The author from that link listed above clearly has a complete misunderstanding of CD audio in particular and digitial sampling and pulse code modulation in general.
There is error correction built into the coding, such that you know if you have a correct bit or not. If not, it's an ERROR, and then one uses ERROR CORRECTION and/or ERROR MASKING (including interpolation/extrapolation and/or muting). PCM is built so that the data is divorced from the vagaries of the transmission or storage media. And they DO use memory buffers, so that only 1 clock--that which drives the D/A at the end--affects timing.
Well, I must say I'm a little disappointed that no one has taken up my offer yet (at least enough to PM me). I'll leave the files up for another week, then take them off so I can keep the space down on the site.
Lot's of talk, no action...Oh well.
+ Reply to Thread
Results 151 to 166 of 166
Originally Posted by adam
Anyway, the winner to my ears is WAVE_B. Considering I'm using the built-in sound card on my PC, that's my best guess. If I'm wrong then I'm wrong.
Provide then an unencoded clip of YOUR choosing and I can then provide a single wave file of 20 matched pairs (with encoded juxtaposed with unencoded audio randomly arranged).
With only six clips to choose from, you've got a 1 in 6 chance of actually just guessing the right answer so it doesn't prove much.
i think ROF should take vitalus up on HIS challange........
Concerning Cornucopia's test:
I do not have the professional material, nor the ears, to be able to state which is the original from listening to the 6 different versions.
Using spectrum analysis, I reject B,D and F, because they show a high frequency cut-off. From the remaining three, A and E look the most identical. So I guess C as the original.
But again, some cheating here.
And I since long compile my favourite songs to 128 kbps MP3 "data" CD's, and am largely satisfied by their quality when played on a JVC-XV-N5.
Originally Posted by cloclo
I believe I did SPECIFICALLY state for you to give your guesses to me via PM, so that others could try unswayed as well.
Did you even listen to them 1st and make a guess by ear before running a Spectral Analysis?
Well, just so anybody else would like to try, I'm not revealing whether he's actually right or not, and/or which file is which, until next week...
Originally Posted by cloclo
What makes this clip so hard to determine is that is has no extreme lows or highs and is not really much of a challenge for an MP3 encoder. And who knows what false frequencies are being created when it's converted back to a WAV. It's definitely a tough call in this case. Despite the spectral analysis, I think B & C are really close. If I had a better sound card, maybe I could give a more definitive answer. But I will stick with my original guess which is B, just so no one can accuse me of being influenced by the spectral analysis. If I'm wrong I know for sure I'll never hear the end of it.
OK. Here is the answer to the MusicCD/DataCD and Uncompressed/MP3 test I created earlier...
- 128 mp3 = 4289KB (D)
320 mp3 = 4298KB (B)
ORIG WAVE = 4295KB (C)
DATACD_RIP = 4293KB (A)
MUSICCD_RIP = 4293KB (E)
MYSTERY = 320KB (F)
I find that I can still leave the files on the server a while longer, so if anyone else wants to take the test, even knowing ahead of time the sources and possibly skewing their answers, they're welcome to.
The spectral analysis that cloclo used did determine which ones used compression, but I find it interesting--particularly as pertaining to the original thread--that (s)he mentions A & E being the most identical, as those are the separate rips from DataCDmedia and MusicCDmedia burns.
I think actually you would find that C, A & E are bit-for-bit identical, with the exception of some [DIGITAL SILENCE] offset bits at the beginning and/or end of the rips (or lack thereof).
edit: Maybe next time (if there is one), I'll do this as a poll so it'll get more attention...
- 128 mp3 = 4289KB (D)
Just to stick my oar in, I think a "by ear" test would be better if it were carried out on music we were all familiar with.
When I want to test some hi-fi gear I take CD's of stuff I know well and love. If I don't like the sound of it, I don't buy it. For example, if know there is thunder at the beginning of Dire Straits' Brothers In Arms after all these years played it back on something and it wasn't there any more, don't you think you would notice ?
