Limitations with MSX-Music

Page 6/7
1 | 2 | 3 | 4 | 5 | | 7

By pser

Rookie (30)

pser's picture

28-10-2021, 19:52

Hi jltursan,
cannot believe the problem could be caps related as the VGM players do sound really terrific!
cheers!

By pser

Rookie (30)

pser's picture

28-10-2021, 19:59

As a spin-off of this project, I have prepared a version of the player that reads the same files and sends the bytes
right to a double 8 bit DAC (a cartridge) just to compare ... and the difference is noticiable. The volume changes
do not happen (obviously) on the DACs
regards
pere

By Grauw

Ascended (10179)

Grauw's picture

28-10-2021, 22:17

If you send me a problematic audio file in 15.7 kHz 8-bit unsigned raw format, I could record it playing with fmpcm on my MSX hardware for comparison, if that would help you.

By the way, is the OPLL you use an NMOS (YM2413) or CMOS (YM2413B) version?

By pser

Rookie (30)

pser's picture

28-10-2021, 23:22

Grauw wrote:

If you send me a problematic audio file in 15.7 kHz 8-bit unsigned raw format, I could record it playing with fmpcm on my MSX hardware for comparison, if that would help you.
By the way, is the OPLL you use an NMOS (YM2413) or CMOS (YM2413B) version?

The board we are using contains an YM2413 (9950 THINK appears below the number)
The files I am using are all of them just 11.025 kHz. I am creating them with Audacity v2.0.1
But it accepts to save an mp3 with 15.7kHz, so I will get the original mp3 of one of those wav files that do 'lose' volume
and will save it that way.
Is it possible to attach a file here? Sorry for the question, I am a complete newcomer here
regards

By Grauw

Ascended (10179)

Grauw's picture

28-10-2021, 23:54

You could also upsample the 11 kHz file using Audacity, it should play the same at a higher frequency. Basically for my convenience I’d just prefer a file that I can play straight away without needing to do additional conversion.

Attaching files is not possible, but you can find my email address in my profile. It should accept files of several MBs (zipped), and otherwise you can send it with wetransfer…

Since this person here also has a bunch of “9950 THINK” labeled YM2413 probably you can’t trust the label, identical production codes are a telltale sign of relabeling by Chinese sellers, so it could be both an YM2413 or an YM2413B.

By pser

Rookie (30)

pser's picture

29-10-2021, 00:18

Grauw wrote:

You could also upsample the 11 kHz file using Audacity, it should play the same at a higher frequency. Basically for my convenience I’d just prefer a file that I can play straight away without needing to do additional conversion.

OK, will send you a 17.5Khz WAV file of any high dynamic changes ...

Quote:

Attaching files is not possible, but you can find my email address in my profile. It should accept files of several MBs (zipped), and otherwise you can send it with wetransfer…

Will try to send it right to your mail

Quote:

Since this person here also has a bunch of “9950 THINK” labeled YM2413 probably you can’t trust the label, identical production codes are a telltale sign of relabeling by Chinese sellers, so it could be both an YM2413 or an YM2413B.

Might be. I will search among my ICs boxes. I am sure I had three different kind of these chips!
Thanks a lot
cheers!
pere

By pser

Rookie (30)

pser's picture

29-10-2021, 00:23

just a note.
I have found the other two kind of chips. ONly one of each but these are labelled that way:
- YM2413 on top and below 0620
- YM2413 on top and below 1025 HAAG
Would you trust any of these?
thanks

By pser

Rookie (30)

pser's picture

29-10-2021, 11:13

Good morning,
I have tested the other two YM2413 chips and one of them has given no output at all :-(
The other labelled "1025 HAAG" seems to show a lower dynamic change effect (a bit better).
Anyway I cannot rely on my current files because they have been severely reworked on Audacity
to reduce drastically their volume changes ...
But keeping into account that right now I could be sending them to the double 8bit DAC, I think it is time
to convert the .mp3 files again and just try to level them to avoid songs at high volume and others with too low
volume.
Probably somewhere, it exists an application that can process a batch of files to level their volume reducing differences,
but I am not aware of any. Does anybody know of something to do that (on Windows 10)?
Thanks in advance
pere

By pser

Rookie (30)

pser's picture

29-10-2021, 21:57

Hello,
jltursan recommended the use of SoX to process multiple files when converting from mp3 to wav
Well, maybe because this tool adds dithering o maybe because I have down converted to 14KHz 8 bits
instead of just 11.025KHz but the result are files that are playing quite well with the first YM2413 I was using
A bit longer but space is not a problem on a card reader!
Not sure, but maybe sending more bytes per second helps the level in certain passages not to fall down ...
I will try to tweek even more the player for maybe I could work with 16K or a bit more sampling frequency :-)
Thanks a lot to all of you for your advices/help!
regards
pere

By Grauw

Ascended (10179)

Grauw's picture

29-10-2021, 23:54

pser wrote:

I have found the other two kind of chips. ONly one of each but these are labelled that way:
- YM2413 on top and below 0620
- YM2413 on top and below 1025 HAAG
Would you trust any of these?

Note that it’s only the label I don’t trust, I believe that the chips themselves are fine legit original Yamaha chips. You just can’t tell from the label whether they are an YM2413 or YM2413B. (Btw yes HAAG is another mentioned in that post I linked so I suspect also relabeled.)

I was asking about the NMOS vs CMOS thing because the NMOS version of the YM2413 does have slightly non-linear output characteristics, as mentioned on Wikipedia and also found in Wouter’s research. But this does not really explain the phenomenon you describe though, so forget I mentioned it.

pser wrote:

Well, maybe because this tool adds dithering o maybe because I have down converted to 14KHz 8 bits
instead of just 11.025KHz but the result are files that are playing quite well with the first YM2413 I was using
Not sure, but maybe sending more bytes per second helps the level in certain passages not to fall down

Strange, it should hold the output level until you send a new value, so I can’t come up with an explanation for why playing back at a slightly higher frequency would make a difference…

Page 6/7
1 | 2 | 3 | 4 | 5 | | 7