Hello, i am following Deejaysystem since his beginning and i am also mixing with it (at home) also since the beginning.

So congrats for the latest beta, there is a MUCH BETTER SOUND QUALITY and by the same way also a better precision, you are the only having these both features.

now there is still a very little lost of quality and i was wondering if you needed someone to beta test the quality ?

thx

Sure, what do you mean exactly with "lost of quality"? i.e. just sounds different, other products have better quality, etc.

Greets!

I mean you have a very little lost of quality when you change the pitch, so you have a perfect quality a Pitch 0 but when you change, there is a VERY VERY little lost but that you can better hear on the breaks of the song.

i will try to make you a little sample if i found time for it

Edy wrote: Sure, what do you mean exactly with "lost of quality"? i.e. just sounds different, other products have better quality, etc. Greets!

i just found to show you with a sample what i mean with this little lost of quality.

it is based on a rip from Boss Nova song, Stone Cold

there you can download it, very little file 300 ko => http://www.geocities.com/retergu/Sample.mp3 - http://www.geocities.com/retergu/Sample.mp3

you here about 5 sec that the speed of the song is faster, 1.00 pitch faster and then you hear a little lost of quality.

so if you need some help to for fixing this problem, i will be glad to help you a bit

i have also another question : you probably also observed that on windows 95/98 /98SE you have a stable beat with the "exact pitch control" option that is on off what not at all the case if of Windows XP.

what do you think make the difference between both os ?

i think you are the only who can answer me on this question

Ok, let's go for it.

About the sample, do you mean that the sample itself is played at 1.00 faster after second 0:05, or should I play it at Djs in order to hear the difference?

I've noticed a bit of distortion, but I think that it's caused by signal clipping instead of pitch changes.

Exact Pitch Control: this option enables internal pitch routines, so the player doesn't depends on soundcard, drivers, etc. to perform the pitch. It should be exactly the same in all OSes.

When this option is enabled, the output is played always at 44,1 Khz. The Djs player applies the pitch value before sending the output to soundcard. When Exact Pitch Control is disabled, then Djs "ask" the soundcard's driver to play at the frequency value according to the pitch. The result is unprecise as often depends on how the driver is implemented, soundcard capabilities, OS / DirectX version etc.

Questions, comments

so that's my reponse

the distortion is only there because i changed the pitch, i tried severals songs and i have the same problem.

on pitch 0 i have a perfect quality and if i change the pitch, there is this distortion that appears, this is only when "Exact pitch control" is activated.

in fact i observed that this problem is only avaible on some songs so i think that this distortion only appears on some type of sound because i don't have this problem with Deejaysystem MK 1 under 98 with changing the pitch on the same file that the sample i gave you.

so do you think it could possibel to fix this very little problem because i know that you want the best sound quality for your software (i red the FAQ :))

if you need someone to test you futur improvement on Deejayssytem about this little lost of quality, i would be glad testing if you want.

Ok, I see. Yeah, I need more testing in order to define where the problem comes from.

Please perform the test in these scenarios:

• Pitch set at a value near +1.0 (i.e. +0.9, +1.1), trying at the same song that distorts with +1.0.
• Look for different positive pitch values where the problem arises (maybe +2, +5...).
• Negative pitch values.

Also, it would be helpful for me to have two versions of the piece of song you linked above: one version fully played at 0.0 and other fully played at +1.0.

Deejaysystem Mk-I is not representative on this problem. Mk-I sets the pitch by directly asking the driver to do so. As this is done throught DirectSound, the pitch effect moves effectively in steps on around 0.25. This means that when you have pitch 0 and start increasing it in 0.01 steps, there's no effective change in the sound until the value reaches a point near 0.25.

It's almost sure that when the pitch label shows +1.0 in Mk-I, you really receive a different pitch value in the sound. That's why the "Exact Pitch Control" option was implemented.