Earlier someone said about not using a very wide spectrum. It's not just that, of course. Test signals can bugger up most lossy compression techniques a real treat. A steel guitar will rasp if a DAC isn't clocking right. It's not having "golden ears" so much as knowing what test to use to prove a point. I suggest you encode an audio test CD with your choice of codec and see what you think of it then - you might be surprised...
Just found this thread. You guys are going to think I'm crazy too. I have a stand alone Sony CD player/recorder W2000ES bought in 2005. The manual states to use only music cds for recording. I used to record on Sony music cds which worked fine. (they have the blueish coloring). Then I switched to Memorex music disc (silver disc). There is a very slight difference in the amount of bass output (a tad less than the Sony music disc). I prefer the sound of the Memorex as it seems to be closer to the original cd and they look like cds put out by record companies. One more thought - I tried the Memorex "colors" music recordable cds. After a few years I can no longer play them on any player. Maybe the color reflecting the laser?
Technically speaking, there are 3 different CD-R dyes that are used. I assume that music cds have no limitations on which dye can be used. I agree that the "Memorex is better" is just the imagine of orangee. The silver dye actually requires more accurate lasers when writing or it will generate more errors and this dye is also unfortunately much more prone to UV damage than the other dyes. So the problem with the bad Memorex CD-Rs may be that they have been exposed to enough UV to damage them or the Sony CD recorder did not have a very accurate laser and degradation over time has now reached the point where the errors are too great for recovery.
The manual says to use only music cds because these have a built in fee that in theory was going to go recording artists in exchange for allowing you, you dirty pirate bastard (no offense, but that's what the RIAA thinks you want your Sony recorder for), to make illegal copies of discs you don't own. Such recorders (I have an old one too from Pioneer) will refuse to record to normal CD-Rs because the recorder knows that those discs don't have the built in fee. If you're curious you can try to record to any normal blank CD-R, but your recorder will refuse to write to such discs. This is why. The pirate comment was not meant to be a serious comment on why you wanted a recorder, but just a somewhat humorous and unfortunately accurate representation of how the RIAA views you whether you like it or not.
It's worthy to note that no RIAA member ever allowed public audits of the 'fees received' for Music CD-Blank sales, although every member has sworn their bloody oaths that every penny will be distributed accurately. Most of this money goes into a retirement or health-care fund. The more commonly referred to title of this fund is "complete rip-off" - it's the monetary equivalent of Jimmy Hoffa's fate.
Too bad we AND the artists can't sue every RIAA member AND the sellers AND the RIAA itself for not making properly audited dispensation of those funds. But even then, the excuses would merely flow faster and deeper.
(I wonder if the RIAA-MPAA have a forum that could host writers with questions originating in Portugese?)
Last edited by OllieTSB; 4th Oct 2013 at 14:55.
Continuing slightly OT: yeah the RIAA is in a pile of deep doodoo of their own making. Their business model has for DECADES been to use Bully tactics to stick it to both their vendors (the musicians) and their customers. Now digital & p2p has brought down that house of cards and there is very little that can be done to put it all back together again. Unfortunately, ultimately we as musicians and/or customers are still on the losing end.
But enough ranting. Back On Topic: It is the nature of the digital medium that "quality" in an pcm/linear setup is SOLELY determined by the A->D, D->A, and prior & subsequent analog & physical/environmental sections (and our ears). The ONLY way a digital storage or distribution pathway can affect quality would be if it were to be at the point of generating uncorrectable ERRORS. So the whole idea of "these CDs sound better than those CDs" is complete unscientific bunk.
Now, I believe jman98 is referring to dye FAMILIES, so there is some possibility for variability there (plus earlier in CD's lifecycle, there were other dye variations). But dye has NOTHING to do with whether a disc is a "data CD" or a "music CD". Only a pre-written bitcode. And "music CD"-only recorders (yes, there are some pro CD recorders that DON'T require music CD: Alesis Masterlink) are set to shut down with the wrong or missing bitcode. Sounds like Cinavia (though the 2 technologies are completely different). Dye has only to do with error-stability & longevity.