ICT3 "E260" wavetable synth sound chip

[Ecernosoft]

Ok, so you guys probably know how keen I am on trying new things.

I was going to post about a possible DVD drive for the 7800, but instead I thought of making a 16 voice wavetable synth chip, where each sound wave is 64 8 bit samples, and you get 256 of them (Stored on a 16KB RAM chip)

 

Is this viable? Because if it is, I have a working design for it.

 

Thanks!

[Jinks]

What did you do to ICT 2? You said ICT 1 will be scrapped. But I thought 2 would be where things get real.  Now the talk of ICT 3 and ICT 4 and chips?

 

What if you made a finished game and just called it ICT?  The other "games" could be just w.i.p. versions to the real game you make? 

 

 

[+x=usr(1536)]

1 hour ago, Ecernosoft said:

I was going to post about a possible DVD drive for the 7800

What are your thoughts in terms of implementing this?  I've been having thoughts of things involving video input on the 7800 as well, so would like to hear more.

1 hour ago, Ecernosoft said:

but instead I thought of making a 16 voice wavetable synth chip, where each sound wave is 64 8 bit samples, and you get 256 of them (Stored on a 16KB RAM chip)

I'm only sort-of following this: the number of voices is understood, but if I'm doing the maths correctly, a sample that's a total of 64 bytes long won't be able to hold much in the way of usable audio data other than short, very low-quality samples of waveforms.  I get that there's 256 of them, adding up to 16K of total data, but are they combined additively, subtractively, or in some other way to create sounds?  The underlying method of how the sound output is actually constructed is where I'm losing the plot on this one.

1 hour ago, Ecernosoft said:

Is this viable? Because if it is, I have a working design for it.

From what's been described so far, there's really no way to evaluate whether it's workable or not.  Post the design; I'm genuinely curious to see what your implementation looks like.

[Astal4]

I am working on a wavetable synthesis chip already for the 7800, also working on a large sample storage for it via NAND chips, however I'm hitting a budget wall. I really really really cannot recommend you go and read documentation on the 7800 enough. Practice your skills, take your time, you're jumping FAR FAR ahead of your skill. A DVD drive on the 7800 would not be a practical idea, I fucked around on the Discord a bit with the idea of a 7800CD addon via adjusting the XTIDE board to 7800 and making an ATAPI driver, but it's just not practical, especially since you can get brand new NAND 512MB flash chips for under $4 in quantities of 100.

 

I could talk about hardware and what-ifs all day, but I would recommend you go and read documentation, practice your skills, etc. People are going to start to get very mean if you continue to ignore them.

[Ecernosoft]

6 hours ago, x=usr(1536) said:

What are your thoughts in terms of implementing this?  I've been having thoughts of things involving video input on the 7800 as well, so would like to hear more.

I'm only sort-of following this: the number of voices is understood, but if I'm doing the maths correctly, a sample that's a total of 64 bytes long won't be able to hold much in the way of usable audio data other than short, very low-quality samples of waveforms.  I get that there's 256 of them, adding up to 16K of total data, but are they combined additively, subtractively, or in some other way to create sounds?  The underlying method of how the sound output is actually constructed is where I'm losing the plot on this one.

From what's been described so far, there's really no way to evaluate whether it's workable or not.  Post the design; I'm genuinely curious to see what your implementation looks like.

1. It wouldn't be video, since the 7800 has no video input to my belief. It would be full DVD audio, DMA control, 96KB of RAM (3 seperate 32KB chunks that can be read/written to by the maria, CPU, or DMA chip but not all at the same time) (48KB for CPU, 48KB for video)

2. No. My chip plays samples.... period. No special effects.

3. Yeah, true, and @Astal4 feircley reminded me he's making his own.

 

4 hours ago, Astal4 said:

I am working on a wavetable synthesis chip already for the 7800, also working on a large sample storage for it via NAND chips, however I'm hitting a budget wall. I really really really cannot recommend you go and read documentation on the 7800 enough. Practice your skills, take your time, you're jumping FAR FAR ahead of your skill. A DVD drive on the 7800 would not be a practical idea, I fucked around on the Discord a bit with the idea of a 7800CD addon via adjusting the XTIDE board to 7800 and making an ATAPI driver, but it's just not practical, especially since you can get brand new NAND 512MB flash chips for under $4 in quantities of 100.

