Digitally Remastered 8 bit released

[Pat the Cat]

https://github.com/Pat-McD/Amiga-DR-8Bit

The FLAC is an example of "perfect" implementation of the concept. In practice it is unlikely the hardware would sound that good, and the other examples are fairer demonstration of what to expect. The MP3s are "warmer" but less defined, the "Oggs" are cleaner but more clicky (no digital delay added to Oggs). If you don't like the Oggs, save them as MP3 and they should sound less clicky.

They sound pretty damn hot, but any credit for that goes to the original musicians. They will be a mixture of MP3, Ogg and FLAC, and won't sound good on a real Amiga unless you are running on a Vampire. But they should give an idea of what a proposed sound processor will produce, when connected to the audio output jacks of a real Amiga.

The whole thing was done as a technology demonstrator for a possible add on, like the Omega and Pyramid sound enhancers. Work with any Amiga for quality boost of standard 8 bit playback, with no accelerator overhead. The difference is, it's a proper digital enhancement, and produces much better output. Not as good as a Delfina for quality, but more connectable for those without a clock port.

The hardware to do this WAS ludicrously expensive 20 years ago, but is now affordable. I'm aiming for a $100 price point or less (UK has never had it so good for manufacturing hardware). Maybe a lot less, I don't have current prices for technology required.

Rasp Pi and USB sound stick, pretty much. See if you can spot the "ringer" that has no origin with the Amiga or with a 4 track recorder. Such hardware would boost the output of music demos recorded with such gear, so it's not really just for Amigans.

Tracks are "Jesus on Ees", most of the rest from DJ Randall on the Suburban Bass release (Cu Amiga I think) and the rest are just stuff I like. Should not be too hard to track down and compare (JoE is on youtube, I know that).

Even if you just like 'em, post and let me know. They took a long time to do by hand.

The software used was Sonique, to extract the left and right audio data from a mod, and Audacity to do the processing. Nothing you could not do without the right hardware plugged into the audio ports of an Amiga.

Enjoy.

[Amiga1992]

All the JEsus On Es files sound horribly distorted, how are they "remastered"?
The random OGG I downloaded sounded OK.

[pandy71]

Originally Posted by Pat the Cat

The difference is, it's a proper digital enhancement, and produces much better output.

Are you able to elaborate on this? What kind technology is used to enhance audio (something like spectral band replication)?

[Pat the Cat]

Originally Posted by Akira

All the JEsus On Es files sound horribly distorted, how are they "remastered"?
The random OGG I downloaded sounded OK.

Yes digital delay cycle was turned off with nearly all the Oggs. Try the Electric.flac.

Delay helps with conventional domestic audio equipment, but it's really the original 8 bit 4 channel mix that is "distorted" in some ways. I also had to give a realistic idea rather than everything simulated perfectly. Real world hardware rarely turns out better or the same as paper design hardware.

I invite people to try with headphones and hear the differences between LSD 8 bit and LSD DR 8 bit. The original Amiga version does indeed sound horribly distorted. The DR 8 Bit sounds wonderful in comparison.

[ Show youtube player ]

I think Akira is listening on a very good amplifier and some very nice speakers. Celestions, perhaps? Most people don't have them, at £18,0000 per speaker. Bose much more affordable and 99.99% as good.

So, hardware maybe needs a delay on / off control. Then again, if you can afford high quality audio gear, you would be using a Delfina and outputing 16 bit stereo in the first place.

Originally Posted by pandy71

Are you able to elaborate on this? What kind technology is used to enhance audio (something like spectral band replication)?

I am happy to elaborate, and NO analog was used. The enhancement done is purely digital. I haven't heard of spectral band replication as such, so it might be the same.

First, I took a mod, and extracted the digital data for left and right channels. This was saved in 16 bit format, using Sonique (can use other packages, just what I had to hand).

So the Amiga 8 bit is now 16 bit, but the sample rate is still whatever was used to make it with. Two huge great wav files.

Then, these were loaded into Audacity, which in itself converts those into 32 bit samples with a sample rate of 44.1 KHz.

After that, digital effects using Audacity were added, but the main point is, a CPU was used to "redraw" the samples from an 8 bit workspace into a 32 bit workspace at a much higher sample rate. I use the word "redraw", you could also use the word "calculate".

A little bit like zooming in on a fractal program.

The final part of the mix is the delay. One side is mixed to the other, either in perfect sync, or with a little delay.

Then you compress and expand it so that it has IMPACT. A real musical technician would just say "compress it to fuck", but I am not a musician, or a studio engineer, I just hung out with quite a few of them.

Finally the combined sound is output to both audio outputs. Effectively, it's just repeated mono, but it still started as 4 separate sounds so the mind doesn't notice, it hears multiple sounds across multiple channels, so does not recognize it as such. A delay takes out the inherant clicks, but will upset musical purists who are used to the snap, crackle and pop of Amiga tunes.

Quieter bits sound louder, louder bits sound quieter.

It's not quite what a pro would use in a sound studio, but in terms of coming up with hardware to do that, it is all just a digital process. No filtering involved. It is broadly similar to the studio recording process, but very much simplified and automated.

