original acoustic instrument is being played (e.g. C beating against C#). There's a real difference there that depends on the physics of harmonic vibration in those physical objects (metal strings and wood) versus digital facsimiles. This may be related to the reasons that some people prefer the sounds they can get from analog synthesizers over those produced by digital synths or VST emulations. The physics just isn't quite the same, particularly in polyphonic playing, even within purely electronic instruments.
There are some people, also, who prefer "lo-fi" or "8-bit" audio over high-resolution digital sounds, at least for some musical applications. Those same folks may prefer slowed-down (slightly lower-res) samples over real-time hi-res notes for similar reasons. I have never really understood "fatness" (or "phatness") as applied to synthesizer voices. Now I can see what is meant, I think, a little better.
I would go with selling some carefully curated Pianobook packages commercially through Spitfire while keeping unpackaged samples (or less extensively curated packages) available for free.
Christian, this channel just blew my mind. It's infinitely more intelligent and in-depth than any other music channel I find searching for solutions or inspiration for my music!
Keep Piano Book Free 100%. It's a godsend for composers who can't afford many great sounds. it gives rise to a great community beyond the market mechanisms which allow it to organically grow without an agenda. it also gets people sharing great instruments for the future, capturing the present for the future like a time capsule. yeah, please keep it free. great idea you had too with offering a greatest hits for purchase each year, perhaps also with a bit wider functionality, more velocity layers, etc... thank you for starting this!
I wonder if this what contributes to the preferred character of the sp1200 in hip hop. The limited amount of storage means records are played fast to record more and then stretched or pitched down..
Chris, I love this vídeo so much! I think it is the most important and relevant vídeo made of sampling ever! Thank you so much!
Dude you gotta stop teleporting, that can't be good for you.
hahahahaha
He British, he apperates.
Teleporting makes me nauseous
Haha. Nice one.
I’m not overweight, I’m a sine wave...;0)
We like fat
would make a great tshirt..
It's funnier cause its Chris Brown.
So you've got sinuses.
Stumbled upon this video. Really interesting and very professional! Thank you.
i absolutely loved the video. beatiful fusion of great topic and creative cinematography. pure joy. can't wait for more.
this is absolutely fascinating! I'll try pitching down other samples just to test this out.
I'm glad this seems to be a channel worth a subscribe.
Oh my god. You made an entire (incredibly detailed) video to answer my flippant comment on the last pianobook video. Bless you.
For a bit of context, I have near zero experience sampling pianos (which is why I'm here, to learn), but some experience sampling synthesizers. I assumed, incorrectly, that sampling the cycle of the fifths as opposed to chromatically was to save time, in the notion of an MVP. Thank you for explaining the intent behind it.
My initial reaction to pitching notes down (which spurred my aforementioned comment) comes from what I was taught in sampling synthesizers. Given that synths are more complex than sine waves at various speeds, pitching them down means reducing the resolution of the sample. It often sounds better to pitch up synth samples as this only increases resolution while decreasing loop length.
Having now tried it for myself, I see how much better it is to pitch down piano samples than pitch them up. I'd warrant that this kind of 'cleans' the sample by reducing its resolution and reducing it's complexity. The lower notes when played on a real piano are very complex, while pitched down higher notes are simpler sine waves that have reduced complexity.
As for your question, I wanted to ask what polls you're referencing? I stumbled onto this video series through TH-cam's recommended section, so I don't know if you're pointing to a Facebook, Twitter post or something?
As for the final product and delivery, I wonder how difficult it would be to adapt the code for the Labs plugin? Perhaps in the same way Spitfire delivers Labs releases you could deliver pianos for the pianobook. I don't mean it as a download within the current Labs, I mean it as a copy of Labs. In the way Labs has many instruments, Pianobook as a plugin would have many pianos. The downside of this is the Labs plugin code is likely proprietary and reduces the community's ability to adapt it, etc. Maybe Spitfire should just open source the Labs plugin code any ways? I think you could/should charge a small amount for the best curated pianos- splitting the cost appropriately between the creator/sampler and Spitfire, if Spitfire is lending programing work.
Beautiful editing :) Instant sub for that level of attention to detail
this is a revolutionary concept in my opinion. i love everything about it. i like the rough suggestion you lay out at the end. looking forward to seeing how things progress.