Hello, there you go with the original sample , download it here

http://www.geocities.com/retergu/Quality.mp3 - http://www.geocities.com/retergu/Quality.mp3

the file is 1 mega big (192 quality) so i hope that it will still stay on Geocities

i can already say you that, i observed that there isn't any difference between the pitches but the distortion does more appears in low volume part of the song and generally when there aren't lot of sounds in this low wolume part of the song

but i will try your scenarios so that we are sure that this distortion take place on every pitch that differs from Pitch 0

Happy to help for fixing this problem

Edy wrote: Ok, I see. Yeah, I need more testing in order to define where the problem comes from. Please perform the test in these scenarios: Pitch set at a value near +1.0 (i.e. +0.9, +1.1), trying at the same song that distorts with +1.0. Look for different positive pitch values where the problem arises (maybe +2, +5...). Negative pitch values. Also, it would be helpful for me to have two versions of the piece of song you linked above: one version fully played at 0.0 and other fully played at +1.0. Deejaysystem Mk-I is not representative on this problem. Mk-I sets the pitch by directly asking the driver to do so. As this is done throught DirectSound, the pitch effect moves effectively in steps on around 0.25. This means that when you have pitch 0 and start increasing it in 0.01 steps, there's no effective change in the sound until the value reaches a point near 0.25. It's almost sure that when the pitch label shows +1.0 in Mk-I, you really receive a different pitch value in the sound. That's why the "Exact Pitch Control" option was implemented.

Maestro wrote: Version MKII 1.01.00 dont have this pitch problem that the new beta has.

every version that does have "exact pitch control" have this problem and without this option you don't have a precise pitch control under Windows xp

I'd bet the problem won't happen on negative pitch values

I'll explain it after you post the results.

Okay, i tested it and nothing to do, the distortion is always there, i tested with the same sample that i gave you.

I tested ALL the different pich from -8 to + 8 wich changing per each 0.1 pitch.

the intensity of the distortion does change a bit if you compare the negative value to the positive one.

exemple, -6 does give a harder distortion that +6, but the difference is very litle

so i am curious to know why you said taht i won't happen on negative values ?

Edy wrote: I'd bet the problem won't happen on negative pitch values I'll explain it after you post the results.

I thought it wouldn't happen on negative values because the way the pitch is applied on the waveform, but now I see I was wrong.

Let's suppose we have ten samples of a waveform, numbered 0 to 9. Their original sequence is:

0 1 2 3 4 5 6 7 8 9

When applying positive pitch, it means that the same 10 samples must be stretched in order to have less length, for instance 8 samples length. This is done by removing samples from the original waveform:

0 1 2 4 5 6 8 9

When applying negative pitch, the 10 samples are stretched in order to have more length, for instance 12 samples. This is done by copying adjacent samples into the new positions:

0 1 2 3 3 4 5 6 7 7 8 9

This doesn't involve sample removal, so I thought distortion wouldn't appear here.

Greets

What could create this little distortion then ?

i am nearly sure that, this distortion does only appears on certain type of sounds. it could be that the "exact pitch control" driver does juste gave a bug on certain type of hertz.

i think i feel which type of sound does create a distortion but i can't describe it.

just a little idea from a complete noob exe editing :)

Now i am curious to know, what do you think about it ?

don't forget, if you need help, simply ask me

Edy wrote: I thought it wouldn't happen on negative values because the way the pitch is applied on the waveform, but now I see I was wrong. Let's suppose we have ten samples of a waveform, numbered 0 to 9. Their original sequence is: 0 1 2 3 4 5 6 7 8 9 When applying positive pitch, it means that the same 10 samples must be stretched in order to have less length, for instance 8 samples length. This is done by removing samples from the original waveform: 0 1 2 4 5 6 8 9 When applying negative pitch, the 10 samples are stretched in order to have more length, for instance 12 samples. This is done by copying adjacent samples into the new positions: 0 1 2 3 3 4 5 6 7 7 8 9 This doesn't involve sample removal, so I thought distortion wouldn't appear here. Greets

Now I easily see what is creating the distortion.

