English Amiga Board


Go Back   English Amiga Board > Main > Amiga scene

 
 
Thread Tools
Old 25 October 2019, 10:27   #861
Steril707
Tigerskunk!

Steril707's Avatar
 
Join Date: Sep 2016
Location: Amiga Island
Posts: 1,461
Not that anyone thinks I would have been against 16 bit sound quality on the Amiga.
What I am saying is, that simply adding a second Paula would have done wonders, and should have been doable without too much trouble, I guess.

I love using my Elektron Octatrack, and this thing also only has 8 Sounds playing at the same time (those are stereo, though) which I use for mixing down whole tracks.
Steril707 is offline  
Old 25 October 2019, 10:45   #862
no9
Registered User

no9's Avatar
 
Join Date: Feb 2018
Location: Poland
Posts: 90
Precisely. We are talking about maximising benefits for the users at minimum expense from company point of view. Not saying that the 1992 computer didn't deserve whole new sound chip.
no9 is offline  
Old 25 October 2019, 10:52   #863
Thomas Richter
Registered User
 
Join Date: Jan 2019
Location: Germany
Posts: 415
Quote:
Originally Posted by Steril707 View Post
What I am saying is, that simply adding a second Paula would have done wonders, and should have been doable without too much trouble, I guess.
I afraid it isn't quite that simple. A second Paula would have required four additional DMA slots to receive sample data, and thus Alice (or Agnus) would have to be expanded as well. In particular, you would need to allocate time slots for the four additional DMA channels from somewhere.
Thomas Richter is offline  
Old 25 October 2019, 10:55   #864
Hewitson
Registered User
Hewitson's Avatar
 
Join Date: Feb 2007
Location: Melbourne, Australia
Age: 37
Posts: 3,709
Quote:
Originally Posted by jizmo View Post
Yes, almost as outrageous than asking a 1985 machine to have four channels of sampled stereo sound – but that's what the original Amiga still delivered.
No. The Amiga delivered 4 channels of mono sound. Two output to the left RCA output, two output to the right. That is NOT stereo.
Hewitson is offline  
Old 25 October 2019, 11:54   #865
Steril707
Tigerskunk!

Steril707's Avatar
 
Join Date: Sep 2016
Location: Amiga Island
Posts: 1,461
Quote:
Originally Posted by Hewitson View Post
No. The Amiga delivered 4 channels of mono sound. Two output to the left RCA output, two output to the right. That is NOT stereo.
Two channels of stereo then...
Steril707 is offline  
Old 25 October 2019, 11:58   #866
no9
Registered User

no9's Avatar
 
Join Date: Feb 2018
Location: Poland
Posts: 90
Quote:
Originally Posted by Thomas Richter View Post
I afraid it isn't quite that simple.

But do you think it is impossible? For C= back then.

For games (sound effects or even some instruments) even half-rate Paula (let's say "28kHz/2" per channel) would still be useful.
no9 is offline  
Old 25 October 2019, 12:05   #867
Steril707
Tigerskunk!

Steril707's Avatar
 
Join Date: Sep 2016
Location: Amiga Island
Posts: 1,461
Quote:
Originally Posted by no9 View Post
But do you think it is impossible? For C= back then.
What I thought, as well...
Would have been more work, but definitely doable.
Steril707 is offline  
Old 25 October 2019, 12:56   #868
Gorf
Registered User

 
Join Date: May 2017
Location: Munich/Bavaria
Posts: 1,144
Quote:
Originally Posted by Steril707 View Post
Not that anyone thinks I would have been against 16 bit sound quality on the Amiga.
What I am saying is, that simply adding a second Paula would have done wonders, and should have been doable without too much trouble, I guess.
A "second Paula" would simply not work in the design of the Amiga-Custom Chips and Commodore had lost the original designer including documents ...

They should have hired Glenn Keller back (as they did with Carl Sassenrath for the CDTV) and upgrade Paula.

The obvious trick would have been to use "double cas" on that 16bit-chip as they did with Lisa - well Lisa is 32-bit wide in addition but uses two fetches per time-slot to get 64 bit each slot.

