Possible OS v1.70 sample load bug! Check please!

Hi guys,

we're having an interesting problem on the Akai mail list with some of the Hollow Sun sounds. After loading, a number of samples just don't play. I found that the problem is a very high sample START value, exactly 256 samples below the END value. So the sample range is very short, and you hear nothing when triggered.

Steve from Hollow Sun tested his sounds with a guy running v1.60 and he had no problems. I'm running a beta v1.70 and have the problems. So has a guy running the official v.170.

As a test, could you please load the samples from the SCI Tom drum machine from Hollow Sun? If you don't have them on your HD already, they're here: http://www.hollowsun.com/vintage/tom/index.html

None of these samples load correctly on my machine. Simply check the sample range in the TRIM/LOOP page. Watch the value of the St: field. It's very high.

I suspect that v1.70 users have this problem, while users running an older OS don't. I don't believe this problem has any relation with Hollow Sun, btw. I just use their samples to test this bug.

hy rob

i m on the official new os1.70 and i just checked them ...

on every sample the sample start is nearly at the end ... ie. tom kick the startpoint is at 2,674 - end is 2,930 ... its similar on every other sample from that folder ... weird! ... i haven't done alot on os 1.70 yet - but on the samples i been workin the past days i didn't encounter that problem ... hopefully sequential tom is just an exeption ...

You'll find that the START value is exactly 256 samples below the END value in the 'broken' samples.

And nope, it's NOT an exception. I know of at least 4 sounds that have this problem. My guess is that it's a serious sample load bug in v1.70, and that it has no relation at all with Hollow Sun.

Serious sh*t.

... thanx rob for the bad newxs ... anyways ... fukk - this really sounds bad!

i haven't done alot on 1.70 yet but dunno - somehow i don't feel it really good after the few times i used it ... what about this > imo screen movements esp. in program mode used to move quicker in the past ... yesterday i had less then 10 drumsamples in my memory - guess thats less then 5 % of ram usage ... and still when workin in programs 4k was very slowly ... is it only my 4k? or has it always been like this? ... dunno ... hopefully i just smoke too much and start imagining things here ...

I haven't noticed any slowness yet. But then again, I'm a new user really.

Maybe it also depends on the amount of MIDI events you're playing? Lots of CCs for instance?

hy rob ... thanx for your answer ... new user? i remember you postin here like ... 1 year ago?! ... it that possible? how long you had to wait for your 4k?

regardin midievents and cc's > nah nada ... yesterday i just had a small session where i did most at my module - 4k has just been used for some drums (with less then 10 samples - all very short samples - only bd, sn, toms and hhs) and triggered 3 lines at my module, no cc's recorded - so i just used some note messages to trigger my module, - 1 sequence - 8 bars ...

last times i often had lotsa ram used - then ok i understand somehow why its slower ... but with such little goin on i remember my 4k bein faster ... could be somethin within my 4k - dunno ram or whateva - however i just want to check - if someone else notices the same i guess its os related ... if not i guess its only my 4k ...

One year ago? Yes, that's possible. That's when I decided to get the 4k, but the $$$$ wasn't there yet. Started reading the manual though. Again.
I have my blue machine for about 2 months now? Not that long.

Regarding the slowness. If I remember correctly, every new OS seems to have a different GUI speed. Sometimes its faster, sometimes slower. I guess it depends on how the load on the main CPU is divided over the different services running in the 4k. And when these are updated/changed/tweaked, we notice the difference in the speed of the GUI updates. Hard to say what's going on there.

I'd like the source code please.

hy rob ... thanx for introducin in the gui stuff ... not that i can make much sense out of it - but sitll ...

congrats for gettin through the manuals without 4k - 2 month ago imo was quite a good time to get a 4k - actually imo since 1.51 everyday is a good day to get a 4k - i got mine on os 1.30 or 1.33 which was maybe a bit too early - i m happy with it since 1.51 which is basically a year from now - but i ll never forget last christmas eve when i uploaded os 1.50 which i still consider the worst christmas present i ever got ... anyways enjoy it and bangin beaz ...

Okay guys, Steve from HollowSun found that the problem resides in both his edit tools and the v1.70 OS. Here's the whole blurb taken from the Akai maillist. (He's replying to an email I sent him after checking a new version of one of his sounds). Here goes:


> the new Melodica loads just fine!
> All samples have a good range, program is fine, all keys
> play here.
>
> So you found a way to fool the MPC4k v1.7 OS?
> I'd LOVE to know the details!
Ok....

I have recently discovered a slight problem here with the wave editor I use
(a very old version of Bias PEAK for OS9).

I typically edit the sound via ak.Sys... the thing is transferred to the
wave editor and I zoom in and trim the sample start as tight as possible to
the attack transient. However, there is also Akai's own separate 'START'
marker and if *my* sample start is after the Akai START marker and I discard the redundant data before *my* sample start, the wave editor sets the Akai START marker to one sample before *my* sample start (still with me ?

No problem... plays ok in the wave editor. So I save and the thing gets
lobbed back at the 5000 where it also plays fine. So I save, zip and upload
to the site.

On closer examination on the 5000, however, samples trimmed this way set the Akai sample START to -1 (which, of course, is a nonsense!). But this has never been a problem before because the 5/6000 can cope with this and the Z4/8 and MPC4000 'trapped' sample starts to be at 0.

However, because of some issues on the MPC4000 where the sample start
couldn't be set after the loop start (even if the loop wasn't active), this
trap was removed in V1.70 and it is a by-product of *this* that has given
rise to the recent spate of problems reported. It seems that my '-1 samples' are placing the sample start just before the sample *end* when loaded in MPC4k V1.70 (as you reported). You won't have the problem in MPC4k V1.60 or earlier - neither will there be a problem in S5/6000 or Z4/8 (and probably not in s/w samplers either as these typically ignore Akai's sample START marker).

> Btw, just an hour ago I loaded your CP70 sound (not the
> cdrom, but the archive sound), and it had the same problem.
It'll be the same problem.

> I guess tons of your (and other) sounds have this.
Possibly - not all the stuff was edited and trimmed this way.

> Are you planning to update all your sounds?
It would be a fairly large undertaking to go through each and every sample
that's available but I will attend to any reported problems wherever
possible.

> Or can we expect an OS update?
The problem has been fixed in the MPC4k OS but it's not for me to say when this will be released as I have no idea! Once this new OS update is
released, however, any of my '-1 samples' won't be a problem as the sample start will be set to 0.

Hope that makes sense! Thanks for checking.


Best regards,


Steve
http://www.hollowsun.com