IPB

Welcome Guest ( Log In | Register )

Using DirectSound SSRC Plugin, I have many questions.
ViPER1313
post Feb 7 2003, 08:20
Post #1





Group: Members
Posts: 271
Joined: 19-August 02
From: Maryland
Member No.: 3109



As the owner of a SB-Live card, I want to improve the quality of my sound playback. I have Peter's DirectSound SSRC v2.2.6 plugin installed with Winamp v2.81. For audio playback, I use a Kenwood 5.1 S.S. Reciever conected to the SPDIF port on the SB-Live (Digital throughput). The amp supports 96khz / 24bit decoding in hardware (and it does it quite well from what I can see, more on that later..) My questions reside on the settings for using all my plugins in conjunction.

Q1) Where should dithering be applied. For example, the vorbis decoder plugin supports dithering of 8/16/24 bit files, while the DS SSRC plugin also supports dithering of 8/16/24bit. Should I enable dithering in all the separate plugins and disable in the the DS SSRC plugin, or disable it in all input plugins (Vorbis, MPC, AAC) and enable it solely in the DS SSRC output plugin. Or should I enable it in both input / output (double dither???)

Q2) Should the input plugins for Vorbis / MPC / AAC be set to 24bit or 16 bit output. If they are set to 24bit, does dither need to be selected on the input side?? (goes along with question 1...) For reference, all files used are in 44100 /16 bit format. Is it higher quality to set all the input plugins to 16bit and the output to 24 bit / 96000 khz, or is it better to select 24bit on all the input plugins and 16bit / 48000 khz in the DS SSRC output plugin?

Q3) If 24 bit is used on the input plugins as well as 24 bit / 96000 on the output plugin, do you need to use dithering at all?

Q4) Is noise shaping a good or bad thing? Should it be enabled in the output plugin?

Q5) What is noise distribution?? The three options for this setting are rectangular, triangular and gaussian. Which of these three settings offers the highest quality.

Any help would be greatly appreciated.

For reference, I have some screenshots of sweep.wav provided below. These are taken by playing the sweep.wav file through my Kenwood Receiverís headphone jack and recording the sound on a second machine using a SB-16 ISA card @ 44.1 / 16bit. Sweep.wav is a 44.1khz / 16bit file.

Input Plugin Settings (Waveform Decoder v2.05a) / Output Plugin Settings
---------------------------------------------------------------------------------------------------------------------------------------
1. 16 bit output / dither -||- No Resample using WaveOut (Pic1)
2. 16 bit output / dither -||- Resample 48/16bit || Dither w/o noise shaping (Pic2)
3. 16 bit output / dither -||- Resample 96/24bit || Dither w/o noise shaping (Pic3)

(If links don't work, try "Open in New Window" or "Save target As" and open them locally. I've had probs with Angelfire in the past...)

Edit - Small errors....

This post has been edited by ViPER1313: Feb 7 2003, 08:22
Go to the top of the page
+Quote Post
 
Start new topic
Replies (1 - 8)
tigre
post Feb 7 2003, 09:29
Post #2


Moderator


Group: Members
Posts: 1434
Joined: 26-November 02
Member No.: 3890



Q1-3) I don't really remember details - maybe I'm totally wrong, but IIRC in Winamp several steps are done with different resolution, so it can happen, that up- and downsampling is done several times, what can (theoretically) harm quality. If you want to be on the safe side, why not use foobar2000? All steps are done at the same resolution (32bit), downsampling to 16/24 bit output with dither only at the end, what should give unbeatable quality. Anyway, so far I haven't heard/read about blind tests proving that foobar2000 sounds better.

Q4)It depends: Noise shaped dither in last step of processing is good, because the added noise is less audible than white noise. But be sure that you apply noise shaped dither only once. Otherwhise it could result in artifacts (or in case of dithering to 8 bit in harm to your ears/equipment as there'll be a lot of very loud very high frequencies). Since you have 24bit output you can safely use no noiseshaping, because the noise will be much lower than the noise caused by 16 bit resolution of the source (CD, lossy compressed files from CD ...).

Q5)Triangular is best (mathematically proven, IIRC). If you do a search about dither/dithering here you'll find detailed information/links.

