News:

2box forum: accident-free since the last one.

Main Menu

DSoundTool 1.11 released

Started by Louis, June 14, 2011, 09:34:30 PM

Previous topic - Next topic

Jman

I did run into similar problems with cymbal tails getting chopped in the new version. And some other strange stuff ... with the new MIDI sequences I was getting random no. of hits .... iE ... make a 33 hit recording and end up with 31 hits or 32 hits .... make a 99 hit recording and end up with anything from 94 to 99 hits ... I thought it might be a sync problem, but I tried synching in the player ... and it didn't change anything ... when I used the earlier MIDI sequences that didn't happen ... not sure what's up on that ....
So, anyway, just reverted to the previous MIDI sequences and splitting out the Wavs to individuals works just fine for me ....
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Louis

Hi Jman,

I did some experiments, and I can confirm that I made a typo in one of the parameters that is involved in splitting of wave files. The result is that samples with very low volume get skipped, so I would suspect that the samples that had lowest volume were skipped.

I was finally also able to reproduce the tail-cut example. I got this when the overall volume in the recorded wave file was very low. I could improve it quite a lot by a few simple adjustments, but not completely get rid of it without making more radical changes. I will test a bit more and release a new version soon.

But in general I would recommend to make sure to record with as high volume as possible (without causing "red peaks"). Also I would recommend to record at 24 bits; this minimizes rounding errors when DSoundTool scales up the volumes and should give a little bit better sound quality at low volumes. The latter also applies if you use "many wave files in 1 directory" method.

Slap the drummer

Hi Louis - I know its not a major point but cd I clarify the 24bit/16bit issue.....

DSoundTool converts from 24bit to 16bit last of all I assume.  And it's best to feed it 24bit so that the volume
adjustments are made at 24bit.... right so far?

So if we don't use the "Max Volume" feature it wdn't make any difference if we feed it 16bit, or does the DSoundTool
still do some "processing" of the samples?

Also, it crosses my mind - if we make volume adjustments AFTER the initial creation of the .dsnd file we might lose
technically lose a small amount of fidelity because the adjustments are now being made to a 16bit file ?


I must say now I'm turning out the dsnd files I really appreciate the way the tool has been designed (with the max
volume features and so on).  The most time is spent bouncing the files in SD.  Creating .dsnd files can literally
take seconds !  Great stuff.

Thanks

Louis

Quote from: Slap the drummer on June 21, 2011, 11:44:41 AM
Hi Louis - I know its not a major point but cd I clarify the 24bit/16bit issue.....

DSoundTool converts from 24bit to 16bit last of all I assume.  And it's best to feed it 24bit so that the volume
adjustments are made at 24bit.... right so far?

So if we don't use the "Max Volume" feature it wdn't make any difference if we feed it 16bit, or does the DSoundTool
still do some "processing" of the samples?

Also, it crosses my mind - if we make volume adjustments AFTER the initial creation of the .dsnd file we might lose
technically lose a small amount of fidelity because the adjustments are now being made to a 16bit file ?


I must say now I'm turning out the dsnd files I really appreciate the way the tool has been designed (with the max
volume features and so on).  The most time is spent bouncing the files in SD.  Creating .dsnd files can literally
take seconds !  Great stuff.

Thanks

Yes, you are right on all points, spot on! So indeed, when you don't use the max volume feature, there is no difference in 16/24 bits; only when you scale volumes. And thanks (again!) for your compliments :D

Slap the drummer

On the subject of red peaks and starting out with as high volume as possible, I've noticed that in SD2.0 that
when you bounce a drum out to a 2 channel WAV, if the signal clips then SD redoes the mixer stage and then
puts out a WAV which peaks back under 0db.

Which is a pretty cool feature.  So I guess we can ramp up the mic volumes without having to worry too much
about clipping......  ?

