Improving 1chipMSX sound quality.
Since the DAC on the 1cM is a simple resistor array, The current version of the code OCM:
1) A mono signal which is fed to both exit;
2) DAC output is not used, the signal is formed 13 - bit PWM module (esepwm.vhd).
1) A mono signal which is fed to both exit;
2) DAC output is not used, the signal is formed 13 - bit PWM module (esepwm.vhd).
13 bit Pulse-width modulation for the audio ?
What you mean exactly ?
You get audio analog signals from PWM by low pass filtering,
is this what OCM does ?
In this case why you say 13 bit ?
What you mean exactly ?
You get audio analog signals from PWM by low pass filtering,
is this what OCM does ?
In this case why you say 13 bit ?
My 1chipMSX[KAI] has improved tone quality.
However, it is not the same as the real thing 100%.
First of all, please test it.
http://www5d.biglobe.ne.jp/~hra/note/onechipmsx/files/emsx_top_20080408.zip
* About DAC of 1chipMSX
The output port has 6bit.
However, only "000000" or "111111" is output.
* About SCC on my 1chipMSX[KAI]
A wavy output is enhanced to 15bit.
* About ESEPWM
ESEPWM is not Pulse Wave Modulator.
It is delta sigma transfom that converts 14bit to 1bit.
* New improvement idea
When the delta sigma conversion is changed to 14bit to 6bit, it is likely to be
improved.
I am busy with addition of the V9958 interchangeability function now
However, it is not the same as the real thing 100%.
First of all, please test it.
http://www5d.biglobe.ne.jp/~hra/note/onechipmsx/files/emsx_top_20080408.zip
* About DAC of 1chipMSX
The output port has 6bit.
However, only "000000" or "111111" is output.
* About SCC on my 1chipMSX[KAI]
A wavy output is enhanced to 15bit.
* About ESEPWM
ESEPWM is not Pulse Wave Modulator.
It is delta sigma transfom that converts 14bit to 1bit.
* New improvement idea
When the delta sigma conversion is changed to 14bit to 6bit, it is likely to be
improved.
I am busy with addition of the V9958 interchangeability function now
Hmmm, in that case I guess there's little to be gained by using all 6 bits then. Unless it's in size. It does make me wonder why the standard SCC sound was so crappy then.
If only there was enough time to investigate
If only there was enough time to investigate
maybe only HARA's onechip has better performance
maybe only HARA's onechip has better performanceIn OCM ported at Altera DE1 for output of sound I used Audio codec in mode 16-bit, 48 KHz Sampling Frequency.
Audio quality can be compared with the output ESEPWM, which is also used in DE1.
Audio quality can be compared with the output ESEPWM, which is also used in DE1.
It is 21MHz output though the output of OCM is 1bit. (1bit DAC)
Because it is one bit output, the quantization noise is recorded.
However, only the noise is modulated to the high frequency region by the delta
sigma modulation.
Noises gather near Nyquist frequency (11.5MHz).
The band in the vicinity of the frequency of the nyquist is cut with the capacitor
connected outside of FPGA.
The sound circuit of OCM is made expecting such a mechanism.
It is necessary to use the input signal of ESEPWM to reproduce with 16bit/48kHz.
However, do not connect directly with 16bit/48kHz Audio Codec.
Please connect the signal that passes the low-pass filter by cutting the
frequency that exceeds 24kHz.
Please pay attention the alias of the signal.
Because it is one bit output, the quantization noise is recorded.
However, only the noise is modulated to the high frequency region by the delta
sigma modulation.
Noises gather near Nyquist frequency (11.5MHz).
The band in the vicinity of the frequency of the nyquist is cut with the capacitor
connected outside of FPGA.
The sound circuit of OCM is made expecting such a mechanism.
It is necessary to use the input signal of ESEPWM to reproduce with 16bit/48kHz.
However, do not connect directly with 16bit/48kHz Audio Codec.
Please connect the signal that passes the low-pass filter by cutting the
frequency that exceeds 24kHz.
Please pay attention the alias of the signal.
Don't connect the voice mixer output of OCM with AudioCodec direct.
The modes of expression of the sign differ.
Fixed version:
http://www5d.biglobe.ne.jp/~hra/note/onechipmsx/files/emsx_top_dev_20080527.zip
http://www.msx.org/forumtopicl8494.html
The modes of expression of the sign differ.
Fixed version:
http://www5d.biglobe.ne.jp/~hra/note/onechipmsx/files/emsx_top_dev_20080527.zip
http://www.msx.org/forumtopicl8494.html

By Edwin
Paladin (826)
09-04-2008, 22:56