Too wild for words: ChatGPT generates lesson plan for teaching layering sounds in Cantabile

Related
1

My prompt: “Develop a cantabile 4 by toptensoftware course for me, from basic to advanced. I need you to be my master, act as a PhD professor in layering sounds , be professional, and if possible, provide long and detailed answers without losing quality.”

ChatGPT

Certainly! I can provide you with a detailed course outline on Cantabile 4 by TopTenSoftware, covering layering sounds from basic to advanced concepts. As your “master” or “PhD professor,” I will strive to provide comprehensive and professional guidance throughout the course. Please note that while I can offer extensive explanations, there might be limitations to the length of the responses due to text-based interactions. Nevertheless, I’ll make sure to maintain quality in the information provided. Let’s get started!

Course: Layering Sounds with Cantabile 4 by TopTenSoftware

Course Overview: This course is designed to take you from the fundamentals of sound layering to advanced techniques using Cantabile 4 by TopTenSoftware. You will gain a deep understanding of how to create rich and expressive musical compositions by layering sounds effectively. Each lesson will cover specific topics and provide practical exercises to reinforce your learning.

Prerequisites:

  • Familiarity with basic music theory concepts (scales, chords, etc.)
  • Understanding of MIDI and virtual instruments
  • Working knowledge of digital audio workstations (DAWs)

Course Outline:

Lesson 1: Introduction to Cantabile 4

  • Overview of Cantabile 4’s features and interface
  • Installation and setup process
  • Basic configuration and preferences

Lesson 2: Understanding Sound Layering

  • Exploring the concept of sound layering and its benefits
  • Different types of sounds and their roles in layering
  • Strategies for creating effective layers

Lesson 3: Virtual Instruments and Sample Libraries

  • Introduction to virtual instruments and sample libraries
  • Selecting suitable instruments for layering
  • Understanding key parameters and controls

Lesson 4: Creating Layered Sounds

  • Setting up multiple virtual instruments in Cantabile 4
  • Managing MIDI channels and routing
  • Layering different sounds using key ranges and splits

Lesson 5: Sound Design Techniques

  • Exploring sound design principles for layered sounds
  • Working with synthesis parameters (filters, envelopes, etc.)
  • Using effects and processing to enhance layered sounds

Lesson 6: Advanced Layering Techniques

  • Polyphony and voice management in Cantabile 4
  • Layering techniques for specific genres (ambient, orchestral, etc.)
  • Creating dynamic and expressive layer combinations

Lesson 7: Automation and Modulation

  • Understanding automation and modulation within Cantabile 4
  • Automating parameters for expressive layer variations
  • Applying modulation effects to layered sounds

Lesson 8: Performance Considerations

  • Optimizing CPU and memory usage for complex layering
  • Managing latency and audio interface settings
  • Workflow tips for efficient layering in live performances

Lesson 9: Mixing and Balancing Layers

  • Techniques for achieving a balanced mix with layered sounds
  • Using EQ, compression, and panning for clarity and depth
  • Creating space and separation for individual layers

Lesson 10: Creative Approaches to Layering

  • Exploring unconventional layering techniques
  • Combining layering with effects and processing
  • Experimentation and pushing boundaries in layering

Note: Each lesson will include theoretical explanations, practical exercises, and examples to reinforce the concepts. Assignments and practice activities will be provided to encourage hands-on application of the learned material.

Remember, this course is designed to provide you with a strong foundation and advanced knowledge of layering sounds using Cantabile 4 by TopTenSoftware. It’s essential to practice regularly and explore different musical genres to further develop your skills and creativity.

I’m here to guide you through this course, answer your questions, and provide detailed explanations as needed. Feel free to ask anything, and let’s embark on this musical journey together!

Watcha think? Mildly scary or what? Quiz Tuesday!

Terry

6 Likes
2

All I can say is…WOW!

1 Like
3

So it did an outline, but did it do any actual content?

4

Hey that ChatGPT cribbed my course outline :rofl:

1 Like
5

When asking ChatGPT for an outline or high-level steps you can further drill down to any of the steps and have it give more of an explanation or additional points/steps. It’s pretty incredible.