This post has been edited by tigre: Feb 7 2003, 09:32


--------------------
Let's suppose that rain washes out a picnic. Who is feeling negative? The rain? Or YOU? What's causing the negative feeling? The rain or your reaction? - Anthony De Mello
Go to the top of the page
+Quote Post
KikeG
post Feb 7 2003, 09:43
Post #3


WinABX developer


Group: Developer
Posts: 1578
Joined: 1-October 01
Member No.: 137



If your decoder supports 24/96, there's no need to use dither. Use 24 bit output in all plugins, and you will get best possible quality.
Go to the top of the page
+Quote Post
ViPER1313
post Feb 7 2003, 21:35
Post #4





Group: Members
Posts: 271
Joined: 19-August 02
From: Maryland
Member No.: 3109



OK, the settings that I'm going with are as follows -

[INPUT PLUGINS]
Nullsoft Vorbis Decoder v1.2.9 - 24bit output / No Dither

MPC MousePack Plugin v0.98 - No dither (Cant use 24bit - stuck using 16 bit)

Nullsoft MP3 Audio Decoder v2.81 - No dither / 16bit output (also stuck using 16bit, and I don't want to use the in_mpg123 plugin, as you cant edit ID3 v1/2 tags in Winamp if you use it)

Nullsoft WaveForm Decoder v2.05a - Don't downmix 24/32 bit files to 16bit and Convert Floating point to 24bit with NO dither.

Audiocoding.com MPEG-4 Plugin - 24bit Output

[OUTPUT PLUGINS]

DirectSound Output v2.2.6 SSRC - Resample Enabled, Sample Rate - 96000Hz, Bits Per Sample - 24, Dither with no noise shaping, Noise Distribution - Triangular, Fast Mode Decoding.


Should these settings be alright? Any single thing that I should do to get better quality? BTW, I tried to disable fast mode decoding, but it causes pops / clicks when decoding Vorbis files, even though CPU usage never went over 10%. I have an AMD Athlon 1400, so the pops / clicks supprise me. Any thoughts?
Go to the top of the page
+Quote Post
KikeG
post Feb 7 2003, 23:18
Post #5


WinABX developer


Group: Developer
Posts: 1578
Joined: 1-October 01
Member No.: 137



If possible, 32 bit output in input plugins should give some theorical advantage over 24 bit, but none in practice I'd say. With 24 bit output in these plugins, use of dither should give also some theorical advantages, but I'd say none in practice too.

For input plugins that are stuck to 16 bit, better use dither. Noise shaping dither is better if you are not doing anything else after the plugin (no resampling, eq. or other DSP). If you do something else, I'd say flat dither (no noiseshaping) would be better in most cases.

Also, there's no need to resample to 96KHz at the output, unless your converter surely works better at 96 KHz. If not, and your converter supports it, use the original sampling rate instead. If you do resample, SSRC fast resampling is OK.

However, there's a high probabilty that no matter if you use dither or not, or 32, 24 or 16 bits in any of the steps, the final perceived audio quality is very similar or the same. Advantages are more theorical than real here, for typical listening conditions.

This post has been edited by KikeG: Feb 7 2003, 23:19
Go to the top of the page
+Quote Post
ViPER1313
post Feb 8 2003, 03:57
Post #6





Group: Members
Posts: 271
Joined: 19-August 02
From: Maryland
Member No.: 3109



QUOTE
Also, there's no need to resample to 96KHz at the output, unless your converter surely works better at 96 KHz. If not, and your converter supports it, use the original sampling rate instead. If you do resample, SSRC fast resampling is OK.

However, there's a high probabilty that no matter if you use dither or not, or 32, 24 or 16 bits in any of the steps, the final perceived audio quality is very similar or the same. Advantages are more theorical than real here, for typical listening conditions.


The reason that I am even going through the trouble to do all of this is because my SB-Live card has a horrible re-sampling algorithm in its hardware. If I left the output at 44.1k, then the SB-Live would resample it to 48khz in hardware. This causes artifacts in the sound, as can be seen in pic one of my first post (It is supposed to be one straight line, as can be seen in pics 2/3). Another question - does the SB-Live allow 24bit / 96000hz sample rates through its SPDIF port? I know that DVD's have 24bit / 96khz sound, and that if I select "USE SPDIF" in a program such as PowerDVD, it will send the 5.1 signal to the receiver (the "Dolby Digital" or "DTS" on the receiver lights up and the speaker picture changes, so I know it is getting the correct signal.) I just hope that other SB-Live people like me can learn something from this thread and get better sound from their cards.

