[Synth-pop] CityNights (it)

Rylo · May 29, 2006, 15:51:21

Hi there, after a few months spent dabbling with Modplug tracker I have decided to take the plunge and upload my first full song. It's not actually quite finished yet, it's a little repetetive and I would really like some feedback as to how I can improve it.

The song was written in Impulse Tracker format and took about 4-5 hours to complete, file size is 904kb and the song length is just shy of 5mins.

Here's the link:CityNights

I hope you enjoy it!

Thanks

Rylo

Sam_Zen · May 30, 2006, 00:43:25

Nice format with attention to open and close as well. Harmonic changes are quite predictable, but that's no problem.
At about 1:03 a melodic sound is introduced, which keeps cominge back later, and seems to be a little low on volume in the balance with the rest. Since the instrument is quite 'harsh' : didn't you dare to make it more prominent, or is this intentional ?

I also spotted an unclean sample, some LF rumble sputters. I suppose from the 'sparkle' sample, if played back low.
While checking this, I found that you could improve things with some 'normalizing' of the material.
It's always better for quality to play a max volume at a lower level, than to amplify a weak sample.

Rylo · May 30, 2006, 12:30:57

Sam, many thanks for the feedback.

QuoteNice format with attention to open and close as well. Harmonic changes are quite predictable, but that's no problem.
At about 1:03 a melodic sound is introduced, which keeps cominge back later, and seems to be a little low on volume in the balance with the rest. Since the instrument is quite 'harsh' : didn't you dare to make it more prominent, or is this intentional ?

I agree the harmony changes are quite predictable. I was aiming for a poppy kind of sound, nothing too complicated, but at the moment I'm reasonably happy with the progression and structure (though I will try and chop it down to 4mins-4 1/2 mins in length!)

The instrument at 1:03 is quite a harsh sounding/loud sample so I did turn the volume down on it so as not to completely dominate that part of the pattern. Perhaps I could do with turning it up a notch?

QuoteI also spotted an unclean sample, some LF rumble sputters. I suppose from the 'sparkle' sample, if played back low.
While checking this, I found that you could improve things with some 'normalizing' of the material.
It's always better for quality to play a max volume at a lower level, than to amplify a weak sample.

I have noticed this as well and yes it appears to be the 'sparkle' sample. I will be editing this sample and giving it more echo and removing the 'clicky' sound you can hear at lower octaves. I'm not quite sure what you mean by 'normalizing'? Could you give me a quick overview or point me in the direction of a thread/tutorial?

So, my next tasks are:

1.Cut down the song length, add more variation.
2.Increase volume on the sample riff that comes in at 1:03
3.Improve the sparkle sample

Any other feedback is more than welcome!

Sam_Zen · May 30, 2006, 23:10:46

Trying to cut down the song length is most of the time a sane goal anyway. I often save the original for archiving, and make a 'short version' of it, with (almost) no doubles in the pattern row, and saved under a slightly different name.

For quality reasons it is best to set the final volume of a sample in the instrument setting, not prepared in the original sample itself. So if you need a soft hihat-sound, don't use a soft sample of a hihat, but make it first as loud as possible without distortion. Then, if you want, you can cutdown the volume in the global setting, or with some volumecodes in the pattern. 'Amplifying' something also can mean : lower the volume.
So you have control over the volume-balance between the instruments inside the tracker-module, instead of depending on the strength of the sample-files. If you import a 'weak' sample, you have to increase the volume for the proper use, but then you will amplify the possible noise in the sample as well.
This 'as loud as possible' is called normalizing. Such a function is available in the sample tag of (O)MPT.
It's preferrable to do this in the sample-making stage, but if you don't have a wav-editor, you could use this function, save the sample as a new one and import it again.
I have published some (more to follow) tips on this page

CrazyAznGamer · May 31, 2006, 01:01:20

The normalization of sample 3 would help it out much. (While you're at it, normalize everything)
It maek 1:03 sound that much more bettar.
(Also, try decreasing instrument 3's line an octave. :wink: )

Rylo · June 08, 2006, 11:08:42

Bttt for one last cheeky bump!

OK, I have almost finished re-editing the song now. I have the length down to under 4 mins and added another melody part. I've tried to normalise the samples as best as I can but it's a lenghty process as all the volume variables in the patterns need to be changed as well.

Only problem I'm having is I can't seem to clean up that damn sparkle sample, so I'm still left with the slight 'clicking' sound in the background.

I just thought I would ask one last time for any feedback on this song before I post up my finalised MP3.

Thanks

Rylo

apple-joe · June 08, 2006, 14:41:15

I like this one. Steady and efficient bass line. The strings seem simple, yet very nice. You seem to be able to create workable things out of few resources, so to speak.

The high pitched sounds are also very atmospheric. Defines the mood.

I don't know if I would make it shorter. Rather spice up some parts of the track. It's nice.

Rylo · June 08, 2006, 16:29:34

Wow, many thanks for the kind words apple-joe! I have added another short melody part to the song and I have also made use of some plugin effects, so hopefully this will 'spice it up' just a bit! :wink:

Thanks for the feedback!

Sam_Zen · June 09, 2006, 00:31:53

I was challenged by the 'damn sparkle sample clicks', so I decided to take a closer look.
It didn't appear to be some distortion like clipping, but a very complicated thing in the wav-file :
Every 'click' is caused by a rather peculiar shift of DC-bias of the wave-form with a certain curve of deviation.

So this can not be solved with some bias-correction, thus the only thing left is some precise filtering.
I gave it a try : Sparkle sample y.

LPChip · June 09, 2006, 08:34:52

When I look to the sample, It almost looks like the sample is looped and then turned around... But It wouldn't surprise me at all if setting new looppoints can remove the clicks. But then again, I haven't heard the sample so I don't know if its possible.