Hi, I want to get Linux and a hardware sampler talking to each other more effectively. Specifically the following: 1. Transfer individual samples between the PC and sampler 2. Access and back up project files saved by the sampler 3. Possibly create a hard-disk partition in the PC which the sampler sees as a SCSI drive The sampler is an Emu ESI-2000 and the interfacing would probably have to be via SCSI because that's all it offers. For the 1st aim, I did find out that the sampler supports SMDI, a protocol for transferring samples over SCSI supported by various sampler manufacturers including Emu and Yamaha. Is there any Linux end-user software that implements this protocol? Or another way of doing it? The only Linux software I have found so far that supports SMDI is OpenSMDI, which implements the SMDI protocol as a free shared library for Linux and Windows. From what I can see there is no front-end on the Linux version: http://home.t-online.de/home/chrisnowak/opensmdi/ If there is no alternative, I wonder how hard it would be to create a simple command-line interface using this? For the 2nd aim I want to read ESI's file system to access individual projects or 'banks'. Currently, all I've managed to do is save banks to Zip disks, and then use the Unix 'dd' command to make images of entire Zip disks. But I can't mount the Zip disks (or disk image files) from Linux, or access individual files or 'banks' within them, because the sampler uses a propietary file system. I believe aims 1 and 2 are both necessary because the SMDI protocol is fairly limited in the sample parameters it supports, so in order to capture all parameters (such as filter settings) and cleanly save whole projects in one go, you have to use the sampler's 'Save Bank' facility rather than transferring individual samples. So is there any Linux software that can read this proprietary file system? The only free (as in beer) software I have found is a Windows program called ESi-Win. Apparently it lets you transfer individual samples over SCSI to and from ESI samplers as well. http://www.simplydata.co.uk/ESi-Win/ This software actually solves both aims 1 and 2, but it's Windows-only. I briefly tried to run it under Wine, but it didn't work straight away. Has anyone got it working under Wine? Has anyone seen it running at all? Both OpenSMDI and ESi-Win seem to be projects that progressed to a reasonably functional alpha or beta level but are no longer maintained - presumably because their authors sold their samplers. I'm beginning to reluctantly wonder if I should do the same... although it seems a shame to do so purely because of issues with file formats etc. Anyway the final problem arises because, even if I solved all the above, I am still reliant on the Zip drive to save and restore banks directly from and to the sampler. This is not good from the point of view of reliablility and cost, especially as I undertand Iomega don't make the SCSI version of the Zip drive any more. To solve this without buying an expensive SCSI hard drive, I wondered if it is possible to create a partition on an IDE drive in the PC that the sampler sees as an external SCSI drive? (Preferably with no possibility of the sampler overwriting other partitions on the drive ;-) The sampler has an option "Ignore SCSI device with ID x", which according to its manual is intended to allow both a PC and the sampler to be masters on the same SCSI chain and share a SCSI drive. But in the setup they describe, the shared drive is a separate box in the middle - can it be done without this? >From what I understand about SCSI, doing this would at least require a SCSI card that has the capability to be 'SCSI target' to another master device, as well as being a master device itself. So, for what it's worth, my SCSI card is an Adaptec 2906 and I'm not sure if it has that capability or not. What about other options? I could get rid of the sampler altogether - then to get the same number of outputs I'd need to upgrade my soundcard - but anyway, what Linux software is there that could replace a hardware sampler while probably retaining a sampler-like way of working? Finally, I suppose I could upgrade the hardware sampler. I guess that more modern hardware samplers integrate better with PCs in general. But are their protocols and file formats any more open and therefore potentially more inter-operable with Linux, or are you just as locked in by proprietary formats and protocols as before? Any thoughts, suggestions, ideas welcome. I can't be the only person who's wants to do this! Thanks James -- James Greenwood | jamesg@xxxxxxxxxxxxxx If you put in the work, the results will look after themselves -- Nick Leeson's mother