Request: Depeche Mode "Perfect" remixes

[Fred Taghon]

I have the Ralphi Rosario 2009 club mixes of Depeche Mode's "Perfect" in 320kps MP3, but am curious if anyone has the actual USA Promo CD http://www.discogs.com/Depeche-Mode-Per ... se/2696902 and are willing to share in WAV or FLAC?

The mixes were also available in WAV format at one time - I believe on Beatport - but they were not available for purchase from U.S. customers so I wasn't able to buy them.

Thanks if anyone can help...

[eddie]

Well this is perfect but not a remix


https://soundcloud.com/user-36975559/ec ... al_sharing

[AudioChimp]

all tracks are here....

https://we.tl/t-rYctEwdZZf

Link available for 7 dats

[Fred Taghon]

all tracks are here....

https://we.tl/t-rYctEwdZZf

Link available for 7 dats

Wow, 8yrs later, request filled! I even ran 'em thru a frequency spectrum analyzer and they all indeed come back as lossless! Are you aware of the source? According to Discogs the USA promo CDR was low quality lossy. As I had originally posted, I seem to recall these tracks being for sale as lossless downloads from Beatport (maybe even Junodownload?) at some point in time, but oddly enough, if you lived in the USA you couldn't buy them.

The USA-only remixes were also affiliated with the Pro-Motion DJ service, which around 2009 often offered free lossless downloads directly from their SoundCloud page, so maybe they came from there?

Regardless of where they come from,.thank you for sharing them!!!

[PKPN]

I even ran 'em thru a frequency spectrum analyzer and they all indeed come back as lossless!

Well, sorta. They are indeed cut on CD sector boundaries and I suspect they are legitimately a lossless copy of what's on the CD-R. However...

1. The Ralphi Rosario & Jody Den Broeder Dub is lossless but is sourced from vinyl; it's full of clicks.

2. The other tracks are all sourced from MP3s. In a hi-res spectrogram of just the first few seconds, the visual artifacts of MP3's frame structure and selective preservation of frequencies are obvious, especially in the uppermost frequency bands. It's mastered loud, no highpass filter was applied during encoding, and it's likely 320 kbps...maybe that's what fooled your analyzer.

[AudioChimp]

AUDIOCHECKER v2.0 beta (build 457) - by Dester - opdester@freemail.hu
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-=== DO NOT EDIT THIS FILE! ===-

Path: ...\RALPHI

1 -=- 01 - Perfect (Ralphi Rosario & Jody Den Broeder Vocal Mix).wav -=- CDDA (100%)
2 -=- 02 - Perfect (Ralphi Rosario & Jody Den Broeder Dub).wav -=- CDDA (99%)
3 -=- 03 - Perfect (Ralphi Rosario & Jody Den Broeder Mix Edit).wav -=- CDDA (100%)
4 -=- 04 - Perfect (Ralphi Rosario & Craig J Vocal Mix).wav -=- CDDA (100%)
5 -=- 05 - Perfect (Ralphi Rosario & Craig J Mix Edit).wav -=- CDDA (99%)

Summary 99.60% CDDA

149670420

this is an audiochecker log of the files

i will upload the original files from the CDr and you and feel free to compare (they are definitly from a lossless source)

remember all these songs are remixed by a dj they sometimes use various sources when creating mixes.

[Fred Taghon]

I even ran 'em thru a frequency spectrum analyzer and they all indeed come back as lossless!

Well, sorta. They are indeed cut on CD sector boundaries and I suspect they are legitimately a lossless copy of what's on the CD-R. However...

1. The Ralphi Rosario & Jody Den Broeder Dub is lossless but is sourced from vinyl; it's full of clicks.

2. The other tracks are all sourced from MP3s. In a hi-res spectrogram of just the first few seconds, the visual artifacts of MP3's frame structure and selective preservation of frequencies are obvious, especially in the uppermost frequency bands. It's mastered loud, no highpass filter was applied during encoding, and it's likely 320 kbps...maybe that's what fooled your analyzer.

Hmmm that's all very interesting, and I admit I didn't see the post and download the tracks until very late in the evening and haven't had time to listen them closely yet. The frequency analyzer app I use is FakinTheFunk for Macs and while it's not 100% fool proof, the only times I've seen it be wrong is when I load in CDs I actually HAVE in my possession from the 80s that occasionally register as 320 kps MP3s lol and we all know MP3s didn't exist in the 80s. But most recordings in that era were mastered for vinyl so the frequency spectrum generally wasn't as wide as more modern recordings. So if anything, it seems to be overly cautious. My 1987 New Order "Substance" discs for instance, have results ranging from 128-320 kps MP3s, with NO tracks being considered as lossless.

