Spikey wrote on 2022-03-18, 15:06:

No, it wouldn't include JV or XV products, as they are not GS compatible and use different samples.

Are the Integra-7 and Sonic Cell GS compatible though?

I guess this is where I'm a bit confused as to the intent of the comparison, since modules like the Integra-7 and Sonic Cell are definitely pro hardware and not oriented towards gaming. Yet they can still be used for gaming given they have GM capabilities, same with earlier Roland sound modules.

In the end, it's all good though.

Shponglefan wrote on 2022-03-18, 15:30:

Are the Integra-7 and Sonic Cell GS compatible though? […]
Show full quote

Spikey wrote on 2022-03-18, 15:06:

No, it wouldn't include JV or XV products, as they are not GS compatible and use different samples.

Are the Integra-7 and Sonic Cell GS compatible though?

I guess this is where I'm a bit confused as to the intent of the comparison, since modules like the Integra-7 and Sonic Cell are definitely pro hardware and not oriented towards gaming. Yet they can still be used for gaming given they have GM capabilities, same with earlier Roland sound modules.

In the end, it's all good though.

Having just looked through the SonicCell's patch list, it isn't, and I'm fairly sure it doesn't have the SC sample set. Therefore, I'm removing it from the comparison. I figured it had a lot in common with the SD's, but it appears to be more like a Fantom and therefore a different beast. Thanks for the comment.

The Roland INTEGRA-7 is obviously not a gaming module, and is not GS compatible, but it does a have a SC-7 style GM sample set, so I'll be including that one.
In terms of the SC-55 section, the comparison is for devices that have the SC-55 Wave ROM. The INTEGRA appears to unofficially, at least for GM tones.

Basically, the goal is to hear the same ROM's play back the samples with different hardware, ROM sizes (compression levels), audio hardware (DAC's, opamps). With the SC-88 and beyond it's a bit trickier because there's a LOT of weird offshoots like the MT-90 series mentioned earlier. Still, I'll try my best.

My personal goal, is if there's a weird Roland keyboard or sound module out there that is Sound Canvas based, and someone is curious to buy it but unsure, they can refer to this database and see what the specs are, what it's equivalent to sample set and audio hardware wise, and how it sounds versus the "regular" Sound Canvas in that horizontal product line. Or at least some of the information if there's no source for info, or a recording. Etc. Then, they can make a purchase decision, or request more information. But not relying on subjective opinions, "YouTube quality" audio comparisons or nonexistent/wrong info.
Hopefully all that makes sense.

Sonic Cell is GM2 and not GS-compliant. In the 2000's Roland abandoned SC/GS in favor of GM2 and HQ sampling for it's hardware and software offerings, to combat the growing threat of Yamaha XG (and others) and the steady decline of Roland's brand and marketshare, but continued to offer some GS hardware and the Roland VSCs through Edirol.

As far as "weird" GS hardware goes, you'd have to look at Roland's digital pianos/keyboards, which have GS-compliant hardware inside. Even some of the newest ones have GS-compliant MIDI to provide a basic sound module alongside the main digital piano function in a separate bank.

RetroGamer4Ever wrote on 2022-03-19, 13:02:

Sonic Cell is GM2 and not GS-compliant. In the 2000's Roland abandoned SC/GS in favor of GM2 and HQ sampling for it's hardware and software offerings, to combat the growing threat of Yamaha XG (and others) and the steady decline of Roland's brand and marketshare, but continued to offer some GS hardware and the Roland VSCs through Edirol.

Correct, which is why I mentioned in the post above yours that I removed it from the comparison.

However, along with Edirol's SD line, Roland has still released GS devices in the 2000s. See: MT-90U and the SD-50 as some examples of this.

RetroGamer4Ever wrote on 2022-03-19, 13:11:

As far as "weird" GS hardware goes, you'd have to look at Roland's digital pianos/keyboards, which have GS-compliant hardware inside. Even some of the newest ones have GS-compliant MIDI to provide a basic sound module alongside the main digital piano function in a separate bank.

I'm not familiar with any current pianos/keyboards, although there certainly could be some current production units with GS support.
Certainly until quite recently, modern arrangers were still using Sound Canvas maps, for example the Roland VA-7 touchscreen arranger. Kind of want a Sound Canvas with extra sounds and a touchscreen, BTW 😀

Even the new E-A7 appears to be an expanded SC-8820/50, although it is no longer GS compatible and doesn't have the MT-32 variations.

Apparently the current BK-5 is still GS compatible. There we go. 😀

Major update to the project document, many sections have been rewritten to reflect changes in methodology and updated information. Overall readability has been improved.
More updates regarding the devices themselves (SCCP section) and the actual MIDI files to be used will hopefully be made this week, as will a separate Excel spreadsheet with people and devices in the project.

Spikey wrote on 2022-03-22, 08:03:

Major update to the project document, many sections have been rewritten to reflect changes in methodology and updated information. Overall readability has been improved.
More updates regarding the devices themselves (SCCP section) and the actual MIDI files to be used will hopefully be made this week, as will a separate Excel spreadsheet with people and devices in the project.

Perusing the updated doc, it looks like the SC-88 isn't listed in the SCCP section.

Also had a question about recording setups. Do you have any preferences as to how these are being recorded (i.e. when it comes to selecting audio recording h/w and signal chain)? Do you have any preference for the audio h/w used?

Shponglefan wrote on 2022-03-23, 01:13:

Spikey wrote on 2022-03-22, 08:03:

Major update to the project document, many sections have been rewritten to reflect changes in methodology and updated information. Overall readability has been improved.
More updates regarding the devices themselves (SCCP section) and the actual MIDI files to be used will hopefully be made this week, as will a separate Excel spreadsheet with people and devices in the project.