Jman

Quote from: Slap the drummer on June 22, 2011, 08:17:59 AM
On the subject of red peaks and starting out with as high volume as possible, I've noticed that in SD2.0 that
when you bounce a drum out to a 2 channel WAV, if the signal clips then SD redoes the mixer stage and then
puts out a WAV which peaks back under 0db.

Which is a pretty cool feature.  So I guess we can ramp up the mic volumes without having to worry too much
about clipping......  ?

I didn't know that .... A lot of producer presets in Sd2.0 peg the Red a lot .... so that explains it ....
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Slap the drummer

Yeah, it seems to calculate an output level (to right on bounce dialog box) and then redo if the level
goes over.

I keep meaning to check that it does just bring the master level down........

logihack

well, i did some good samples yesterday...now i know that the problem with tails of the sound must be in logic, because everything works fine in toontrack solo...but the progress is so slow - much slower than working with logic, where you dont have to record every sample

Jman

Anyone had a (Java) Negative Array Size Exception message that prevents you from completing a dsnd? That's what I got today....
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Louis

Yes, it is a bug that I also have found and fixed. I had planned to release a new version last week. Unfortunately this did not happen; I spent the last 10 days at the hospital because my 12-year old son was diagnosed with diabetes. My son is ok and we came back home today.

I must say that we got really great support, the health care in Sweden is really fantastic for children.

You understand that I was not in the mood for any programming... I am sorry for any inconvenience that this bug causes. If I remember well (but my memory is dim) the bug occurs if the wav file that is to be converted is bigger than a certain size; you could try to split the file using soundforge before converting to dsnd.

If everything goes fine, perhaps I will release next week, there is only some tweaking and testing left, but I do not want to make any promise at this point.

Jman

Quote from: Louis on July 08, 2011, 10:55:04 PM
Yes, it is a bug that I also have found and fixed. I had planned to release a new version last week. Unfortunately this did not happen; I spent the last 10 days at the hospital because my 12-year old son was diagnosed with diabetes. My son is ok and we came back home today.

I must say that we got really great support, the health care in Sweden is really fantastic for children.

You understand that I was not in the mood for any programming... I am sorry for any inconvenience that this bug causes. If I remember well (but my memory is dim) the bug occurs if the wav file that is to be converted is bigger than a certain size; you could try to split the file using soundforge before converting to dsnd.

If everything goes fine, perhaps I will release next week, there is only some tweaking and testing left, but I do not want to make any promise at this point.

Sorry you have had such a shake up in your life ... it really brings you back to what is most important in life when something like that happens ... best to your son and you and your family Louis...

I did go ahead and split out the waves in SF the other day and it worked just fine ... so it evidently is a bug only when large single Wav file is used..... take care of the fam .... no worries, no rush ... best wishes, J
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Louis

At last DSoundTool 1.12 is released. This is a bug fix release:

    * Fixed an error when creating a dsnd file from a large wave file.
    * Further improved silence detection when creating dsnd files from a single wave file.

The silence detection is not 100% perfect: a little bit "tail cutting" of certain long-lasting cymbal sounds can still occur if the samples are recorded at very low volumes. But for most sounds (also most cymbal sounds) I think/hope that this version will be good enough.

Any feedback would be appreciated!

Jman

Cool, Thanks Louis, give er a try next dsnd's I make ..... and again really appreciate what you've done with this. If it weren't for your program I wouldn't have these killer new drum sounds in my module!  :patbat2box:
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Jman

#28
I haven't done any new cymbals yet, but did some toms and it worked great using the "all samples are contained in one .wav file" ... all tails were in tact .... Plus with your newest MIDIs I am getting a wide dynamic range. I think you have this updated tool working quite well! Thanks for all your work Louis! J
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/

Jman

I just generated a couple new MIDI sequences from version 1.12 and the velocities go from loudest to softest whereas before it was softest to loudest.... is that on purpose?
I could tell you where to stick that piezo! :D ;)
http://stealthdrums.com/