Fantastic insights on sampling. Timely, too. I just finished sampling my electric guitar for a project, and noticed how much better notes played on the high B and E strings sounded when pitched way down to the lower octaves, vs the low low E string. There is something about it being stretched out that allows appreciation for some of that hidden detail, as if you zoomed in on a distant landscape.
Every sound in the video, from the SM7b voice recording, through all the piano and sampled pianos, all of it sounded...really good. Even the wind sounds into the outdoor rig sounded musical. And then I heard the words “Spitfire Audio” and realized why I simply must find more money. I have the free LABS...most of the things in there sound better than anything I have in NI Komplete. I just need the $$$ and some GAS counselling.
Loved that transition at the start. I'm stealing it.
Hey Christian, loved this one and been mulling it over this weekend. I think maybe our preference for pitch shifting downwards rather than up comes down simply to our love of resonant sound the physical vibrational aspect of music. I feel that when we hear the sound of a piano, we hear the physical body of the instrument in resonance. This resonance is a quality of the tone that we are drawn to and as we shift that sound downwards it as though the pianos size and vibrational depth is increased. The tonal weight below and around the vibrating string is heightened. Of course the pleasing aspect of this shifting is limited as we soon notice the sound is being artificially altered which fast becomes a negative trait. I notice the importance of this physical resonance to the sound of the piano when moving to the upper ranges. When we play a high note on the piano, the instrument still has this large resonant body to give a feeling of depth and weight to that note. If we were to achieve this same note by speeding up a sample of a lower note, we would be in effect shrinking the size of this resonant chamber which I feel we inherently interpret as a negative change in the sound. I do think as well this changes from instrument to instrument. Sometimes I prefer the sound a guitar slightly sped up, kick drum or snare (I guess we tighten up attack of the transients which for some sounds would be preferable). Synthesisers as well seem more immune to pitch shifting than other instruments. Perhaps this is because we as listeners are not perceiving the sound as a resonant physical body, and therefore no perceived reality is being contradicted when we artificiality change the sound.
Regarding your sampled piano in particular, I think that when considering why the sampled version sounds more pleasing and more suitable than the real deal, that there are considerations of genre and taste here to think about, rather than a purely objective comparison. I think we can agree that the defining characteristic of the modern zeitgeist piano style is a desired reduction in harmonic content. Whether by playing whisper soft or placing muting felts over the strings, essentially the goal is usually to reduce that hi-frequency resonance that a piano naturally has. When pitching sound downwards, the tone is dulled and the transient attack reduced, helping us greater achieve this lovely soft ghostly sound. Whilst dulling a tone through pitch shifting really lends itself to Satie, I’m not sure it would apply as well to say, Rachmaninov.
Secondly, I feel that what we love to hear in sound and instruments is consistency across their tonal ranges. No lumps and bumps. If we take one note and spread that tonally through pitch shifting, each of these new notes share the same tonal fingerprint and I feel its this consistency that our ears love.
Oh my lord I’m going to watch all of your videos, you are amazing
Very interesting concept! He has such a "BBC accent" and presentation it is great. And I mean that as a complement.
Your delivery in this reminded me very much of James Burke's connections! (not sure if you are old enough for that to have been an influence - very much meant as a compliment) :-)
I think giving everyone the opportunity to use Pianobooks for free would be nice as it's been a community effort. Some people in the comments have talked about open-source software models, which I think is a pretty good idea. If Pianobooks also get a curated, paid version, it would make sense for the funds to be reinvested in the project.
On the topic of file formats, I don't know what's the standard at the moment, but it would be helpful to have formats compatible with Ableton - it recognizes EXS with some annoying steps, and it can also recognize Ableton Packs seamlessly. Sadly they're a bit trickier to put together - they can only be put together by companies, not individuals. Maybe I'm weird, but I really hate having to use plugins, as Live Suite and Max for Live effects and instruments cover pretty much everything I need in a more elegant way.
I've been thinking along the same lines. Currently I'm leaning towards something more 'universal' like .sfz soundfont files. Most daw/samplers will recognize them at least on some level.
Sharing everything - open source and free would be really great for the musical community and i'm totally down with once its set and curated to offer a paid version of it as long as its not expensive, id think 60$ or less would be a fair.
And as to what to do with the money - i think half should be used to keep the project going open source because it still takes money keep it going - its going to need a community website and forum and plenty of bandwith for downloading - its a reinvestment into the project and the other half should be used to get musical instruments into schools in both the UK and the USA that do not have them so that the next generation can learn and keep the project going for decades to come.
