English Amiga Board Amiga Lore


Go Back   English Amiga Board > Support > support.WinUAE

 
 
Thread Tools
Old 14 March 2017, 17:52   #21
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Posts: 442
Quote:
Originally Posted by demolition View Post
The low pass filter can be considered a part of the DAC circuit, also known as a reconstruction filter:
https://en.wikipedia.org/wiki/Reconstruction_filter

Mirror images of the frequencies in the audio signal are created around the half of the sampling frequency of the DAC and some kind of low pass filter must be used if you want to remove these, or at least attenuate them.

This theory also applies to the samples being played back which are played at a lower sampling rate than the DAC/Paula itself, thus some kind of sample rate conversion is involved (and nearest neighbor is a very crude way of doing it). The overtones from the sample rate conversion is probably the main reason why you would want a low pass filter, however ideally the filter should be applied separately to each channel with a cut off frequency tracking each sample's sampling frequency, so having just one filter and being able to turn it on and off was probably as good as what could realistically be done.
Doesn't the Creative Sound Blaster (or GUS or any PC sound card) work along similar principles on playing back samples at different sampling rates?

The only problem I have with Amiga audio is that low KHz samples can sound bad when played back at lower frequencies. You get this distortion that is audible at low frequencies and gradually fades out as the frequencies get higher. The bass notes in the first minute of the first tune of Odyssey by Alcatraz is a prime example (listen for the first minute): [ Show youtube player ]

Sometimes I swear I can hear other musical notes in particular samples that I think weren't meant to be there, but I think that's down to the sampling source.

In all, I've always been impressed by Paula, but for someone to come along and say it wasn't as good as it should've been is an insult to Jay Miner and the other Amiga engineers.
Foebane is online now  
AdSense AdSense  
Old 14 March 2017, 21:32   #22
demolition
Unregistered User
demolition's Avatar
 
Join Date: Sep 2012
Location: Copenhagen / DK
Age: 37
Posts: 3,254
Quote:
Originally Posted by Foebane View Post
Doesn't the Creative Sound Blaster (or GUS or any PC sound card) work along similar principles on playing back samples at different sampling rates?

In all, I've always been impressed by Paula, but for someone to come along and say it wasn't as good as it should've been is an insult to Jay Miner and the other Amiga engineers.
Yes, other sound cards (all DACs) work along the same principles, however SB does not do HW sample playback at different sample rates as far as I know. It is all mixed in software, so the end quality is very much dependent on the playback routines used in the SW. GUS can do it in HW, and if you compare the GUS sound to Paula, you can hear that the GUS has a cleaner sound with fewer overtones because the filtering is more accurate/correct.

I'm sure that no Amiga engineers would be insulted by me calling Paula imperfect because they were fully aware of that when they designed it. Nothing is perfect and the product they ended up with was a result of various compromises which they thought provided the best overall performance, and it so happened that many people liked the choices they made. This is how all engineering works, even serious stuff like the space shuttle program was littered with compromises and imperfect solutions.
demolition is offline  
Old 14 March 2017, 22:01   #23
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Posts: 442
Human technology is not perfect and full of compromises. Oh, well.
Foebane is online now  
Old 14 March 2017, 23:06   #24
Akira
Registered User

Akira's Avatar
 
Join Date: May 2001
Location: New York
Posts: 17,689
Quote:
Originally Posted by Foebane View Post
Why did Commodore even PUT filters into the Amiga machines to begin with? I doubt it was an Amiga decision.
I am pretty sure you are wrong in this assumption.
If I am not 100% wrong, the sound hardware design was at least in some way influenced or initiated by Bob Yannes, and when he left C= and went to form Ensoniq, he did the exact same thing on the Mirage, only that instead of a shitty cheap fixed filter, he put in a really nice Curtis controllable low pass filter. The way the Mirage works is very similar to the way the Amiga works.

Why is it there? Sampling at such low resolution brings a shit ton of noise and overtones and a low pass filter makes this issue a bit better.

Quote:
Originally Posted by Foebane View Post
Amiga sound is perfect, I never hear any "overtones". At best, there is that irritating distortion you get with some lower Khz samples, but that was down to memory constraints.
As said by demolition, even if you don't hear them (are you a spectrum analyzer? ) they are there and don't help, and also the "distortion" you hear is also due to low quality DAC. This is an effect where a low pass filter can help a lot in smoothing it out. The kind of DAC the Amiga has really limits the dynamic range, anything subtle is just a mass of noise. So if you have to sample at a really high volume, an analog low pas filter can help you mellow it out, analogically, without digital artifaction.
That distortion you mention on tha Alcatraz demo, is because someone had to make that bass audible.

Try to load a perfect-looking sine wave as an Amiga sample into a tracker and try to make it be heard, it'll be real fucking hard unless you start distorting it until it approaches a pulse more than a sine.