Double-Cas could have been introduced as early as 88 or 89, since RAM was fast enough by than ... so the chipset could have stayed 16-bit but Paula and Denise would double the throughput.
So 8 channels for Paula and 64 colors in high-res via Denise...

That is what 16bit ECS in 1988 should have been - followed by 32bit AAA in 91 or 92 skipping AGA entirely.

(the A3000 should have come with VRAM for Chip-Ram instead of Amber Flicker-Fixer allowing 1024x768 in parallel to old modes - see discussion about the use of UHRES here on EAB)

How Glenn Keller did Paula:

[ Show youtube player ]

Last edited by Gorf; 25 October 2019 at 13:35.
Gorf is offline  
Old 25 October 2019, 14:53   #869
8bitbubsy
Registered User

8bitbubsy's Avatar
 
Join Date: Sep 2009
Location: Norway
Posts: 1,360
Quote:
Originally Posted by Bruce Abbott View Post
Computer Games World in October 1993 had this to say about the Gravis Ultrasound:-

"Due to the need for many TSR's, lack of publisher support, and poor Sound Blaster emulation, we find it hard to recommend this card to anyone other than a Windows MIDI musician."

So which PC's came with it? None, you had to pay US$199 if you wanted one. But due to lack of compatibility you probably also had to buy a somewhat more compatible card like the Sound Blaster 16 (US$249). So now you are up for US$448, plus lots of hassle and still no guarantee that all your games will work.

32 16-bit 44.1kHz hardware channels would be pretty useless if none of your games used them, and even more useless if it couldn't even emulate the original Paula sound properly. Not 'upgrading' the sound in the A1200 might have disappointed some people, but it had one huge upside - compatibility. Amiga sound cards were available, but since all Amigas already had a Paula inside we didn't have to worry about compatibility. And 27 years later we still don't have to worry about it.
Yeah, in terms of games it wasn't that successful, but it sure was in the tracker scene, which was about making music.

Quote:
Originally Posted by Steril707 View Post
The Amiga was kind of the birth place for a lot of 90ies sample based music, Jungle and DnB for example.
It was a cheap sampler for its time, and usually people really didn't care that much about the 8 bit, or even embraced the rawness of its sound.
The hiss you hear (especially on bass or low-amp samples) is basically errors in the quantization rounding. In my opinion it doesn't show the "rawness" of the sample, instead it's just a bunch of errors you wouldn't want to hear, but the noisefloor of 8-bit PCM is really high so yeah, that happens...

Quote:
Originally Posted by Hewitson View Post
No. The Amiga delivered 4 channels of mono sound. Two output to the left RCA output, two output to the right. That is NOT stereo.
I'd say it's still stereo. If you use two channels (L+R), you can actually freely pan around left and right by changing the volumes accordingly.

Last edited by 8bitbubsy; 25 October 2019 at 14:59.
8bitbubsy is offline  
Old 25 October 2019, 15:20   #870
daxb
Registered User
 
Join Date: Oct 2009
Location: Germany
Posts: 2,535
Quote:
Originally Posted by jizmo View Post
I'd say there were easily ten times more hobbyist musicians using Amiga primarily for composing, remixing and sampling than there were people using Amiga for spreadsheets, desktop publishing or even creating art in DPaint.
Yes, I would say it was the largest computer music scene worldwide at that time. Just have a look at Aminet uploads each month. Or other CD releases with a lot of mods. Then there are a massive amount of never released stuff. Or just have a look at https://amp.dascene.net/

The 8 Bit 4 channel Paula might be enough in 1993 but 2-3 years later it wasn't. I used it for making music (Protracker and later DigiboosterPro) a big amount of my live time. I still listen to recorded music through Paula each day.