 

I could talk about hardware and what-ifs all day, but I would recommend you go and read documentation, practice your skills, etc. People are going to start to get very mean if you continue to ignore them.

Huh? Am I ignoring you guys???

Heya, I absolutley don't mean that! 

And also, nice! I'm excited for your chip.

[Silver Back]

I think he means your ignoring the advice being given freely. Your dealing with people who are life long software and hardware developers, some who was even around when these consoles were new. They are willing to help you but you jump around a lot and it gets exhausting.  Especially when these people have been working on this stuff as a passion for years and here comes somebody who moves onto the next thing almost daily. 

 

They've mostly been there and done that, so really listen to what their telling you as far as the trouble their having doing some of these things.  Focus on developing a really tightly coded game that could be finished and I bet the community would not only be impressed but rally more behind any other idea that you come up with. Show the ability to see something to the end. 

 

 

[+x=usr(1536)]

1 hour ago, Ecernosoft said:

1. It wouldn't be video, since the 7800 has no video input to my belief.

True, but I have some ideas as to how it may be possible to work around that.  One of them hinges on MARIA's ability to switch off some of its own pins, however.  If that isn't possible, it'll have to be external.

1 hour ago, Ecernosoft said:

It would be full DVD audio, DMA control, 96KB of RAM (3 seperate 32KB chunks that can be read/written to by the maria, CPU, or DMA chip but not all at the same time) (48KB for CPU, 48KB for video)

OK, that clears it up a bit.  When you mention DVD audio, you are referring to the 24-bit / 96KHz audio-specific format, correct?  Is this for output, input, or both?

1 hour ago, Ecernosoft said:

2. No. My chip plays samples.... period. No special effects.

So what you're proposing is basically a sampler.  Nothing wrong with that.  Two recommendations:

1. Read up on additive and subtractive synthesis.  This can be done with samples as well as FM, etc. sounds.
2. Look into music trackers and how they work.  What you're proposing sounds very similar to the ways in which they play back and mix samples.  This is fine, but there's likely some legwork you can save yourself by seeing how they handled it.

The one thing that I keep coming back to in my mind is the sample length that you've chosen.  64 bytes doesn't allow for a lot of resolution simply because more resolution requires more storage space - and you'll run into this pretty quickly.  There will also need to be a way to pitch-shift the samples in playback, which is where I see a) hitting the wall with sample size / resolution and b) the need for a tracker-alike engine to handle that plus the mixing for final output.

 

None of these are unobtainable goals, and you may have already figured some of them out.  What core(s) are you thinking of basing this around?

[Ecernosoft]

2 hours ago, x=usr(1536) said:

True, but I have some ideas as to how it may be possible to work around that.  One of them hinges on MARIA's ability to switch off some of its own pins, however.  If that isn't possible, it'll have to be external.

OK, that clears it up a bit.  When you mention DVD audio, you are referring to the 24-bit / 96KHz audio-specific format, correct?  Is this for output, input, or both?

So what you're proposing is basically a sampler.  Nothing wrong with that.  Two recommendations:

1. Read up on additive and subtractive synthesis.  This can be done with samples as well as FM, etc. sounds.
2. Look into music trackers and how they work.  What you're proposing sounds very similar to the ways in which they play back and mix samples.  This is fine, but there's likely some legwork you can save yourself by seeing how they handled it.

The one thing that I keep coming back to in my mind is the sample length that you've chosen.  64 bytes doesn't allow for a lot of resolution simply because more resolution requires more storage space - and you'll run into this pretty quickly.  There will also need to be a way to pitch-shift the samples in playback, which is where I see a) hitting the wall with sample size / resolution and b) the need for a tracker-alike engine to handle that plus the mixing for final output.

 

None of these are unobtainable goals, and you may have already figured some of them out.  What core(s) are you thinking of basing this around?

I…. Don’t know. (I honestly don’t know what that is )

[Ecernosoft]

2 hours ago, x=usr(1536) said:

True, but I have some ideas as to how it may be possible to work around that.  One of them hinges on MARIA's ability to switch off some of its own pins, however.  If that isn't possible, it'll have to be external.