This is absolutely fascinating. Thank you so much for sharing your work and ideas! That final piano pad sound that you landed on was superb and I wish I had it!
For an idea as to how to distribute to Pianobook, see the Blender Foundation's model. The Blender Cloud is a paid service that distributes training and various other things, and they frequently crowdfund things and take donations, but the main final product is packaged, and completely free and open source. Anyone can contribute anything, start their own branches, and go into detail, completely changing the product if they need to, but the blender cloud offers training made by qualified professionals.
It may not seem immediately relevant, but the models followed by this and other FOSS projects are definitely worth looking into, and it may be worth asking questions of those involved in them to see what they do/think, to get a good background of undeniably similar things to what you're doing onto which you can build your own model.
I think the Piano Book should be the sort of "open source" form as it has been from the start until today. There are just sooo many commercialised pianos out there that I don't really see the point of adding more to the pile. So keeping it within the community and giving it away freely could be the start of building a bigger community for the next instrument that's going to be sampled. And I really like where you're going with all this. It is a fantastic project from the very start.
i can't believe how much inspired im today to make Music after Watching your videos , I Don't know why . keep Making these gems .
I am so glad I stumbled onto this video. At one point my face was literal inches away from the screen, felt like a revelation of sorts. The appeal of the whole lo-fi community, DJ Screw's influence, hell, even bass-oriented music in general, started making a lot more sense all of a sudden. I always wondered why I enjoyed the sound of pitched-down samples so much more, and always gravitated towards more "sine-like" instrumentation in my productions. Maybe the lack of "information" that's contained within those sounds makes it more pleasant to perceive? Maybe because the most pure of tones are essentially sine waves? Either way, amazing video, it's an instant sub from me.
What your talking about for piano book is how the FOSS (Free Open Source Software) world works. Everything is out there free to download and use, but there are companies that offer packaged version, support, etc for a cost. A lot of companies who use FOSS software as a way to contribute will let their programmers spend some time working a FOSS project as their way to contribute back. So I would say look at the FOSS model for ideas on piano book.
Toortog Pown i wish there was more open source plugins and audio work
Hi Christian, the fact that AS you talked about a charitable endeavour with piano book you got a rainbow says it all😀. A charity based around music would be great, piano lessons for blind kids, that kind of thing would be lovely.
I think you got it right as a community project that's free and then the best of 2019/2020 etc etc on Spitfire with the money going to charity for the polished version.
The production of your videos is frankly some of the best on TH-cam. The algorithm did well today
3:59 anyone know what that controllers left to the computer keyboard are? I faintly remember something like this being announced a few years ago, individual control bits you can stick together to do form a custom controller, but I cannot remember the product or company name.
Nice pad sound with a sustained note you got, I always wanted to get that Weather Report pad sound !
You Sir are a wonderful man. I've felt this for a while.
first of all this is an amazing video dealing with some many interesting aspects of sampling, instrument playing, technology and psychoacoustics. so thanks ever so much!
regarding painobook, something about your openness and creating this community makes me want to keep it a community thing. i.e concentrate efforts to keeping this knowledge, open source in a way, open to endless manipulations and upgrades done by the members of the community.
if going for the "best of boxset" I can see some kind of model of 50% towards keeping the initiative a live and well maintained (investment in a future interface or someone to upload and take care of the site for example) and 50% going to charity (one for example that makes music lessons more accessible for kids). I know I'll be happy to contribute to such a cause.
Randomly stumbled across this and enjoyed it immensely. Instant hit on the subscribe button!
The word Genius comes to mind. You Sir, Are what the sample world needs. I use drum and percussion samples. And tuning them down does make them sound better. But you know why! Wow! Just Wow!
Brilliant explanation and demonstration. Thank you Christian
Could someone please explain what he is doing from 6:25 to 6:50??? I'm using Ableton I don't know how I could achieve this sound, I think it would be with the Sampler device? Thank you
He just sampled a piano note with a sustain pedal, so in ableton you can use the loop option in the simpler to loop the tail of the sample so it never stops, with medium release, then add some plate reverb and cut off a bit the lows u should get something similiar
Hi Christian.
I'm still working on my own piano samples to contribute. I guess my first thought would be just to make it all open, but that's from my perspective as someone who has a little to give, but much more to gain. It's easy to give something that's isn't as good as something else that's better.
Perhaps a minimum entry "fee" of one piano gains one access to others, but perhaps not all?
Or, roll it into Labs and make a pianobook instrument with options for other pianos to be added through that system.