Perusing the updated doc, it looks like the SC-88 isn't listed in the SCCP section.

Also had a question about recording setups. Do you have any preferences as to how these are being recorded (i.e. when it comes to selecting audio recording h/w and signal chain)? Do you have any preference for the audio h/w used?

Hey man! Yeah, I didn't get it fully sorted, as per my post you quoted. I was very busy with University assignments last week. Hopefully another update shortly.

For setups, I don't have control over people's setups, so it makes sense to go with the best options you have, without applying any processing to the recording (I'll do the final editing). 5 seconds of silence at the start and end preferred.

MIDI files coming ASAP.

If anyone has access to an SD-50, that would be a good piece of hardware to test, as one of the last Roland SC-type modules. I'm thinking about getting one, as it works with Windows 10/11 and is a good playback device for GS MIDI files and games.

I currently own one, so it's going to be included for sure. Your question is specifically one of the types I'm trying to answer with this project, nice one. 😀

Hoping to release the MIDI's this weekend.

Spikey wrote on 2022-04-22, 03:12:

Hoping to release the MIDI's this weekend.

How many MIDI files are you thinking?

Do you plan on including virtual synths? Like the Windows GS Softsynth, Edirol Hypercanvas, SaoundCanvas VA or Mac OS DLS Synth?

Official or compatible GS synths are totally fine, whether virtual or hardware. I just don't want others that aren't. That can be in the GM project. 😀 The Mac one in your list probably is the only one that falls into the latter case.

OK, the files are finally here.

The files chosen are all Roland files from the 90s, "All Stars", "Cool" and "Elise". The three were chosen because they display GS characteristics and a variety of tempos (All Stars), problem cases like Bass and Lead* (Cool), and solo piano for easier audio quality/sound comparison (Elise). They also all contain an embedded GS Reset at the start.
* Bass and Lead is an example of a synth instrument that sounds different from the 55 to later models due to different waveforms used. (If I've somehow got this wrong, and Bass and Lead is not one of them that uses different waveforms, let me know please! 😀 )

They can be found here:
https://drive.google.com/file/d/1c0Lo2RjKDF1C … iew?usp=sharing

Instructions:
1. Please record the MIDI files digitally from an audio interface/sound card at either 24 or 32-bits/96 kHz sample rate - or if not possible, let me know and we will work out a nearest-neighbour compromise - with no editing of the MIDI data, no changed settings on your module, no additional processing enabled on your audio interface or sound card, and no processing of the digital recording (I will be doing all of that).
2. Also, please include at least 5 seconds of recorded silence before the MIDI plays, and 5 after the MIDI stops playing (and audio/reverb/etc can no longer be heard).
3. Please ensure your Sound Canvas is set for maximum volume output, unless the recording clips, and then preferably limit it on your audio interface/sound card as necessary. Volume changes on the Sound Canvas itself is to be used as a *last resort* and only to the amount that stops the resulting recording clipping, as any reduction in volume reduces the dynamic range of the synth when recorded. If you have had to use any of these changes (i.e., any scenario other than full volume output of your Sound Canvas or full volume input on your audio interface), please include that information.
4. Finally, upload the unprocessed file as a lossless WAV/FLAC file (zipping/etc is advisable) to a file sharing host that has no ads or restrictions, such as Google Drive, Mega, or Mediafire, and send it to me via this forum, or via email at agillett@NOSPAMsierramusiccentral.com (delete the NOSPAM part).

Note: If there's a 4th file you want to record, or any additional files that were designed for Sound Canvas playback (designed for either SC-55, SC-88, SC-88Pro, SC-8820 or SC-8850, and variations), feel free to record them and I will use some under an "Extra" or similar section. (Of course, extras need to have a GS reset at the start, and you need to specify reverb/chorus settings used, what MIDI device recorded from, etc. I'll also need the MIDI file for inspection as well as the digital recordings recorded to specifications above.)

OK, hopefully we can finally get moving. Sorry for the delays everyone, I was sick and then had exams, and then got sick again. Ah, fun times. 😉

NOTE: The project document will hopefully be updated this week.

I’ll probably just play the files directly off the floppy drive then. I’ll have to poke around and see what best I can use for recording this weekend.

I downloaded the MIDIs and tried giving them a quick listen via Media Player Classic. While ELISE.MID played fine, ALLSTARS.MID and COOL.MID wouldn't play. The player shows them as having zero length.

Never had this issue before with a MIDI file and media player classic.

Shponglefan wrote on 2022-04-23, 19:36:

I downloaded the MIDIs and tried giving them a quick listen via Media Player Classic. While ELISE.MID played fine, ALLSTARS.MID and COOL.MID wouldn't play. The player shows them as having zero length.

Never had this issue before with a MIDI file and media player classic.

If you look at Play->Filters in MPC-HC menu then you can see that it falsely identifies these files as video files. I do not know why since these are perfectly legit Midi files with perfect headers etc.
BTW, I re-exported them with FSMP that uses running status so while the information is perfectly preserved, the file structure and header are slightly modified (since the files became smaller). This way MPC-HC can play them:

Falcosoft wrote on 2022-04-23, 21:01:

If you look at Play->Filters in MPC-HC menu then you can see that it falsely identifies these files as video files. I do not know why since these are perfectly legit Midi files with perfect headers etc.
BTW, I re-exported them with FSMP that uses running status so while the information is perfectly preserved, the file structure and header are slightly modified (since the files became smaller). This way MPC-HC can play them:
ALLSTARS_COOL.zip

Thanks for doing that, the re-exported midis play property in MPC-HC.

I wonder why it was not recognizing them as MIDI files. I've played hundreds of MIDIs with it and never had that issue before.

I played the files on the old VSC VSTi (FSMP) and they sounded wonderful.