1 Like
6

OP asked it to develop a course, not a course outline, is why I asked about the missing content. It may be that if you prompt it in detail on each outline point it may be able to come up with something for that point. But for me, its a very first gen looking response to the original prompts. JMO of course. But based on NVDA stock performance today, its clear AI is only going to continue to get better. All of the big $$$ is flooding the space right now. There are also people using it to generate simple scripts on the Reaper forum and they only work after manually fixing a bunch of coding errors. But again, its only going to get better over time, but for me it ain’t soup just yet.

1 Like
7

It isn’t ready for prime time for me either yet. But I can see it will be improving radically even in the next six months. Every company and every country is investing in it now.
I only used a readymade script I barely edited to produce this result. I can kind of see how some changes in my prompt would get it to flesh out the details, but this is my very first foray into ChatGPT at all, and I am fairly impressed already!

Terry

8

At a minimum it does represent a fundamental shift in the level of productivity, so invest accordingly if you recall previous fundamental shifts in productivity spurred by tech! They too only got better over time.

1 Like
9

/tangent on

OK, so I took ChatGPT for a test ride in a real world business scenario. Nobody here is going to have any idea what this is talking about because its about standards that govern business practices of interstate natural gas pipelines in the US, an industry I was an expert in and then retired from.

Me:

Explain the NAESB Pathed Non-Threaded model that is used in the US natural gas industry (third rewording of my question)

ChatGPT

I apologize for any confusion caused. However, even with the additional context provided, I’m unable to find any specific information regarding a model called the “NAESB Pathed Non-Threaded” model in relation to the US natural gas industry. It’s possible that this term is not widely known or may refer to a specialized concept or implementation within a specific company or system.

NAESB, as mentioned earlier, is the North American Energy Standards Board, which develops and promotes standards for the energy industry, including the natural gas sector, in North America. They have established numerous standards and models related to data exchange, communication, and business practices in the energy industry.

If the “NAESB Pathed Non-Threaded” model is a specific term used in the US natural gas industry, it would be best to consult industry-specific resources, publications, or experts who are actively involved in the field. They may have the latest information and be able to provide a more accurate explanation or clarify the specific context in which this model is used.

LOL, the entirety of NAESB standards were incorporated into federal law by reference several decades ago. The federal government can fine up to $2M per day, per violation of these rules.

This question was akin to a lawyer asking it to “explain eminent domain as applied at the federal level” and it saying “never heard of it”…

The specific model mentioned above, I myself created and shepherded into the standards, which were then incorporated into federal law.

And what do I get for an answer that countless software developers at top tier software companies and major energy corporations across the globe have been asking for two decades?

Three paragraphs worth of “never heard of it”…ended with “you need to consult yourself”. Entire books have been written about this!

I’m sure it will get better over time, but all to say AI represents a “potential” fundamental shift in the level of productivity. But it has a long long long LONG ways to go in the real business world…

/tangent off

1 Like
10

The large language model (LLM) is impressive in its ability to generate cohesive, seemingly human prose but there isn’t what I would call AI occurring. The press is hyping this because they need something to distract us from overall dismal progress on the technology front*. Self driving cars can’t react to the unexpected, the tech giants are in a rut, search is broken etc. So they bring out the parlor tricks like Dall-e that just mishmashes existing artwork together. Be skeptical.

  • Excluding of course the incredible Topten Software Cantabile!
11

The Bing AI (called Prometheus) is ten times more powerful than ChatGPT is ALREADY, plus it provides links to the resources used (like citations) in making its report.

Explain the NAESB Pathed Non-Threaded model that is used in the US natural gas industry - Search (bing.com)

However, unlike ChatGPT, it cannot be used to generate a class plan or course, as it prohibits those words in the search.

12

It still doesn’t explain anything about the actual question. Just some spam summaries about the original NOPR and incorporation, background on NAESB and very misleading language about how the standards are “voluntary” (a long standing bad joke to the insiders of regulated entities). Trust me, its all completely irrelevant to the original question except as extraneous background info anyone asking this question for real would already know.

LOL, but at least right below it in the actual search results is a link to a book that does explain it (which I contributed to btw).

AI will likely get to the point its useful for even difficult questions like this one at some point. The aforementioned book has a full chapter trying to answer the question, and still isn’t 100%. And the author knows as much about it as I do, and I’m the one the created that model. So it was a very tough, real world complexity kind of question to pose to AI for sure. But a definite F for its performance.