Awesome video Christian can't wait to play the outcome of this project. :)
OMG! LOL! I watched this video twice before the lightbulb went on and I realised that we worked together a while back when you were working with Marcus (can't remember his surname!)! Nice to see you doing well Christian. Great video!
Geoffrey! d’Cruze was the surname! Lovely to hear from you, I hope you’re keeping well.
Wow... a very interesting take on sampling. I like your psycho-acoustic analysis, it's intriguing and I'm wowed.
Wrt your questions about reconciling community contribution with commercialisation: have you explored the use of Creative Commons licenses? A possible application is where contributors submit samples either as a CC-BY-NC (non-commercial) or CC-BY (commercial allowed). Others can utilise the samples as per the license. You're also free to curate the samples that allow commercialisation. Over at a site ccmixter.org, they have adopted a variation called CCPlus, where it's like a royalty payment arrangement if your non-commercial samples are requested for commercial use. Note: you'd still have to deal with a potential issue of a copyright dispute wrt a contributed sample, but if you're going for commercialisation then I suppose treat the rights clearances process as no different from any commissioned project.
That was a really nice take on the topic, thank you! I feel kind of bad about it, but i have to take one thing away from you: we don't hear the low end before the top end. The sensory cells that perceive high frequencies are located at the front end of the cochlea, while lower frequencies are percieved further inside the cochlea. This means, we hear the top end a couple of microseconds before the low end. Not that anyone could actually hear that. The difference is way too small.
On the topic of free vs paid product for Pianobook, this certainly brings to mind open source software. It seems that Pianobook has a few key values:
- Sharing/community: Contributors can share what they have and use what others share.
- Education: Learning sampling, preparation and production techniques.
- Creativity: Sparking ideas with sounds to which you might not have access, otherwise.
These are all values that open source software includes. As long as the community is engaged, this could be self-sustaining. Why not filter decisions through these values? This could be turned into a commercial product later with simplicity being what people are paying for, as you suggest, Christian. I've seen a few (software) projects become sour and go off the rails by considering commercial avenues too soon. It tends to narrow the thinking and prevent the values above from being the main drivers.
As always, great stuff.
Warren Harrison sure thanks Warren the initial motivation was me making a commercial library and showing how I did it, and then I thought I could involve the community so I’m rolling back the original idea as the commercial aspect doesn’t fit for me anymore. I think your advice is sound though.
@@TheCrowHillCo That makes sense. I missed that point about the original idea.
This is just amazing I'm so happy i found your channel!!!
As a huge fan of resampling, this video is one of the best things I've seen.
2:12 I prefer the live piano sound. The sampled version sounds much warmer but as if it's missing the full range of freq. especially in the high end. The resonance on the live piano has been greatly stripped on your sampled sound, initially. Then when you 'phatten' the sound it's starting to get better. That's how my ears/brain interpret. Great video btw.
So running a sample into a sampler and the pitching down and then resampling and pitching back up ?
I'm just amazed you managed to record anything audible on Arthur's seat. Great video!
Yes yes yes! Great video!
Wonderful content. Subscribed and waiting for more!
P.S. Where can I get that beautiful sustain pedal pad? 😍
Love the experiment. Noticed the same when I pitched a piano down a few octaves. I'd like to see the experiment done with higher resolution recordings tho. My theory is that when you record at 96kHz and pitch it down you'll get a brighter sound than with a 44.1kHz recording treated the same way, since you are basically capturing more high end frequencies.
It depends on the ability of the microphone to capture high-range frequencies - the highest pitch that can reliably captured is half the sample rate, and response curves are normally given up to about 22kHz. If the microphone can't respond to frequencies higher than that, then it doesn't matter what sample rate you've used, those frequencies won't be present
@@charlesgaskell5899 that's true aswell. of course you should take that into consideration. Still I find it interesting to see how big of a difference there is comparing the end results side by side
I like the idea of Piano Book being a commercial product. On the one hand I can see how if successful it could open a new direction for Spitfire as a company in regards to its products and services; and on the other seeing and maybe even being involved in the creation of an instrument library is of huge educational value as a composer. How you would marry that all up with rewarding the contributors on a commercial venture I have no idea, but I think its worth exploring further.
Man, this is fantastic stuff! Just found your channel yesterday and really enjoying your content.
Love what you're doing here! I was only dimly aware of it until now. Pianos are amazing instruments. By the way, I'm going to remember that reply about rainbows, ready for when I move north in a few weeks (going to Aberdeenshire).