For games it wasn't or still isn't important to have more then 8 Bit and 4 channels, because it only makes a "small" part of a game in most cases. Having music and sfx at the same time doesn't make a game good. Today you can listen to your favourite music while playing a game.
daxb is offline  
Old 25 October 2019, 16:28   #871
idrougge
Registered User
 
Join Date: Sep 2007
Location: Stockholm
Posts: 3,929
Quote:
Originally Posted by Bruce Abbott View Post
But only because it was there first, not because it was a better machine for music.
Ah, but it was if you were using MIDI. Unlike the Amiga, the ST had a cheap, sharp, high-resolution monochrome screen without any flicker or 15 kHz whine which would be unwelcome in a studio setting.

Quote:
Amiga did have that, but expecting it from an unexpanded A1200 (still at the same low price) is asking a bit much. To complete the system you would need a CD writer, a huge hard drive, and probably more RAM and a faster CPU, as well as the 16 bit codec - all of which was available to those who were willing to pay for it.
Or you could get a Falcon, which had SCSI as standard, a wide range of MIDI and hard disc recording software, VGA screen and a DSP for effects.
idrougge is offline  
Old 25 October 2019, 16:35   #872
idrougge
Registered User
 
Join Date: Sep 2007
Location: Stockholm
Posts: 3,929
Quote:
Originally Posted by no9 View Post
I agree with all that you wrote there. And I think 16-bit sound have it's use but only in music niche. So it is difficult to argue for '16-bit is always more useful and versatile than 8-bit in sound' since it was pointed many times in this thread that in game business it is not. Amiga games was far from utilize already existing potential of Paula if we talking on bits and sample rates. But those same games were constantly choking on 4 voices sound limit since a decade because that dilemma ('music or sfx?') was well known even on 8-bit computers.
Indeed, I think most users would prefer 8 channels of 8-bit sound over 4 channels of 16-bit sound, especially as you could combine those channels into 4 channels of 14 bits anyway.
idrougge is offline  
Old 26 October 2019, 21:29   #873
Bruce Abbott
Registered User

Bruce Abbott's Avatar
 
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 328
Quote:
Originally Posted by idrougge View Post
Ah, but it was if you were using MIDI. Unlike the Amiga, the ST had a cheap, sharp, high-resolution monochrome screen without any flicker or 15 kHz whine which would be unwelcome in a studio setting.
The A1200 can do the same resolution and frame rate (640x400 at 72Hz) as the ST, but with a lot more colors.

So it boils down to simply that Atari had it first.

But I think think the key factor was having MIDI ports built into the ST. It certainly wasn't its sound capabilities, or the disk drive (which was originally only 360k), or the non-upgradeable memory, or the clunky operating system, or the crappy keyboard and horrible mouse. But a musician saw those MIDI ports and thought - "this is the machine for me!" (even though other computers could do it with a $15 cable).

So the ST got a foothold in the music market despite having much worse sound than the Amiga.

Quote:
Or you could get a Falcon, which had SCSI as standard, a wide range of MIDI and hard disc recording software, VGA screen and a DSP for effects.
The Falcon was in a different league to the A1200, but it had its problems...

The Atari Falcon030: A Closer Look
Quote:
As you may recall, I sold off all of my Atari equipment and bought an Amiga 1200 back in January, after the Falcon had been delayed several times. (Its original release date was September, 1992). It seemed only natural to take a look at the machine I had intended to purchase, when I decided it was time to move on. Now I'll take a closer look and decide if I made the right decision...

It has two problems right now... First, since there are no screen accelerators yet, loading MultiTOS only slows the system down further. Secondly, using it without memory protection will eat 800K of RAM. Enabling memory protection will eat another 800K. So on a 4 meg machine, MultiTOS is nearly useless...

I can pick up some used 4 meg SIMMs on Usenet a lot cheaper than buying one of Atari's boards. So, for about $1500, I could have a Falcon with 14 meg of RAM and an external 240 meg hard drive, or an internal 120 meg unit. But unless I come into some serious money...
No wonder it was discontinued only one year after being released.
Bruce Abbott is offline  
Old 26 October 2019, 21:54   #874
Bruce Abbott
Registered User

Bruce Abbott's Avatar
 
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 328
Quote:
Originally Posted by 8bitbubsy View Post
The hiss you hear (especially on bass or low-amp samples) is basically errors in the quantization rounding. In my opinion it doesn't show the "rawness" of the sample, instead it's just a bunch of errors you wouldn't want to hear, but the noisefloor of 8-bit PCM is really high so yeah, that happens...
It's noticeable if the sample is recorded with too low sample rate, or if samples are mixed at too low volume.