OK, that clears it up a bit.  When you mention DVD audio, you are referring to the 24-bit / 96KHz audio-specific format, correct?  Is this for output, input, or both?

So what you're proposing is basically a sampler.  Nothing wrong with that.  Two recommendations:

1. Read up on additive and subtractive synthesis.  This can be done with samples as well as FM, etc. sounds.
2. Look into music trackers and how they work.  What you're proposing sounds very similar to the ways in which they play back and mix samples.  This is fine, but there's likely some legwork you can save yourself by seeing how they handled it.

The one thing that I keep coming back to in my mind is the sample length that you've chosen.  64 bytes doesn't allow for a lot of resolution simply because more resolution requires more storage space - and you'll run into this pretty quickly.  There will also need to be a way to pitch-shift the samples in playback, which is where I see a) hitting the wall with sample size / resolution and b) the need for a tracker-alike engine to handle that plus the mixing for final output.

 

None of these are unobtainable goals, and you may have already figured some of them out.  What core(s) are you thinking of basing this around?

Thanks for the advice.

 

I think if this were to work, it would  be a great alternative to the 2151 for those who want better than pokey.

 

Edit: I didn’t know a DVD can do 24 bit 96 kHz audio! Wow!

I was expecting 16 bit at 48 kHz.

Edited November 4, 2022 by Ecernosoft

[Ecernosoft]

10 hours ago, x=usr(1536) said:

What are your thoughts in terms of implementing this?  I've been having thoughts of things involving video input on the 7800 as well, so would like to hear more.

I'm only sort-of following this: the number of voices is understood, but if I'm doing the maths correctly, a sample that's a total of 64 bytes long won't be able to hold much in the way of usable audio data other than short, very low-quality samples of waveforms.  I get that there's 256 of them, adding up to 16K of total data, but are they combined additively, subtractively, or in some other way to create sounds?  The underlying method of how the sound output is actually constructed is where I'm losing the plot on this one.

From what's been described so far, there's really no way to evaluate whether it's workable or not.  Post the design; I'm genuinely curious to see what your implementation looks like.

64 bytes looping over and over, like on the PCE with it’s sound chip. Only that was 5 bits per sample with 32 samples.

[+x=usr(1536)]

On 11/4/2022 at 9:47 AM, Ecernosoft said:

I…. Don’t know. (I honestly don’t know what that is )

I'm not sure what it is that you're referring to.  Enlighten me

On 11/4/2022 at 9:50 AM, Ecernosoft said:

I think if this were to work, it would  be a great alternative to the 2151 for those who want better than pokey.

Possibly, but almost any sound chip can be interfaced to the 7800; it's just a question of what sits between it and the software to make it happen.  Also, check out BupChip (as used in Rikki & Vikki) if you haven't already.

On 11/4/2022 at 9:50 AM, Ecernosoft said:

Edit: I didn’t know a DVD can do 24 bit 96 kHz audio! Wow!

I was expecting 16 bit at 48 kHz.

In all honesty, I was a little surprised to see you mention it as DVD-Audio never caught on as a music format and was pretty much dead by the mid-2000s.  Regular DVDs are 16 bit at 48KHz, though.

On 11/4/2022 at 9:52 AM, Ecernosoft said:

64 bytes looping over and over, like on the PCE with it’s sound chip. Only that was 5 bits per sample with 32 samples.

That's going to sound...  Harsh.  Note that I'm saying this as someone who has been listening to Industrial for over 30 years and has worked pretty extensively with 8-bit samples in trackers, so harsh music has a certain appeal to me.  But it needs to be music, not just a conglomeration of noise.  With the types of samples you're considering using, I suspect that the end result may not match expectations.

 

Even if dealing with samples that are basically just different waveforms pitch-shifted appropriately, you still need to account for things like panning, volume levels, time expansion or compression (which is pitch-related).  I'm not saying that the idea is bad, just that there are a lot of variables to take into account.

 

Try this: grab a tracker (on the PC, I was always a fan of Scream Tracker and Impulse Tracker; YMMV), generate some samples of the type you intend to use in Audacity (pure square, sine, triangle, sawtooth, etc., waveforms recommended), and do some experimentation with them in whatever tracker you select.  This should give you some idea of why I'm mentioning a lot of these concepts, because they'll apply to the proposal you're making for your sound device.

