New testing image Bookworm Oram 64bits

Have you checked the subsnapshots are loaded too?

Thanks

I can see them, they are imported, but tagged as new. You can check it :stuck_out_tongue_winking_eye:

I have recreated all the subsnapshots, don’t worry!!

I do worry! We need auto conversion to work so need to understand the failure modes. Are you saying that the snapshot loaded but MIDI CC wasn’t changing the subsnaphots? Can you give an example of an action and expected result? I don’t know what parameters are expected to change in your snapshot.

Thanks Brian. I’ve gone ahead and recreated the subsnapshots so tha I can test Oram on V4 in a rehearsal and check it in my real use situation.

Engines are loaded and the latest status of each chain is also kept correctly, but none of the subsnapshots called with Midi Program Change instructions are called. The funny thing is that when you go to the subsnapshots menu you can see all the Midi PCs used in the snapshot after the user area which is empty, but all of them are present taggged as new.

I can attach screen captures but you can check it youir self if you load it.

I have 38 subsnapshots startig from 1 :slight_smile:

Best
Pau

PS: I have attached the snapshot in the first post so that you can check it.

I can’t get wifi working either.

Using the zynthian menus I can type the password but shifted characters are missing.

Using the webconf menu I can see all the networks but I am not offered a menu to enter the password, I just can see the available networks.


Pau

What characters are missing, Pau?
Have you checked the extra character screens when pushing the “alt” key in the virtual keyboard?

Regards

For instance “@”.

But I see very strange chars with alt which you would seldom use :slight_smile:

But @ is there. I added some weeks ago. It’s in the fourth screen, alt+shift. Top row, char #3 from left.
Perhaps it should be moved to a better place.

They are used in other countries. People like to use strange chars for wifi passwords.

Anyway, I would add any missing char if you tell me.

1 Like

It may be sensible to add internationalisation. People probably use characters from their locale.

1 Like

I’ve been testing oram in a rehearsal and I have good news as everything worked fine, so I’ll be using it for real performances.

Let me summarise the problems I’ve found:
-cannot configure WiFi using webconf: you are not offered the chance to enter a password
-missing characters to enter password in V4 (@jofemodo has offered to add the missing chars)
-subsnapshots are not rightly imported, so the migration is not strightforward
-sfizz seem to have some issues with soundonts and xruns are common if you play fast. I’ve installed oram in a Sandisk extreme because Sandisk ultra was unbearable.
-linuxsampler enters and endless loop with SCC Expressive strings (it has always been like this). When you play certain soundfonts linuxsampler tends to get stuck and you cannot use the chain anymore.
-VNC is not usable with a web client, only VNC client is available now :slight_smile:

For the rest is a huge step forward, kudos for the hard work!

Best
Pau

2 Likes

We probably will remove webconf WIFI configuration. It’s quite redundant and maintaining duplicated admin interfaces takes effort and time that we could use in a better way.

I’ve explained where to find the “@” above. Please, tell me what other characters did you miss.

This will be investigated :wink: We want full compatibility.

Do you mean that sfizz works worst in Oram than stable? Strange because most engines run faster/better because the 64 bits change. So it should be investigated.

I will check this soundfont and linuxsampler. Could you tell me other soundfonts that are “breaking” linuxsampler.

Uhmmmm?? I can’t reproduce this on my zynthians. I will check tomorrow with a fresh burned image.

Cheers

1 Like

@jofemodo see issue #1052. It looks like noVNC is not installed or configured in the image and the install in the update script is not being called.

1 Like

Ups! It seems that i deleted the package by error when trying to free some space to make the image to fit in 16GB SSDs. It’s restored again. Simply update.

Regards

1 Like

Well, i’ve removed the “WIFI config” from webconf. As i explained, having duplicated interfaces is hard to maintain, specially non-trivial ones like this.

Regards,

1 Like

Hi @Pau!

I’ve been working with your snapshot, trying to fine-tune the import code. Please, update, try with the original snapshot. and send feedback. I hope it’s now much better :wink:

Thanks!

1 Like

Hello and thank you for your help.

System up to date and it still does not work, subsnapshots don’t seem oi be imported.

I attach two pics:
-the first one is the list of empty ZS3 after restoring the snapshot
-the second one correspond to the new snapshot I’ve created that includes the 32 subsnapshots that I have added again. I attach it as a reference.

004-marti~2024.zss (1.7 MB)

I’ll test anything you send :slight_smile:


You are right, it’s there!
But the font does not help much, it’s much smaller than the rest :slight_smile:
Sorry for the noise.

1 Like

It looks like the MIDI channel is not being correctly set. In the import, MIDI channel 1 (#0) is used but in your reconstructed snapshot you use MIDI channel 15 (#15).

The names of the ZS3s may also be confusing. In the imported snapshot, they are all, “New ZS3” where as in your recreated one they have the name: CHAN/PROG.

I agree that the OSK button font is odd and I could only just recognise the ‘@’ symbol. Also, it is rather well hidden on the 4th layer!

[Edit] The snapshot you provided does have ZS3 triggered from channel 1 (#0) so the translated snapshot looks correct. Indeed, if I send program change on channel 1, the snapshot responds. (I don’t know if it gives the desired effect - you can tell us.)

1 Like