With samples recorded at full volume the dynamic range should be at least as good as cassette tape, and with their playback volume controlled by Paula it's even better. If you can hear quantization noise it's probably due to poor samples or a bad player.
Bruce Abbott is offline  
Old 27 October 2019, 12:46   #875
daxb
Registered User
 
Join Date: Oct 2009
Location: Germany
Posts: 2,535
I agree to what 8bitbubsy says. Samples are not always using 100% volume amplitude. I would say the most don't. So you can't get always 8 bit resolution. Further, if you have a full amplitude sample and do a volume fade out in Protracker for example you will hear noise at the end part. The mentioned bass samples by 8bitbubsy often or always have a lot of noise with full 8 bit resolution. 8 bit really sucks in this cases.
Code:
 8 bit = 256 possible values
16 bit = 65536 possible values
24 bit = 16777216 possible values
Count the possible values if you only have 4 bit for the amplitude.
daxb is offline  
Old 27 October 2019, 13:21   #876
Gorf
Registered User

 
Join Date: May 2017
Location: Munich/Bavaria
Posts: 1,144
Quote:
Originally Posted by daxb View Post
I agree to what 8bitbubsy says. Samples are not always using 100% volume amplitude. I would say the most don't. So you can't get always 8 bit resolution. Further, if you have a full amplitude sample and do a volume fade out in Protracker for example you will hear noise at the end part. The mentioned bass samples by 8bitbubsy often or always have a lot of noise with full 8 bit resolution. 8 bit really sucks in this cases.
Code:
 8 bit = 256 possible values
16 bit = 65536 possible values
24 bit = 16777216 possible values
Count the possible values if you only have 4 bit for the amplitude.
that is still only a question of the sampling frequency in the end: you can sample with just 1-bit in high fidelity if you drive your frequency up in the MHz range. The noise is huge, but it is all very far in an unhearable high band, that gets filtered.
See "Super Audio CD".

For 8-bit and a moderate sampling frequency, you can also resort to sound-dithering to compensate for the lower resolution ... but that is of course lot of extra work and was newer used for Amiga samples as far as I know...
Gorf is offline  
Old 27 October 2019, 13:49   #877
chb
Registered User

 
Join Date: Dec 2014
Location: germany
Posts: 250
Quote:
Originally Posted by daxb View Post
I agree to what 8bitbubsy says. Samples are not always using 100% volume amplitude. I would say the most don't. So you can't get always 8 bit resolution. Further, if you have a full amplitude sample and do a volume fade out in Protracker for example you will hear noise at the end part. The mentioned bass samples by 8bitbubsy often or always have a lot of noise with full 8 bit resolution. 8 bit really sucks in this cases.
On the Amiga when using four channels in tracked music, you have additional 6 bits for volume per channel, so best practice is using samples normalized to 100% volume and then use the volume registers to adapt to your needs. A fade out done this way should not at additional quantization noise, as you always use the full 8 bit range.

If you want to sample a whole musical piece with both very loud and rather gentle parts (e.g. a classical piece) within the same eight bit range, you will run into those problems you described, unless you're using an envelope follower or something like that to dynamically adjust the volume.