Fascinating. I would hypothesize that the upper harmonics of low-frequency notes interfere with the fundamentals of higher notes, so cutting off the high end of your low notes makes them blend better.
There was Crowdchoir library. It basically gave free version of Kontakt instruments to all contributors
When the Casio Sk-1 sampler came out back in the 80's, it had 2.5 seconds of sample time. So I would sample records at 45 rpm or 78 rpm and play back the sample on the lower keys to recover the original tempo. The sound was wider and thicker and crunchier since the sample rate was 8 bit. But you are right...anything sampled then played back pitched down is phat to death. Although some like that scratchy alias effect, I try to avoid it at all cost. I dig dirty samples but aliasing turns a good sample into trash.
Hey Christian! I was wondering if you ended up releasing that pad you made in the video? I really liked it!
Even as you were talking, especially about the different contributors, I couldn't help thinking of the way that a lot of open source software is managed. Might be worth looking at some of their models.
Odd question; what keyboard racks are you using?
Seriously thought I was watching a documentary
Great job
I'll tell you what I thought it would be: a site called Piano Book with a Kontakt Library for download and an "About" page. That would already be more than a minimal viable product. You've already gone way the hell beyond that and into space past Voyager 1!
I first have to digest this awesome, eye-opening video of yours, but here are some immediate thoughts:
a) We could change the name of the community project so that if we go forwards with future projects (please let there be future projects) that maybe do not include pianos, there is a sense of continuity. Music Books, maybe?
b) Great explanation for why samples sound bad when you pitch them higher. Another reason for only transposing downwards, is that if you pitch samples upwards, you will hear the shift from one sample to the next much better
c) If the project is to be sold through Spitfire Audio, which I am strangely uncomfortable with (I should be stoked; I'm a fanboy), then it shouldn't cost much at all, maybe the price of a double album, and the money should go towards future projects and compensation for your work and your teaching.
d) there needs to be some defined parameters for wiggle room, so that people (hopefully there are some women invested in this as well?) can work around any problems their instrument may present, or so that they are able to include elements that they think will make it unique.
So, if the particular keys chosen by you, (I forget which ones you chose) contain that one key where there is always a weird buzz that you can't get rid of, you could decide that no, in this particular library it's not Bb and F, but B and F.
And if people want to include a particularly interesting IR file, or a sample that otherwise triggers with every note, there is room for that.
Does that last one even make sense?
that piano pad was gorgeous... is there any way we can get our hands on it?? :-)
Yep we like the Interpolation of the downpitching more than the original mess, cleaner + kind of more powerful / solid in a way. Relating that to our origin in the uterus is interesting, good reflections. If you want to see how powerful can be that effect get, instead of kontakt, use a emu ultra, a ASR or one of the few first akais.... now THAT is amazing.
Really cool. I have no idea about the project to which you are referring but the pitch demo and lecture were insightful. Sub'd
the ultimat answer is what feeling you want to achieve in the soul of the listener ... raw sounds of pianos round the world could be free ... what takes time and skill is to process the data the result is always a synthetic sound (you could buy); a piano sounds is different in different rooms even in a analog situation (chamber concerto) the sound will be different depending on the place you sitt. ... its like so many things all is about the context
Bloody great video, Christian. I'm inspired to go and sample my tiny toy piano today. Now I just have to figure out how...
an absolute diamond. Psychoacoustic fascinating research, interesting presentation of the material. Thank you.
is that pad you made with the sustain pedal going to be made available?
Best sound channel on TH-cam, along with @Benongear.
Loved the video Christian! Really good production value as well. 🍻
What are you using for your SM7B? (preamp, interface, cloudlifter, is it set to flat?, etc.). Sounds really clean! Thanks
Interesting discussion! I will always try pitching a sample up and down, usually do pitch down if anything. Sounds like a great project, what about other instruments? or non-instruments...
I am interested in contributing a Wurlitzer grand that I love. But I don't use EXS or Kontakt. I use Grace Sampler, ReaSamplomatic, and occasionally Ableton Sampler. Raw WAV samples is what I prefer, so I can compile them myself. Do you think there would be any interest in compiling for Grace Sampler, or simply providing the Raw WAV samples?
I'm so happy I've found this channel
Pretty interesting! Love the way you are fascinated by the subject!
This was a fascinating journey. Thank you for it.
What is the small grey console on your left??