In brief, the Amiga sound is FAR from perfect, but that isn't necessarily a problem.
If you want to wonder how the AMiga sound hardware could have been with not as many budget restrains put on it, check out Yannes' Mirage and see how far the Amiga sound could be taken with very few differences (mainly an improved frequency response and the controllable, resonant and ultimately expensive Curtis low pass filter)
Akira is offline  
Old 15 March 2017, 06:45   #25
demolition
Unregistered User
demolition's Avatar
 
Join Date: Sep 2012
Location: Copenhagen / DK
Age: 37
Posts: 3,254
Quote:
Originally Posted by Foebane View Post
Human technology is not perfect and full of compromises. Oh, well.
Not just human tech. Nature is also not perfect and is full of compromises and suboptimal solutions. The point is that it doesn't need to be perfect - it just needs to be good enough.
demolition is offline  
Old 15 March 2017, 11:41   #26
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Posts: 442
Quote:
Originally Posted by Akira View Post
Why is it there? Sampling at such low resolution brings a shit ton of noise and overtones and a low pass filter makes this issue a bit better.
I always thought the filter muffled the audio, so I wondered why it was there. Then again, 8-bit at 8KHz (average) is rather low.

Quote:
Try to load a perfect-looking sine wave as an Amiga sample into a tracker and try to make it be heard, it'll be real fucking hard unless you start distorting it until it approaches a pulse more than a sine.
If sound samples had to be distorted before they sounded good on Paula, then how did the sound sampler cartridges of the day work? I had one, and I sampled many movie and TV show clips, and I always figured the poor-to-average quality was down to the 8-bit nature at such low sample rates.
Foebane is online now  
Old 15 March 2017, 15:26   #27
Akira
Registered User

Akira's Avatar
 
Join Date: May 2001
Location: New York
Posts: 17,689
Quote:
Originally Posted by Foebane View Post
I always thought the filter muffled the audio, so I wondered why it was there. Then again, 8-bit at 8KHz (average) is rather low.
Yes, you are right, the filter will muffle the sound, but what this filter is trying to muffle is stuff that doesn't need to be heard.
So generally speaking, yeah, an Amiga sounds a bit muffled on the high end, a bit "muddy". But so did oldschool 8 and 12 bit samplers like the Mirage or the Emu SP-1200.
Quote:
If sound samples had to be distorted before they sounded good on Paula, then how did the sound sampler cartridges of the day work?
I don't really understand your question. Yes, the low volume parts, will be grainy/noisy/shitty, even if you sample and playback at the highest rate possible on an Amiga. Yes, resolution and sample rate of the samples has to do with it too.
But even if you had a 16-bit sampler like a Clarity 16, I am inclined to think that as long as you reproduce them with Paula you will run into the same issues.

I'm not a deep tech here, so I cannot give you a more in-depth explanation, but someone on the board like pandy71 would be able to do so if that's what you want.
So don't take my word for it, but I think this is because the Amiga uses pulse width modulation to reproduce audio. Do you have a Raspberry Pi with an audio out? If you do, you can see what I mean. Anything reproduced at top volume will sound OK, but start lowering the volume, and you will hear the problems. The Raspberry Pi has a really crud piezo pwm sound thingy which works in a similar way.
Akira is offline  
Old 16 March 2017, 08:38   #28
Foebane
Registered User
 
Join Date: Sep 2011
Location: Cardiff, UK
Posts: 442
Quote:
I don't really understand your question. Yes, the low volume parts, will be grainy/noisy/shitty, even if you sample and playback at the highest rate possible on an Amiga. Yes, resolution and sample rate of the samples has to do with it too.
I know Paula is 8-bit, basically meaning up to 256 possible speaker positions, so I know the resolution would be very poor indeed, compared to CD Audio (16-bit, 65536 possible speaker positions).

Quote:
But even if you had a 16-bit sampler like a Clarity 16, I am inclined to think that as long as you reproduce them with Paula you will run into the same issues.
Of course, because they'd be downgraded (de-rezzed, if you like) from 16-bit to 8-bit. Then again, the 14-bit Paula audio might mitigate that problem a bit.

Quote:
but I think this is because the Amiga uses pulse width modulation to reproduce audio.
I always heard it was PCM, or Pulse Coded Modulation.
Foebane is online now  
AdSense AdSense  
 


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

Similar Threads
Thread Thread Starter Forum Replies Last Post
"Approximate A500/A1200 or cycle-exact" at +500% CPU Speed Foebane support.WinUAE 2 28 January 2017 08:11
Information about filter EMI "TDK 682" Franchute13 support.Hardware 10 21 January 2015 12:23
How do you make a filter setting "stick" to a configuration? ral-clan support.WinUAE 0 11 September 2011 20:51
Ian "tinkering with Amiga's since 1986" Stedman's A500/A600/A1200 ATX power adaptor Paul_s MarketPlace 0 05 October 2010 22:39
is it good or bad to use a a1200 PSU"brick" with an A500? lost_lemming support.Hardware 7 22 February 2010 21:44

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 00:33.


Powered by vBulletin® Version 3.8.8 Beta 1
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Page generated in 0.15383 seconds with 11 queries