Last edited by chb; 27 October 2019 at 14:03.
chb is offline  
Old 27 October 2019, 16:44   #878
daxb
Registered User
 
Join Date: Oct 2009
Location: Germany
Posts: 2,535
Quote:
Originally Posted by Gorf View Post
that is still only a question of the sampling frequency in the end: you can sample with just 1-bit in high fidelity if you drive your frequency up in the MHz range. The noise is huge, but it is all very far in an unhearable high band, that gets filtered.
See "Super Audio CD".
Theoretical maybe but you know the practice. I've never seen samples with such high frequencies.
Quote:
Originally Posted by chb View Post
On the Amiga when using four channels in tracked music, you have additional 6 bits for volume per channel, so best practice is using samples normalized to 100% volume and then use the volume registers to adapt to your needs. A fade out done this way should not at additional quantization noise, as you always use the full 8 bit range.
This is similar. Theory against practice. You have samples with several different volume in tracked music. You can't always avoid this. And sometimes you fade out the whole track e.g.. I sampled a lot with my TST2 sampler. The source isn't always optimal. Optimal volume isn't always possible. Further you have to fight against the limited hardware and software. Yes, you can get really good quality, but that doesn't work always.
daxb is offline  
Old 27 October 2019, 16:59   #879
AmigaHope
Registered User
 
Join Date: Sep 2006
Location: New Sandusky
Posts: 646
Quote:
Originally Posted by idrougge View Post
Ah, but it was if you were using MIDI. Unlike the Amiga, the ST had a cheap, sharp, high-resolution monochrome screen without any flicker or 15 kHz whine which would be unwelcome in a studio setting.
That was a bit of an oversight on Amiga Inc.'s part back when they designed OCS. They could have easily hacked in a 31Khz screenmode which is essentially what ECS was. ECS didn't require any more silicon than OCS, it was just tweaking OCS to add "free" features without having to do any redesign of the architecture.

That's why ECS has its 64-color palette limitation in the 35ns pixel modes, it's pulling the palette fetch in the exact same way as OCS, just splitting the bits. There's no reason it couldn't have used a 4096 color palette with some small redesign, but the 64-color hack worked with almost no changes to the design. C= probably could've had it added in the second batch of chips for the A1000 if they'd asked for it (back when they added the EHB tweak, which was a similar zero-cost hack they added after the first batch of Denise).

The most fascinating thing is that Paula was flexible enough to output higher sample rates with zero changes, once ECS let you change the horizontal sync. Another free performance hack that required zero changes to the audio silicon.
AmigaHope is offline  
Old 27 October 2019, 18:12   #880
chb
Registered User

 
Join Date: Dec 2014
Location: germany
Posts: 250
Quote:
Originally Posted by daxb View Post
This is similar. Theory against practice. You have samples with several different volume in tracked music. You can't always avoid this. And sometimes you fade out the whole track e.g.. I sampled a lot with my TST2 sampler. The source isn't always optimal. Optimal volume isn't always possible. Further you have to fight against the limited hardware and software. Yes, you can get really good quality, but that doesn't work always.
I totally agree with you that recording samples in 8 bit is very difficult, as you need perfect recording levels not to loose even more of the already limited dynamic range. However, that does not apply to playback at a comparable scale. You can e.g. record a sample with 16 bit and easily covert it to an 8 bit sample with full volume, without the need for perfect levels - you have enough bits left for normalization. That's similar today - almost everything in the studio is recorded at 24 bit (and processed in 32 bit floating point), to get some headroom for recording and processing; but for the final mix 16 bit is sufficient (even more than). So that rather supports the argument that the grainy, noisy sound heard so often on the Amiga stems from limitations of the sample recording process, not the playback hardware.
chb is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
A1200 RF module removal pics + A1200 chips overview eXeler0 Hardware pics 2 08 March 2017 00:09
Sale - 2 auctions: A1200 mobo + flickerfixer & A1200 tower case w/ kit blakespot MarketPlace 0 27 August 2015 18:50
For Sale - A1200/A1000/IndiAGA MkII/A1200 Trapdoor Ram & Other Goodies! fitzsteve MarketPlace 1 11 December 2012 10:32
Trading A1200 030 acc and A1200 indivision for Amiga stuff 8bitbubsy MarketPlace 17 14 December 2009 21:50
Trade Mac g3 300/400 or A1200 for an A1200 accellerator BiL0 MarketPlace 0 07 June 2006 17:41

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 04:24.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.
Page generated in 0.13942 seconds with 15 queries