The Omega and Pyramid were, from what I gather, analog frequency filters and boosters. They did take noise out but they also dumped some frequency responses, which is why there are two versions of the Omega.

This is better than either, as a concept, but you can do EVEN BETTER with just software. alone. Right at the beginning, you seperate each of the four channels seperately and clean them up sepeately. Then mix them. You just can't do it in realtime with hardware without taking inputs from the top of the Paula chip or similar. Far easier to build a box that plugs into the sound phono sockets on the outside. Works any Amiga. Works any stereo audio output.

Whether "works" is what you want is different, it's not aimed at deep pockets. It was my idea of how Omega SE worked when I first heard about it. Turns out, my idea was better, just far too expensive to implement at the time.

[jarp]

So effectively you just applied heavy compression and 100% crossfeed?

[Pat the Cat]

Originally Posted by jarp

So effectively you just applied heavy compression and 100% crossfeed?

Not quite. It's a subtle process finding the rigtht values and procedure that work with Amiga sound outputs.

You can't just compress as a first step, that really screws things up.

It's not like this is fresh work. It's something I did years ago. Don't quiz me on the details, I didn't write them down at the time.

I guess you could reverse engineer the Flac with the original to work them out, but that's just as much work as playing with Audacity or similar until you get something that works.

Anybody thinks they can do better, post a link to a lossless version of Electric Bad Boy that sounds better than mine. That could be done, but it's not just a copy and paste exercise. Talk is cheap.

[Locutus]

And here i was thinking Amiga mods should just sound as they come out of the A500/1200 audio outputs. Dont i feel stupid now

[Amiga1992]

Originally Posted by Pat the Cat

I think Akira is listening on a very good amplifier and some very nice speakers. Celestions, perhaps? Most people don't have them, at £18,0000 per speaker. Bose much more affordable and 99.99% as good.

Nope. Laptop speakers. Sound is really bad (don't want to ruin my monitor speakers with it tbh).
I think your problem is the source. This doesn't sound like this at all on an Amiga. The audio I tried it's clipped to fuck, so of course it is distorted. Did you not say you recorded from some sort of modern PC/Mac player software?

[Pat the Cat]

Originally Posted by Akira

Nope. Laptop speakers. Sound is really bad (don't want to ruin my monitor speakers with it tbh).
I think your problem is the source. This doesn't sound like this at all on an Amiga. The audio I tried it's clipped to fuck, so of course it is distorted. Did you not say you recorded from some sort of modern PC/Mac player software?

YOU ARE LISTENING ON LAPTOP SPEAKERS? Plug headphones in, for goodness sake. Or a Hifi stereo amp. OK, it will sound OK on a huge laptop with a built in sub bass unit, but 99% of laptops do not have those and screw up anything played through them to a certain extent.

No, the data started as a set of 16 bit samples extrapolated from an original MOD file, just using software alone (Windows Sonique, in this case) on the original mod file. No hardware resampling was done AT ALL with these remasters.

No point, it would have made things worse. That's actually the question mark in getting a hardware package to sound as good as the FLAC. I suspect you need to sample at huge sample rate in 32 bit to really make it work. Get true 16 bit stereo from four 8 bit channels.

That's pretty much how CBM planned to improve the Amiga sound, with an onboard DSP doing the work. That's what they planned for AAA, but I don't know if AAA even had a Paula replacement. It didn't really need one unless you were a professional sound engineer. Or say, a naval sonar operator or other Electronic Warfare person. All my opinion, of course. Dave Haynie would know, he was the chief. He's probably unaware of the military applications. Tech has moved on, but at the time I guess some Admirals or higher crapped themselves when they worked out the potential, and what it meant to the arms race. with the Berlin Wall tumbling down. CBM had links to USN from the early days. And those "Admirals or higher" might not have been USN. Oy vais, they might have technically been Jewish, but then again so am I on a DNA test. Goes with the surname if your ancestors have been playing straight, on the level.

I'm not pointing the finger at ANYBODY, OK? American, British, whoever. Water under the bridge, all that stuff. If it really happened that way, I'm not bothered. I can't change the past.

[Locutus]

what? are you sure you have taken your meds?

[Amiga1992]

I know laptop speakers are terrible, that's not what I normally use (have both high quality Hi Fi and a pair of good reference monitors), but that has nothing to do with audio clipped to hell and distorting. At least not in my case.

Your wave file above, my recording, taken just now directly from my Amiga 600 through a MOTU828 interface, below (all I did was normalize it and I accidentally mixed it to mono :P)


Clearly whatever you did has clipped the sound and this creates distortion. It has nothing to do with my speakers or with the Amiga's original sound. Look at how much is lost if you compare, your wave looks weirdly brickwalled. You said you didn't want to compress to shit? It looks exactly like if you did, just look at those damn waves. The quieter parts are so loud it makes no sense (compare with my wave).

I mean look at all the shit you have lost, you killed all the dynamic range.

Also your recording has a weird pitch, not the original. You must have altered it with some of your process.
Here's the recording from my computer (ie the only irrefutable reference anyone should ever need):
https://we.tl/XH1yjyy6jP

Have to agree with Locutus anyway, about whatever you were saying on the end of your message, which has fuck all to do with this board. Focus on your "remaster" topic.