Imagine you have a pure sine wave in form of PCM data. If you strip one of its values then a vertical "jump" will be put at the waveform. On the other way, if you dupplicate one of its values, then an horizontal "step" will be generated as result.

Overally this is a very small distortion, but it can be more noticeable in some kinds of sounds than others.

The solution would be to substitute the current fast (assembler coded) pitch routine with a slower but more accurate resampling routine. As currently there are lots of works to do at the software and this is a minor issue, it could take time to be implemented in the software...

...BUT: if you (or someone else) could provide me with a ready-to-use resampling routine as source code, then I'd gladly put it into the Deejaysystem application. As "payment" I can offer a free license for the application. The author will also be properly credited.

Anybody interested please send a PM or an email via the "Contact Us" form.

Greetings!

I do also a little bit understand what created this distortion.

Do you think that this little distortion will always exist even if you use the futur implemented ASIO ?

to find someone that could give you "a ready-to-use resampling routine as source code" you could maybe put on the Deejaysystem site a post about it and also to post topics on other forums.It is sometimes amazing to see how people can help or want to try to fix something on softwares

i will check often to see if there is any progress about this

Edy wrote: Now I easily see what is creating the distortion. Imagine you have a pure sine wave in form of PCM data. If you strip one of its values then a vertical "jump" will be put at the waveform. On the other way, if you dupplicate one of its values, then an horizontal "step" will be generated as result. Overally this is a very small distortion, but it can be more noticeable in some kinds of sounds than others. The solution would be to substitute the current fast (assembler coded) pitch routine with a slower but more accurate resampling routine. As currently there are lots of works to do at the software and this is a minor issue, it could take time to be implemented in the software... ...BUT: if you (or someone else) could provide me with a ready-to-use resampling routine as source code, then I'd gladly put it into the Deejaysystem application. As "payment" I can offer a free license for the application. The author will also be properly credited. Anybody interested please send a PM or an email via the "Contact Us" form. Greetings!

Yes, it will occur also with ASIO as the distortion is produced before the audio reaches the output device.

I'd like to make Deejaysystem accept audio plugins, so anybody could contribute. It's another "to-do" feature.

Greetings!

That is a excellent idea, so i hope that a lot of people will do great plugins. if you take as exemple Winamp, see his evolution and count how many plugins so i think there is hope for Deejaysytem.

don't forget to cummunicate this project at the good time and the good place  to have lot of reactions, i am still ready if you need beta testing

Edy wrote: Yes, it will occur also with ASIO as the distortion is produced before the audio reaches the output device. I'd like to make Deejaysystem accept audio plugins, so anybody could contribute. It's another "to-do" feature. Greetings!

cool thread. you guys rock.

Hey Edy, It;s me Dj_Shawn from way back in the day!

You guys ROCK!! I am still to this day ussing DJS for ALL my gigs and night club's....

Now with the new beta VJ-11 I'm about to jump threw the ceiling! WOW!!!! this is going to rock my world once it get's pefected and I can mix videos on the fly to a Big Screen Projector!.. I'll pay whatever you want once this get's perfected.. Also Adding or making a plugin for "Master Tempo" whould just send me to heaven and back!...

You guys have always been the best for Live Computer MP3 Mixing and playing... Now it's finnaly gonna get even better!!! Can't wait!!

I have tried the beta, and yes there are few bugs still left to work out, but from what I've read on here and my personal experience trying the beta, It looks very promising!!. If and when you guys pull this off and get it perfected, I want to be the first one to BUY IT!!!..

Thanks for the GREAT WORK!!! Keep it up...

Hey Shawn!! Nice to see you again!

Sure! We're working hard on VJ-II, and soon it will have special mixing effects like alpha, 3D transitions, and more.

I'm about to release a new beta. When you try it, then please send me all the bug reports and comments. Developing a video player is a bit harder than an audio player mostly because there are many different video formats and codecs out there, and each one can have a completely different behavior in Djs VJ-II.

Thanks for being there!

Best regards!