![]() |
![]() |
#1 |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
Amiga Anti-aliasing filter query
Hi, i appreciate this section is for music making, but i thought people here may know best.?
What is the deal with the Anti-aliasing filters in the Amiga line-up? It appears the A1200 is fine in it's stock form but the others are somewhat crippled by the low-pass filter.. I have looked into it, but things get a bit technical re: modding cap and resistor values... Is this fixable 100% on the A600 or is it just best i get an A1200? Cheers. |
![]() |
![]() |
#2 |
Alien Bleed
Join Date: Aug 2022
Location: UK
Posts: 4,853
|
Do you mean the the reconstruction filter or the lowpass filter?
|
![]() |
![]() |
#3 |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
To be quiet honest, i'm not sure. All i know is there is a difference (high frequency loss) between models, resulting in duller audio. But i'm all ears to hear about both. Cheers! EDIT *having a quick look about i think it would be the "reconstruction filter"* Last edited by Finku; 03 October 2023 at 06:58. |
![]() |
![]() |
#4 |
Alien Bleed
Join Date: Aug 2022
Location: UK
Posts: 4,853
|
It probably is fixable but I imagine it's fiddly work. Not sure I'd want to do it. I don't recall finding the A600 audio that bad to be fair. Depends on what sort of samples you're going to use and whether or not you intend to re-eq the output.
|
![]() |
![]() |
#5 | |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
Quote:
From what i can tell the A1200 has the most "open" sound so i will go for one of them. Of coarse all channels will eventually be tracked out to a modern daw for polishing, or at the very least, rolled in glitter...! |
|
![]() |
![]() |
#6 |
Registered User
Join Date: Aug 2006
Location: Scunthorpe/United Kingdom
Posts: 2,168
|
Original A1000 had a permanent low pass filter. Amiga500/ECS/A600 was switchable (Power LED would change brightness) and the A1200 didn't have it, iirc
https://en.wikipedia.org/wiki/Original_Chip_Set#Audio |
![]() |
![]() |
#7 |
Registered User
Join Date: Sep 2009
Location: Norway
Posts: 1,725
|
Every single Amiga except Amiga 1200 has a fixed low-pass filter tuned at around 4420.97Hz (A1200 = around 33-34kHz). This is why the audio is dull on every other Amiga than A1200. This has nothing to do with the toggleable "LED" filter, and it's one of the reasons why A1200 is my favorite Amiga.
EDIT: Worth mentioning that A1000 and very early A500 machines (rev 3) have a permanently enabled "LED" filter which can't be turned off, so they sound even more dull. Dunny: A1200 has the "LED" filter too. Last edited by 8bitbubsy; 06 October 2023 at 20:04. |
![]() |
![]() |
#8 |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
[ Show youtube player ]
|
![]() |
![]() |
#9 |
Registered User
Join Date: Aug 2023
Location: Scotland
Posts: 35
|
There's an excellent discussion of the filter differences between models in Henryk Richter's Dear Paula, starting on p10. The A1200 does have a fixed LPF, but it's at a sufficiently high frequency that it has very little effect on the audible range.
I first saw this in lft's Paulimba article, which is also worth a read if you're interested in the effects of the filter on sample playback. |
![]() |
![]() |
#10 | |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
Quote:
Great stuff thanks! A lot goes way over my head though.. seems odd they jumped from a very conservative 4khz to 27/34khz. You would think maybe try 14khz and see... Makes me think it would be great to have a variable capacitance, resistace box tapped into the Amiga output circuit.. |
|
![]() |
![]() |
#11 |
Registered User
Join Date: Sep 2009
Location: Norway
Posts: 1,725
|
I'm guessing 34kHz on the A1200 (it's never 27kHz AFAIK, the schematics are wrong) is about as high as you can go while properly filtering out the volume PWM artifacts (which I think resides in 55kHz+). Remember, this is a 1-pole 6dB/oct lp filter, so the filtering slope is not steep. You need to set the cutoff lower than the stuff you actually want to filter out.
EDIT: Speaking of wrong schematics, there was supposed to be an Amiga 600 rev2 with A1200-like cutoff, but yet again are the schematics wrong, and that revision has 4420.97Hz (or so) like the rest. Last edited by 8bitbubsy; 07 October 2023 at 10:09. |
![]() |
![]() |
#12 | |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
Quote:
From the "Dear Paula" article: • Amiga 500/2000: 4.42 kHz (100 nF, 360 ?) • A600: 4.42 kHz (100 nF, 360 ?) 11 • A1200 Rev1d: 27.7 kHz (3.9 nF, 1.5 k?) • A1200 Rev2: 34.4 kHz (6.8 nF, 680 ?) • A4000: 4.52 kHz (47 nF, 750 ?) *A500/600/1200: R331/C331,R321/C321, A4000: R430/C430,R440/C440 *a previous version of this text had wrong parts listed, these here I’ve checked |
|
![]() |
![]() |
#13 |
Registered User
Join Date: Sep 2009
Location: Norway
Posts: 1,725
|
A1200 R1D.x is also 34.4kHz, just like R2. I have confirmed this myself by checking various motherboards physically.
Never trust the Amiga schematics. ![]() |
![]() |
![]() |
#14 |
Registered User
Join Date: May 2023
Location: UK
Posts: 13
|
|
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
A cool new filter, anti-dither | turrican3 | support.WinUAE | 2 | 13 August 2022 22:59 |
Anti Aliasing function | 21stcentury | support.WinUAE | 7 | 17 March 2022 21:16 |
Possibilities for integer audio scaling to minimize aliasing | Dr.Venom | support.WinUAE | 5 | 13 September 2020 20:33 |
Full-scene anti-aliasing (FSAA) | Keylonta | support.FS-UAE | 1 | 11 December 2016 23:16 |
|
|