[Pat the Cat]

Uhuh. Tried the headphones?

Come to that, try sending it all through a statistical wavelength analyzer. The Flac. A comparison with the audio waveform of the original Amiga mod should reveal some very interesting facts, if you want to rate true sound quality and which frequencies are present where.

If you've got one of those lying about, of course. Bit tricky if you haven't. Software can do that, but I don't know what to recommend for doing that, I just use the hardware built into my head. It's called brain and ears. It works pretty good, as a combination.

Check out the waveform of a Prodigy track some time. They really know how to do this stuff.

Originally Posted by Locutus

what? are you sure you have taken your meds?

Let's put it this way, a person in my position, with my interests, skills, and history, has had some very interesting job offers. I never pursued some of them, always wanted to retain my independence and ability to comment on subjects.

I've asked for a couple independent experts to comment, they may do so, they may not, but they'll probably at least read it.

[Amiga1992]

Originally Posted by Pat the Cat

Uhuh. Tried the headphones?

Come to that, try sending it all through a statistical wavelength analyzer. The Flac. A comparison with the audio waveform of the original Amiga mod should reveal some very interesting facts, if you want to rate true sound quality and which frequencies are present where.

If you've got one of those lying about, of course. Bit tricky if you haven't. Software can do that, but I don't know what to recommend for doing that, I just use the hardware built into my head. It's called brain and ears. It works pretty good, as a combination.

Let's put it this way, a person in my position, with my interests, skills, and history, has had some very interesting job offers. I never pursued some of them, always wanted to retain my independence and ability to comment on subjects.

I've asked for a couple independent experts to comment, they may do so, they may not, but they'll probably at least read it.

Clearly your ears and eyes are malfunctioning if what you hear and see posted in my previous message doesn't make you realize what you did has actually destroyed the sound.
I gave you both a graphical comparison and an actual soundfile coming from actual hardware, the REAL deal. But I dunno what the hell you are talking about, truly rambling right now, perhaps it's way past your bed time, being that you are on GMT.

[demolition]

Doesn't need to listen to it. That plot clearly shows how compressed/clipped the track is. Perhaps this is intended, but no point in pretending it is not dynamically compressed since it clearly is.

I also have a really hard time understanding what you're writing..

[Pat the Cat]

Originally Posted by Akira

Also your recording has a weird pitch, not the original. You must have altered it with some of your process.
Here's the recording from my computer (ie the only irrefutable reference anyone should ever need):
https://we.tl/XH1yjyy6jP

Have to agree with Locutus anyway, about whatever you were saying on the end of your message, which has fuck all to do with this board. Focus on your "remaster" topic.

Excuse me, why should I have to sign up for terms of service for a scientific proof of fidelity with a Public Domain work?

I'm not even going to DL that. People can get the original mod. Some of us can even rip it from the original disks - which I didn't do, I must admit. I used the mod I got from a collection, so it might not even be the original.

But it's still fucking kickass. They all are. Spotted the ringer yet?

[Amiga1992]

There's nothing to download if you just want to look at the clear waveform evidence.
It's not like you are going to be able to hear that your audio is at the wrong pitch anyway. But that's OK, I guess.

Gotta get those eyes and ears checked, man. Getting old is a bitch.

[Pat the Cat]

Well, anybody thinks it's crap is entitled to say so.

Go study Prodigy. Those guys don't need sound engineers when they do music. They ARE sound engineers, of incredible stature. Study from the best. You might learn something.

[Pat the Cat]

Originally Posted by demolition

Doesn't need to listen to it. That plot clearly shows how compressed/clipped the track is. Perhaps this is intended, but no point in pretending it is not dynamically compressed since it clearly is.

I also have a really hard time understanding what you're writing..

OK. I don't actually care what algorithm it's using. And my bandwidth is rapidly heading to zero right now for some reason. Certainly not because I'm downloading Akira's file. Might do it when it doesn't cost me bread (I pay per byte, mobile, not cheap).

Any algorithm can be implemented in real time, if the overall dynamics are established, if you throw enough CPU and memory at it, or use a DSP, can be done with hardware.

That's what I'm aiming for. How cheap that is, I'm still not sure. Just that it's cheaper than it was.

That would be awesome for every Amiga sound ever done. It would be awesome for the Amiga software base. It would be an awesome piece of hardware, and not very expensive even to make yourself. A Pi cannot sample on it's own. A USB sound card can. That's cheap kit these days. Arduino is another route, but it's not got the CPU or connectivity unless you go wireless Due. Maybe £20, including a wifi on the sampler needed on the Amiga end.

Think about this stuff. I'm not pretending what I've done is the best possible. The FLAC is an example of the best of what I've come up with.

[demolition]

Lots of Prodigy's music is also heavily clipped, but they never claimed that it isn't. That's how it is intended.

[prowler]

I have moved this thread from the News forum to Amiga scene, because, IMO, the tone of the conversation here is not typical of EAB, nor therefore suitable to be broadcast across the whole of the Abime network.