[Ecernosoft]

26 minutes ago, x=usr(1536) said:

I'm not sure what it is that you're referring to.  Enlighten me

Possibly, but almost any sound chip can be interfaced to the 7800; it's just a question of what sits between it and the software to make it happen.  Also, check out BupChip (as used in Rikki & Vikki) if you haven't already.

In all honesty, I was a little surprised to see you mention it as DVD-Audio never caught on as a music format and was pretty much dead by the mid-2000s.  Regular DVDs are 16 bit at 48KHz, though.

That's going to sound...  Harsh.  Note that I'm saying this as someone who has been listening to Industrial for over 30 years and has worked pretty extensively with 8-bit samples in trackers, so harsh music has a certain appeal to me.  But it needs to be music, not just a conglomeration of noise.  With the types of samples you're considering using, I suspect that the end result may not match expectations.

 

Even if dealing with samples that are basically just different waveforms pitch-shifted appropriately, you still need to account for things like panning, volume levels, time expansion or compression (which is pitch-related)  I'm not saying that the idea is bad, just that there are a lot of variables to take into account.

 

Try this: grab a tracker (on the PC, I was always a fan of Scream Tracker and Impulse Tracker; YMMV), generate some samples of the type you intend to use in Audacity (pure square, triangle, sawtooth, etc., waveforms recommended), and do some experimentation with them in whatever tracker you select.  This should give you some idea of why I'm mentioning a lot of these concepts, because they'll apply to the proposal you're making for your sound device.

Yeah, I don’t think this is going to actually happen.

And with 16 voices, it would need to run at at least 1 mhz.( to get up to 64 kHz)

so, what’s 1.79 mhz / 16?

[Astal4]

Focus on ONE project with no special features first. Learn the hardware and do it start to finish. In later stages you can go back and add features. Really focus on something and see it through.

[Ecernosoft]

3 hours ago, x=usr(1536) said:

I'm not sure what it is that you're referring to.  Enlighten me

Possibly, but almost any sound chip can be interfaced to the 7800; it's just a question of what sits between it and the software to make it happen.  Also, check out BupChip (as used in Rikki & Vikki) if you haven't already.

In all honesty, I was a little surprised to see you mention it as DVD-Audio never caught on as a music format and was pretty much dead by the mid-2000s.  Regular DVDs are 16 bit at 48KHz, though.

That's going to sound...  Harsh.  Note that I'm saying this as someone who has been listening to Industrial for over 30 years and has worked pretty extensively with 8-bit samples in trackers, so harsh music has a certain appeal to me.  But it needs to be music, not just a conglomeration of noise.  With the types of samples you're considering using, I suspect that the end result may not match expectations.

 

Even if dealing with samples that are basically just different waveforms pitch-shifted appropriately, you still need to account for things like panning, volume levels, time expansion or compression (which is pitch-related)  I'm not saying that the idea is bad, just that there are a lot of variables to take into account.

 

Try this: grab a tracker (on the PC, I was always a fan of Scream Tracker and Impulse Tracker; YMMV), generate some samples of the type you intend to use in Audacity (pure square, sine, triangle, sawtooth, etc., waveforms recommended), and do some experimentation with them in whatever tracker you select.  This should give you some idea of why I'm mentioning a lot of these concepts, because they'll apply to the proposal you're making for your sound device.

Well, a DVD can hold 4.7 gigs

[Silver Back]

I think laser disc would be more appropriate for the 7800 as that was something that was teased early in its life. Why add a DVD player?  I mean you could do hard drives and new boards with the most powerful gpu and modern tech available but at some point it loses the charm of being a 7800. 

Pushing the limits of the existing hardware and trying to implement the lost add-ons that Atari originally planned is more appealing to me.  

If I want to play a console that plays DVDs I have the Xbox and PlayStation for that. Maybe you lost me somewhere and I'm not understanding what your saying.

[Ecernosoft]

5 minutes ago, Silver Back said:

I think laser disc would be more appropriate for the 7800 as that was something that was teased early in its life. Why add a DVD player?  I mean you could do hard drives and new boards with the most powerful gpu and modern tech available but at some point it loses the charm of being a 7800. 

