quality of trackers sound-output

[l8]

i whant to test different trackers rendering abilities. MPT, madtracker, renoise, Bero...and so on and so on...
can anybody help me with understainding these stuff:

bero tracker: SYNC 64...1024, FAST SYNC 64..1024 or pholyphase. choosing SYNC 1024 - overloading my cpu..and nothing`s happening..

(i saw posts connected with bero, that`s why i am asking it here, i am just wondering)

[speed-goddamn-focus]

It's sinc, not sync.  Higher means less aliasing but slower.

[l8]

It's sinc, not sync.  Higher means less aliasing but slower.

less? but why slower?

[LPChip]

It's sinc, not sync.  Higher means less aliasing but slower.

less? but why slower?

You can compare this with an image made with a digital photocamera or scanner.

These devices allow you to set its quality by setting a DPI. (camera has mega pixels)

What basically happens is that the higher you set the quality, the bigger this image will be.

For instance, if you put it awefully low, you will get an unsharp image and only the global image is visable. The more you enlarge this image, the more details are being captured.

To give a small example:

Code Select Expand


small: (eg low quality)

-##
#--
#--



Code Select Expand


big: (eg high quality)
----#######
--##-------
-#---------
-#---------
#----------
#----------
#----------
#----------
#----------


You can see that it was curved here, but it could also be else.

The higher the resolution is, the more precise it becomes to the original. But because it has this resolution, it means that it has all these bits to identify how it was. This results in lots of calculations, which makes it slower.

Same happens with audio. You can see the waveform as an image. The higher the resolution is, the more detailed the waveform is. For example: downsample a waveform alot of times, and see how it changes in shape. You can see that you loose valuable data, and the sound changes accordingly. This process is inreversable. This basically also happens when choosing a resampling mode. The sound is being adjusted to that resampling mode. The higher it is, the better high quality sounds will become. If you only work with low quality stuff that is below your resampling method (they must be really bad) then it might be preferable to choose a lower one. (chiptunes often are made with no resampling because the original sample is made to sound like that, and no conversion is wanted.

Note that the anti-alaising that is being done can improve the sample's quality, so unless you really want no resampling, choose a good one.

[Sam_Zen]

Keeping quality demands a high resolution. But high-resolution data demands more space. And more space demands more speed to handle it. The human range sets the boundaries in this. It would be totally useless to make sure, that in a composition a frequency of 56 kHz can be heard. Only some pets would get a bit alert maybe.
As usual, it's a matter of finetuning with the personal equipment. In the end, only the quality of the suggestion counts. If it works, it's sufficient.

The choice between using 22 or 44 kHz samples for example. This choice does mean a double filesize, more memory.
But due to the nice resample algoritms of MPT, the 22 kHz samples sound e.g. like a 33 kHz sample. Which appears
to deliver an acceptable cd-quality, with no clear differences in hearing a file with the 44 kHz samples.

But apart from this kind of efficiency, the quality of the output starts with the quality of the samples.
And in those years I've seen too much tracker-compo's going by, where maybe the score was ok, but where the collection of used samples consisted of a bunch of sloppy crap.

[speed-goddamn-focus]

But due to the nice resample algoritms of MPT, the 22 kHz samples sound e.g. like a 33 kHz sample. Which appears
to deliver an acceptable cd-quality, with no clear differences in hearing a file with the 44 kHz samples.

Once again, this depends on the frequency contents of the sample.

[l8]

i render 1min  (9:04-10:04) of high quality-xm (an-dream.xm) via these trackers:

96khz,32bit float, maxed output settings:
skale
mpt
bero
xmplayer (not a tracker)

imho results:
1.skale (clear, bright, attacking sound, i think that is happening due to applying dynamic compression when rendering to wave, and the sound "looks" better...)
2.mpt/xmplayer (rather good)
3.bero (very strange sound, f#cked up some high frequencies and feeling of presence in solo,
i use polyphase filter)

renoise, psycle, buzz - cannot play correctly it/xm. they spoil panning, volume and pitch of samples. renoise have the best sound output among these trackers.

48khz, 16bit (max capabilities of these trackers)

it (64khz)
schism
milky

imho results:
1.milky goes first, very good sound
2.then it
3. and schism - somewhat dirty sound..i dont like it at all

the results of test you can see here

http://www.l8.democoder.ru/tmp/test.png
(1.8mb png image)




this is my site, u can try my song made in schism
http://www.l8.democoder.ru/trk/mountain.teraphy.of.pain-l8.ogg (3mb)
solo in the end of track taken from pink floyd, but i`ve played it myself..

rather ditry sound...cause of using schism..

[speed-goddamn-focus]

http://www.l8.democoder.ru/tmp/test.png
(1.8mb png image)

Nice graphs, but it's quite impossible to say anything of the quality just by looking at them, don't you think?

[l8]

http://www.l8.democoder.ru/tmp/test.png
(1.8mb png image)

Nice graphs, but it's quite impossible to say anything of the quality just by looking at them, don't you think?

i`ve got the waves samples, but i don`t have such many hosting space in the  internet for them.

my subjective opinion about output quality if these trackers i wrote higher.

[speed-goddamn-focus]

my subjective opinion about output quality if these trackers i wrote higher.

yes I saw that, I just wondered if there was something in the image I didn't understand... subjective opinions are what matters the most tho, so biggup to you!

On another note I guess MPT is hard to compare since it allows sample specific resampling modes.

[anboi]

all the resample modes of mpt are not quite good enough for my liking - there are a lot of ringing high bits coming in when pitching things up or down a lot. i find that the old xmms is better than the polyphase some of the time - usually on bass stuff. mpts resampling is still pretty good though.