Edited cause I always seem to have one or two small errors in the origional..... headbang.gif

This post has been edited by ViPER1313: Feb 8 2003, 04:00
Go to the top of the page
+Quote Post
Timothyw
post Feb 8 2003, 14:13
Post #7





Group: Members
Posts: 14
Joined: 1-February 03
From: St Albans/Brighton (Sussex Uni)
Member No.: 4828



QUOTE (ViPER1313 @ Feb 7 2003 - 06:57 PM)
QUOTE
Also, there's no need to resample to 96KHz at the output, unless your converter surely works better at 96 KHz. If not, and your converter supports it, use the original sampling rate instead. If you do resample, SSRC fast resampling is OK.

However, there's a high probabilty that no matter if you use dither or not, or 32, 24 or 16 bits in any of the steps, the final perceived audio quality is very similar or the same. Advantages are more theorical than real here, for typical listening conditions.


The reason that I am even going through the trouble to do all of this is because my SB-Live card has a horrible re-sampling algorithm in its hardware. If I left the output at 44.1k, then the SB-Live would resample it to 48khz in hardware. This causes artifacts in the sound, as can be seen in pic one of my first post (It is supposed to be one straight line, as can be seen in pics 2/3). Another question - does the SB-Live allow 24bit / 96000hz sample rates through its SPDIF port? I know that DVD's have 24bit / 96khz sound, and that if I select "USE SPDIF" in a program such as PowerDVD, it will send the 5.1 signal to the receiver (the "Dolby Digital" or "DTS" on the receiver lights up and the speaker picture changes, so I know it is getting the correct signal.) I just hope that other SB-Live people like me can learn something from this thread and get better sound from their cards.

Edited cause I always seem to have one or two small errors in the origional..... headbang.gif

I'm pretty much in the same boat as you with regard to running a sblive and a surround amp (yamaha dsp-ax620) and i think you've got a little confused regarding some of the facts.
Sadly the SB live simply cannot handle 24bit or 96khz sampling. The S/PDIF will only ever output at 16bit with 48khz sampling. This is the main criticism of this card. As such, my settings for winamp are dither only on the ssrc output plugin, with it set to resample to 48khz 16 bit. I very much doubt that you will hear any quality increase (beyond placebo) by setting the input plugins anything higher than 16 bit. Which is why I don't bother.
Regarding dvds, DVD-audio uses 24bit with 96 khz sampling, DVD-video does not. It uses Dolby Digital or DTS, which are output to your amp by encoding them as a 16bit 48khz pcm signal that your amp will recognise as a dolby digital or dts stream. If you connected this to a pure pcm dac you would see it was getting a 16 bit 48khz pcm input, and you would hear horrible square wave's etc.
Go to the top of the page
+Quote Post
KikeG
post Feb 8 2003, 16:22
Post #8


WinABX developer


Group: Developer
Posts: 1578
Joined: 1-October 01
Member No.: 137



Yes, I'm sorry, I forgot you were talking about a Live card. Yes, the Live can only output 16 bit 48 KHz on its digital output, so SSRC with these setting is needed at the last step. Dither with noise shaping should be the better to use in SSRC, but flat dither would be ok too.

This post has been edited by KikeG: Feb 8 2003, 16:24
Go to the top of the page
+Quote Post
ViPER1313
post Feb 8 2003, 20:47
Post #9





Group: Members
Posts: 271
Joined: 19-August 02
From: Maryland
Member No.: 3109



QUOTE
Sadly the SB live simply cannot handle 24bit or 96khz sampling.


Thanks for the info. I'll just set the SSRC plugin to 48000hz / 16 bit on the output. Sigh, I hate the SB-Live. dry.gif (Although the down-sampling from 96khz/24bit to 48khz/16bit seems ok - look at pics 2/3 - there is little or no difference in the quality.)

QUOTE
Regarding dvds, DVD-audio uses 24bit with 96 khz sampling, DVD-video does not.


You learn something new everyday tongue.gif . Thanks for everyone's help.

Edit - For some reason the ' in everyones was a "?" - grrrrr unsure.gif headbang.gif

This post has been edited by ViPER1313: Feb 8 2003, 20:49
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 23rd October 2014 - 21:08