Pushing the limits of the existing hardware and trying to implement the lost add-ons that Atari originally planned is more appealing to me.  

If I want to play a console that plays DVDs I have the Xbox and PlayStation for that. Maybe you lost me somewhere and I'm not understanding what your saying.

You're right. It was just an idea.

 

Lazerdisk would be so cool!

 

 

(Thinks of replacing ICT4 with ICT LazerDisk edition)

Edited November 5, 2022 by Ecernosoft

[+x=usr(1536)]

5 hours ago, Ecernosoft said:

Well, a DVD can hold 4.7 gigs

Or 8.5GB if dual-layer.  Not clear as to what this has to do with what we've been discussing so far, though.

30 minutes ago, Silver Back said:

I think laser disc would be more appropriate for the 7800 as that was something that was teased early in its life.

As much as I agree with you on this, it's really impractical to do in 2022 for a variety of reasons.  Without going too deeply into specifics as to why, the audience where 7800 ownership, working laserdisc player ownership, and being interested enough to buy a device that would let them merge the two probably numbers less than 10.  Using something like a DEXTER would be much easier and far preferable, but if you're going that far it makes more sense to come up with a general-purpose video streaming input solution that doesn't cost $500 and up.

 

[Ecernosoft]

3 minutes ago, x=usr(1536) said:

As much as I agree with you on this, it's really impractical to do in 2022 for a variety of reasons.  I won't get into specifics here since that's a huge rabbit-hole to go down, but using something like a DEXTER would be much easier and far preferable.

That's why I brought up the DVD player! To remedy the Laserdisk!

[+x=usr(1536)]

On 11/4/2022 at 8:11 PM, Ecernosoft said:

That's why I brought up the DVD player! To remedy the Laserdisk!

Laserdisc was mentioned after you had suggested DVD.

 

This line of reasoning is becoming way too fragmented to follow.  BTW, you do understand that laserdisc is an analogue format, yes?

[Ecernosoft]

16 minutes ago, x=usr(1536) said:

Laserdisc was mentioned after you had suggested DVD.

 

This line of reasoning is becoming way too fragmented to follow.  BTW, you do understand that laserdisc is an analogue format, yes?

Oh....  Well, no ICT LD then.

 

The reason I thought of a DVD player was so that games like ICT can use more memory and still be cheap so that Al doesn't have to spend a fortune on giant ROM chips, as well as amazing sound and visuals.

Edited November 5, 2022 by Ecernosoft

[+x=usr(1536)]

20 minutes ago, Ecernosoft said:

The reason I thought of a DVD player was so that games like ICT can use more memory and still be cheap so that Al doesn't have to spend a fortune on giant ROM chips, as well as amazing sound and visuals.

All of this can be accomplished in solid state these days with better reliability, flexibility, and for much less cost (and aggravation) than it would take to integrate a physical DVD drive.

 

Besides, DVD's a dead media format.  Sure, they're still being mass-produced and probably will be for another decade or so, but newer technology has already overtaken it.

[Ecernosoft]

Just now, x=usr(1536) said:

All of this can be accomplished in solid state these days with better reliability, flexibility, and for much less cost (and aggravation) than it would take to integrate a physical DVD drive.

 

Besides, DVD's a dead media format.  Sure, they're still being mass-produced and probably will be for another decade or so, but newer technology has overtaken it.

A hard drive. For the 7800. Wow.

[+x=usr(1536)]

2 minutes ago, Ecernosoft said:

A hard drive. For the 7800. Wow.

No, solid state storage.  Note that I named a technology, not a specific type of product.  An SD card is solid state, just not using the same storage methodology as an SSD, CF card, or EEPROM.  Any of those (and many others) would qualify.

[Ecernosoft]

1 hour ago, x=usr(1536) said:

No, solid state storage.  Note that I named a technology, not a specific type of product.  An SD card is solid state, just not using the same storage methodology as an SSD, CF card, or EEPROM.  Any of those (and many others) would qualify.

Oh, you meant an SD card... oh!

[+x=usr(1536)]

10 hours ago, Ecernosoft said:

Oh, you meant an SD card... oh!

No.  Again: I was referring to any device capable of solid-state storage. That leaves the choice of media completely open-ended.