[AudioChimp]

Just so you can compare the tracks i posted in previous post to the ones ripped from the original CDr....

https://we.tl/t-4QqO2qfW1l

these ones are lossy sourced yet the original tracks from the CDr frome here -

https://www.discogs.com/release/2696902 ... en-Broeder

[AudioChimp]

2. The other tracks are all sourced from MP3s. In a hi-res spectrogram of just the first few seconds, the visual artifacts of MP3's frame structure and selective preservation of frequencies are obvious, especially in the uppermost frequency bands. It's mastered loud, no highpass filter was applied during encoding, and it's likely 320 kbps...maybe that's what fooled your analyzer.

i had the tracks all from varying lossless sources and added 2 seconds to each track to provide unity to the start of all the tracks (the 2 second segments were copied and pasted from the end of each track), using audacity i then selected the 2 second part i added and generated 'silence' so no noise would occur from the copied segment, ... no other part of the files were added to or altered

here is an analysis one of the tracks you mentioned, i am not an expert on audio but is this a sign of an mp3 sourced file?

[PKPN]

here is an analysis one of the tracks you mentioned, i am not an expert on audio but is this a sign of an mp3 sourced file?

Even when looking at the entire file's spectrogram all at once like this, I do notice the telltale "gritty" look of lossy coding in the upper frequency bands. If you were to put just the first few seconds of the file into a file and run Spek on that, you could verify; you will see something similar to what I posted, where the upper part of the spectrum has had a lot of sound removed, and what remains behind is in discrete chunks, all the same size (the little dashes in the image I posted). This gritty texture actually extends all the way down into the lower bands but it's most noticeable at the top, due to the way MP3 deprioritizes that range. Audio that has never been subjected to lossy coding does not exhibit these characteristics.

remember all these songs are remixed by a dj they sometimes use various sources when creating mixes.

Yep, yep... absolutely correct. Some remixes are made from audio taken from different sources, e.g. some samples from MP3, some from CD, with freshly synthesized beats laid on top. This can explain why a spectrogram can have the look of lossless and lossy at the same time. That may have happened here. It's hard to say with certainty.

My point is just that this is one of those times when the fake-spotting tools are not entirely correct when they say this audio is "100% lossless". My experience with promo CD-Rs from this era (mid-2000s to mid-2010s) is that this kind of thing happened a lot, where they contain audio that's clearly from different, not entirely lossless digital sources.

i had the tracks all from varying lossless sources and added 2 seconds to each track to provide unity to the start of all the tracks (the 2 second segments were copied and pasted from the end of each track), using audacity i then selected the 2 second part i added and generated 'silence' so no noise would occur from the copied segment, ... no other part of the files were added to or altered

Depending on your Audacity settings, you may have accidentally added dither (very quiet, high-frequency hiss) to the non-silent parts when saving. In terms of what you hear, it's harmless, but for collectors it's not as ideal as sharing the completely unedited rip. This is one of the pitfalls of audio editors; they work with 32-bit floating-point samples, and then assume you want dither when quantizing to 16-bit integer. You might, depending on what you did to the audio. But if all you did was make simple cut-and-paste edits of audio that was quantized to begin with, then you probably don't want dither when saving.

As for the 2nd rip you shared, yes, that's quite lossy, visibly worse than the first batch. It was converted to MP3 during/after the rip.

[Fred Taghon]

I have found all of this to be very educational. I know a good deal about audio quality and all that jazz, but you have shared some info that has helped me understand this stuff even better! I did purchase the lossless digital download of the Fragile Tension/Hole To Feed single a year or so ago, so I already had the 2 full length Ralphi club mixes from THAT. So out of curiosity I put them through the spectrum analyzer and they do visually look much "smoother" overall than the files audiochimp was kind enough to share. So your analysis is very likely correct. And I agree, the Dub is definitely recorded from vinyl, there's absolutely no denying that. Besides the pop and clicks, an easy way to tell is to rapidly jump playback from the beginning section to the end and listen to the degradation in the percussion, especially any high frequency signals like high hats and cymbals. This change in quality is inherent in all vinyl because the closer the stylus gets to the center of the record, the more condensed the grooves are (or something along those lines, I can't remember the technical terms of it all - but outermost tracks on a record always sound better than the innermost tracks).

[alexx]

Ooh, I just missed these! Any chance of a reupload? Thanks very much.