[Orpheus]

heh he... wanna have some fun with this...

well, my original version of Tears for You was extremely noisy. so, since then, I have tried to reduce the noise a little. so, i used a filter that uses a sample of "silence" as a model. then it tries to subtract that sampled model from the whole selection. well, the problem is, i cannot get a clean sample, as there is no noise "silence" that does not contain any note at all. so, the best that i could do was pull a 1/10sec sample from the very last part of a measure's rest, while a note was being held with the pedal down. it's the best i could do.

so, with that sample i proceeded to process the file... and voila, no hiss for most of the song! i was so happy! but that was monitoring on my near-field system. when i put on the headphones, it was another story:

i assume the imperfect sample caused an irregularity. listen to the original (the one i posted here earlier):

http://www.fallingtodreams.com/music...0for%20You.mp3

now, the noise should be very evident. lots of it. and it sucks.

so, here's the processed file:

http://www.fallingtodreams.com/music...REMASTERED.mp3

listen carefully. you hear something weird!? ...there's some kind of high frequency distortion/noise. on loudspeakers, you can't hear it at all. or if you have really badass ears and a quiet as hell room, maybe. but on good headphones (i use a Grace 901 + Sony CD3000's), it shouldn't be too hard to hear.

so, tell me... what you think?--which version should i use? also, if you have EQ suggestions (even just for tonal reasons), let me know.

keep in mind, 99% of the public ain't as picky as us.

orpheus

 

[Canman]

Dean, I've only listened on my computer speakers so far, but I wanted to tell you that I enjoyed your composition. The dreamy harp-like intonations put me in a nice relaxed state. Time to go to bed. Will look for high frequency noises mañana.

 

[lan]

Yeah I hear something weird. It's called MP3

No but really though, it's tough to judge because it's an MP3. I dont know if it was your processing or the encoding.

 

[Orpheus]

it was encoded at 128, so, most probably not the MP3 effect. yeah, MP3 does sound somewhat similar, but no... the weirdness is not due to the MP3 encoding processing.

but to really tell, just listen to the first version, and the second version. and difference between the two is the result of the processing. they otherwise should be identical.

 

[ooheadsoo]

Hey Dean. My guess is that the silence sample you are using doesn't embrace the entire spectrum of the hiss. Notice at 1:18, when you start to play louder, the hiss comes back. Moving on from that point when you start to diminuendo, I think it sounds a bit more apparent at least to my ears that the high frequency bits are remanants of hiss left over from the processor's subtraction.

This is on my meta42+revo+senn600.

 

[itza2mer]

I agree with ooheadsoo, he's spot on.

Stick with the original, hiss and all.

I listened with Foobar 2000/Revo 7.1/Marantz 2220b/rewired-modded AT A900 with Zu Mobius cable.

 

[eric343]

Quote:

Originally posted by Orpheus it was encoded at 128, so, most probably not the MP3 effect. yeah, MP3 does sound somewhat similar, but no... the weirdness is not due to the MP3 encoding processing.

128kbps MP3, and you're saying the MP3 encoding is not the problem?!

Try LAME, --alt-preset-extreme or whatever, (or better yet, something entirely lossless!) then send it around. 128kbit is not anywhere close to "high quality" much less "mastering grade"

 

[itza2mer]

Is there any way for you to post a wav file somewhere, so that we can take a stab at cleaning it up just for fun?

 

[Orpheus]

Quote:

Notice at 1:18, when you start to play louder, the hiss comes back.

yes, but that's not because the sample's incorrect. if you listen to the original, at that point, there is much more hiss than at other points. the hiss increases and decreases in various areas, and is not at a constant volume. i set the noise remover to attenuate the noise by 30db. it also has attack and release functions. due to those variables, at that point in time the hiss comes back. i can attenuate the noise even more, but then the overall sound suffers--the reverb sounds unatural, and it just doesn't sound as good anymore.
Quote:

I think it sounds a bit more apparent at least to my ears that the high frequency bits are remanants of hiss left over from the processor's subtraction.

yes, exactly. well, almost exactly. what it is is a function of the removal of what the processor thought was noise. but as i said before, the noise sample was not 100% noise, and included a small amount of tone that was not noise. and i believe this is why we're left with those high-frequency artifacts.
Quote:

Stick with the original, hiss and all.

yup, that's one answer. but why do you say this?--that's what i'm looking for. again, the noise is inaudible with normal speakers. in fact, Nora Jones famous CD has more audible distortion than that!
Quote:

128kbps MP3, and you're saying the MP3 encoding is not the problem?!

but it's not. yes, MP3 results in lower grade sound. but to me, it isn't that noticeable. i know you disagree... but really, the MP3 effect is much less apparent than the artifacts that you should notice comparing sample #1 to #2. again, #1 and #2 were from the same recording. they are identical except one was processed to reduce noise, and they other wasn't. therefore, if you compare one to the other, you should be able to tell what's cause by MP3 lossy compression, and what's the effect of the noise reduction.
Quote:

Is there any way for you to post a wav file somewhere, so that we can take a stab at cleaning it up just for fun?

heh he, i dunno how many people want to sit around downloading a 50mb file from the net. ...and besides, it would take me quite a while to upload it too.

 

[ooheadsoo]

Yeah, I noticed that the hiss changes according to how you're playing your keyboard. A time consuming idea is to take a new sample for smaller segments.

Gee, you guys listen at really high volumes, because for a good 30 seconds, I was wondering..."What hiss?," until I turned it up good and loud. I'd recommend sticking with the original because it's more consistant. It sounded odd to me at 1:18 when the hiss suddenly appears, though it does that too with your original, but on a more consistent basis. You can train yourself to tune out a bit of hiss, like you may have to do with vinyl.

 

[lan]

I also listen at low volumes and couldn't notice the hiss much. I had to turn up to 3 oclock to hear it.

 

[itza2mer]

I listened at a louder than normal level. There are several new and remastered CD's that have hiss as loud (or louder) than the hiss on yours. The original sounded 'better' to me (hiss and all). I guess it sounds 'warmer' with the hiss, and 'dry' without.

 

[Edwood]

How about posting an APE or FLAC instead of a WAV?

I'll try to remember to have a listen tonight.

-Ed

 

[Orpheus]

heh he.... alright. i still don't see the point. again, you have the original and processed files. even if mp3 loss is a concern, subtracting the original from the processed file would result in just the processing.

but anyway... send me a link to a utility that can convert the original .wav's to the lossless format of your choice.

dean

 

[Orpheus]

Quote:

I also listen at low volumes and couldn't notice the hiss much. I had to turn up to 3 oclock to hear it.

Quote:

Gee, you guys listen at really high volumes, because for a good 30 seconds, I was wondering..."What hiss?," until I turned it up good and loud.

you know what guys.... i just did a bit more listening of the original, and hmm....... the hiss really ain't that loud on headphones. it's really weird, cause when i play it on the near-fields, it's quite evident. when letting friends play it on their low-fi systems, i hear lots of hiss too. really strange. usually headphones are better for picking out details like this........... i dunno.

i think i'll keep the original then. cause as you guys noted earlier, the processing also leaves the music a little dry'er. this is also probably because the noise sample included a bit of the reverb tail, which would result in less ambience.

heh he....

so, any EQ settings you guys want to try out?--or leave it?