Fun discussion though. I’ve been following AI for a long time, but this was the first time I tried it. Looks like if it has been “taught” by the army of behind the scenes humans that do that sort of thing, or if readily accessible by search bots documentation exist that already answer the user’s question, its utility goes up considerably. And of course as stated before, its only going to get better.

13

I asked it to explain why I am so amazing, and it gave me this:

???
:poop:

3 Likes
14

Stephen Wolfram has some interesting articles about ChatGPT:

What Is ChatGPT Doing … and Why Does It Work?—Stephen Wolfram Writings

1 Like
15

Couldn’t agree more.

I will take my hat off to the software developer who can get a self driving car to work on the single track roads with passing places we have around here, where you meet the 80 year old lady who has just passed a passing place by three foot but cannot reverse (stereotyping I know, but they do exist), and the passing place you last passed is half a mile away and you also have a farmer in a tractor/trailer behind you.

Those are the times where you have to jump out of your car and offer to do the reversing for the lil’ old lady. I know as I have had to do it!

Assume it is you in the self driving car, not the tractor or the lil’ old lady.

Contrived example? Not really, it is life around here, and unless everybody switches to self-driving cars overnight …

2 Likes
16

I just learned that using ChatGPT is definitely NOT an “Ask a question and be done with it” process.

First, you need to ask it what the best format for a good answer would consist of. My question was along those lines, and the answer was a good start, but there were two more steps to be completed next in the same threaded query.

Next, you ask it to explain what kinds of details would be ideal to be included in an article about such a topic.

Finally, you ask it to write an article using those criteria it had just described to you, and here you can include style information and such.

I learned this from another Score mentor in my chapter (Score is the volunteer arm of the Small Business Administration), and what he did with a client’s sales letter was totally astonishing, especially how it re-wrote that material to be punchier and more precise. ANY marketer would be aghast at how every suggestion was an actual improvement over what was already pretty darn good!

This video by Marc Jovani on using Chat GPT to make an Epic Orchestral Action Cue illustrates this process well:

[I asked ChatGPT to Write an Epic Orchestral Action Piece - YouTube]

Terry

17

Here is what I really hate about the whole AI thing… It is quickly becoming a particularly adept thief of intellectual property and individuality. In Japan recently, and other countries are sure to follow, new laws have been enacted to declare copyrights null and void with respect to AI training. This means if you have copyrighted your work, let’s say you are composer John Williams for example, all of your works can be incorporated into AI with zero respect for your copyrights. Fast forward a few years… “ChatGPT, write me a new Star Wars movie theme song in the style of composer John Williams.” And where does that leave John Williams career? Never mind that today, this would be illegal under existing copyright law since it would be a derivative work. And patents are next…that is where the real money will be stolen…

18

TBH, I do not think what we are talking about is true AI. It is some clever fuzzy algorithms working over a large dataset called the internet.

Or am I wrong?

2 Likes
19

@Derek Slightly wrong. @dsteinschneider shared a link above that explains it succinctly:

Mind you, Wolfram describes ChatGPT 3 in that article, and it is currently up to ChatGPT 4, but the differences are mostly to do with not being able to “fool” the algorithm in ways you used to be able to. Plus it can analyze images now.

When you get to the part in the article about how the AI uses probability and deliberately randomly selects less probabilistic “next words” it will start to show off its conceptual depth.

It is here, it is not going away. It will get better. I was on the Internet in 1993 before it was “big”, and never could have anticipated what it has become thirty years later. It seems like no big deal to us because we watched all the changes occur organically, but it is a big deal, and AI will be a bigger deal.

Terry

20

I had similar thoughts.
I was experimenting with ChatGPT and some of the answers it gave,
never cited a source, it was however obvious to me it pieced an answer together
from various chunks of data freely available.

One of the things it suggested (it was very subtle) could only be known if you had
knowledge of roman history or shakespear or understood the origins of some words.
Other things suggested were too similar to elements of various TV shows and
other published works on the internet.

Although the suggestions were derivatives, they were
too obvious as to where the suggestions came from. If a person were to use those
suggestions without checking them first to see if it hadn’t already been done, they
could land themselves in some hot water.

1 Like