Need your opinion regarding GR8NET capabilities

Page 7/55
1 | 2 | 3 | 4 | 5 | 6 | | 8 | 9 | 10 | 11 | 12

By edoz

Prophet (2169)

edoz's picture

10-09-2015, 18:16

Wow! What a good manual! I wonder if it would be possible to use the cartridge only for network in a slot expander ? And does it support msx computers that runs the slot on 7 mhz ?

By hit9918

Prophet (2866)

hit9918's picture

10-09-2015, 18:31

Adding a low pass filter? That means low sounds get louder, right?
But the example case grauw showed, the original had more HIGH frequencies.

Now, if that came in spite of a low pass filter, it is the SCCs tendency to make high pitched things.
Like a glitch of sample OR previous sample.
I recommend to investigate that issue whether that OR does add the original feel.

By hit9918

Prophet (2866)

hit9918's picture

10-09-2015, 18:34

Quote:

Actually great thanks to Konami for making good samples!

They made samples that work around the OR issue! Things would best be checked with non konami samples.
How does an original SCC sound when playing a wav file?

By Eugeny_Brychkov

Paragon (1081)

Eugeny_Brychkov's picture

10-09-2015, 18:36

edoz wrote:

in a slot expander?

Yes.

edoz wrote:

And does it support msx computers that runs the slot on 7 mhz ?

This is good question. To my understanding external slots anyway run 3.58 even if CPU/machine is running 7.

By hit9918

Prophet (2866)

hit9918's picture

10-09-2015, 18:50

@eugeny, is clear which issue I mean? I found one hit about it from 2008:
http://www.msx.org/forum/development/msx-development/scc-sou...

Quote:

If you check the sampled output, you see spikes in the waveform

And there was a thread with pics from osciloscope, that was before the forum remake, the urls no more work.

By Grauw

Ascended (8318)

Grauw's picture

10-09-2015, 19:40

hit9918 wrote:

Adding a low pass filter? That means low sounds get louder, right?

High frequencies get attenuated, rather. Because the human ear can't hear more than 20 kHz, the low-pass filter maxis mentioned improves audio quality by filtering out pitches outside the audible range which conventional audio equipment can not reproduce well without causing audible distortion. It does not filter out audible pitches like the higher harmonic frequencies in SCC sounds.

hit9918 wrote:

Now, if that came in spite of a low pass filter, it is the SCCs tendency to make high pitched things.
Like a glitch of sample OR previous sample.
I recommend to investigate that issue whether that OR does add the original feel.

Those higher harmonic frequencies in SCC sounds are not occurring by some strange analog or internal glitch phenomenon, rather they are a very logical result of the stepping in SCC waveforms due to the limited length of 32 samples. This is why interpolation, which removes the stepping, eliminates those frequencies. (I can elaborate if you want.)

That "OR effect" you mention is not well-described and I'm pretty sure none of the FPGA SCCs like the MegaFlashROM and 1chipMSX implement it, while they sound fine. (Well... haven't done a thorough comparison, but never heard complaints.) So I don't think it is needed to go into this much detail emulating the SCC to get a good result.

Eugeny_Brychkov wrote:
edoz wrote:

And does it support msx computers that runs the slot on 7 mhz ?

This is good question. To my understanding external slots anyway run 3.58 even if CPU/machine is running 7.

They should, but they don't.

They don't because decoupling the external bus would be next to impossible in a 7 MHz kit.

Official MSX machines with turbo all do have a 3.58 MHz external bus though, to my knowledge.
(Panasonic MSX2+ and turboR computers.)

By Eugeny_Brychkov

Paragon (1081)

Eugeny_Brychkov's picture

10-09-2015, 19:47

Grauw wrote:

Official MSX machines with turbo all do have a 3.58 MHz external bus though, to my knowledge.
(Panasonic MSX2+ and turboR computers.)

Thanks, this is what I assumed. It is not current priority to make the stuff working @ overclocked machines. My focus is on reliability and usefulness.

By Eugeny_Brychkov

Paragon (1081)

Eugeny_Brychkov's picture

20-09-2015, 01:44

Hello, I need your opinion please.
If you read new version of manual, you see that now you can have up to 4 GR8NET adapters installed in the system. In single adapter version I named BASIC devices as TCP0 ot TCP1, UDP0 or UDP1 (adapter has 2 user sockets). Now when there could be 4 adapters installed, what in your opinion is the best naming convention?
I came to to the idea of renaming sockets to A and B, and numbering adapters 0-3. For example, TCP2B will means adapter's #2 socket 1. TCPA will be adapter 0's socket 0. While TCP0 will be invalid because it is missing socket identifier.
Any better ideas?

By Prodatron

Paragon (1788)

Prodatron's picture

20-09-2015, 01:57

Hi Eugeny, yes, I already read this and was quite impressed. It's very good that this is scaleable but I wonder what is the advantage to have multiple GR8NET cards connected at the same time. Currently I wonder if it's possible to write a flexible (SymbOS) driver which supports more than 4 open sockets. Anyway I still want to have 2 or more of your very cool cards! Tongue

By Eugeny_Brychkov

Paragon (1081)

Eugeny_Brychkov's picture

20-09-2015, 08:11

@Prodatron: thank you. I need to ensure it is possible to extend single adapter configuration, and that it works.

Alternatively, I can name adapters A, B, C and D and leave sockets 0 and 1 then TCPB2 (or TCP2B) will be socket #2 of adapter #2, and TCP0 will be valid device (socket 0 of adapter 0) while just TCPC will be invalid device name.

Page 7/55
1 | 2 | 3 | 4 | 5 | 6 | | 8 | 9 | 10 | 11 | 12