DSD Vs. PCM Very ineresting article

samandhi

I find it espcecially interesting that in order to do any sort of EQ, or reverb, or etc... the recording must be first encoded to PCM, then re-encoded back to DSD:

http://www.ayre.com/insights_dsdvspcm.htm

ForSerious

The only advantages I could see DSD having is if it was smaller or less taxing to decode than PCM, and had surround sound channels.
I still think there is awesome musical potential in surround sound, where each channel has different instruments playing. So far though, all surround sound recordings I've heard just have standard generated reverb or an panoramic surround effect in the rear channels.

The ultimate in music recording formats would be one where each instrument is stored in it's own spot so that the listener could turn on or off what is being played back. Creative lab's sound blaster live value soundcard had a little demo with one short song recorded in this fashion.

BruceBanner

Samandhi: For a digital EQ to edit the digital recording, yes you are right... you need uncompressed data.

... but having full controll of the amplifier with a frequency based digital potentiometer you can completely do an EQ... like the FiiO-DAPs.. only by loosing some potentiometer steps.... even if it's just an decoded MP3-128 signal ;)

samandhi

ForSerious replied at 2015-10-1 04:36
The only advantages I could see DSD having is if it was smaller or less taxing to decode than PCM, a ...

  That is one of the reasons that I actually liked DVD-Audio better than SACD. While it is still PCM (not that I think DSD is any better sounding than PCM) it comes closer to that discreet thing you were talking about, because I agree that would be VERY good stuff.

Samandhi: For a digital EQ to edit the digital recording, yes you are right... you need uncompressed data.

... but having full controll of the amplifier with a frequency based digital potentiometer you can completely do an EQ... like the FiiO-DAPs.. only by loosing some potentiometer steps.... even if it's just an decoded MP3-128 signal ;)

  Yeah, I agree. I didn't explain myself better. Concerning my OP, I meant that all this was mandatory on the recording side. After that, anything is possible.

samandhi

  I guess, the reason I find it interesting that VERY little music is actually recorded, mixed, mastered, etc in pure DSD, is because; whenever you convert from PCM to DSD or vice versa, you are now talking about a lossy format. The whole point, from the end user listener, of DSD (or even high def PCM) is that it is of the highest quality available (no loss of information). By converting (in any stage of the recording), you are losing some of that information, which makes the point of being able to use DSD on a DAP (for instance) kind of pointless.

  I would also argue that because of this, when downloading DSD files from here (for example) you are not getting pure DSD music. Somewhere in the chain it has been converted, which in most cases, is worse than just getting 24/192 PCM recordings (because it is pure through the whole chain). This is NOT to say that it is impossible to produce music using pure DSD, but the equipment is so very expensive to do this, that I think it highly unlikely that ANYONE does this.

  I (personally) think that DSD is a sham started by Sony marketing because the patent licensing of CD is no longer under their control. I think that DVD-Audio would have done better, had the organization that implemented it were more... well, organized.. heh

  On a personal player, I find that DVD-Audio loses its advantage (for me), but in a good surround system, I think that it sounds better (again, to me) than does SACD. Anyhow, just my thoughts. Feel free to comment.

MisterC925

I've found analog -> DSD (ie older pre-digital recordings) are amazing in DSD, as are orchestral performances recorded live as DSD with no post processing.   

Since most studio recording these days is recorded and mixed in PCM, I don't see an advantage for DSD on material that was originally PCM.

So keeping the material as close to it's original format is best IMHO.

samandhi

MisterC925 replied at 2015-10-2 00:53
I've found analog -> DSD (ie older pre-digital recordings) are amazing in DSD, as are orchestral per ...

  I would agree with that. If a recording were recorded in DSD and kept in DSD, I would find it very good. I also think that if it is PCM, keep it PCM and it will be better for it.

ForSerious

From what I've read, the process of analog recording adds pops and clicks even in the master recording, just because it is not a perfect medium. ...Something about lots of CDs, when they were first introduced, sounded bad because there was no remastering done on them. Would this remastering process be considered an impurity?

samandhi

  That is a great question. I suppose it depends on who you are and what you are looking for in a recording. If you are a purist, then, yes I would say it IS an impurity. Those that believe that vinyl is the most pure of sounds would probably agree with this.

  If you are one who just wants to use their ears only for judgment (not relying on graphs and tables), then they might say no, it's not.

  The problem is that there are 2 camps, analog, and digital, and both have good arguments. My problem is when someone complains about something not being so pure (on paper), but don't really have the equipment to be able to tell the difference anyhow. If one is going to make an argument one way or the other, you had better be able to back it up via experience, rather than just deciding "that sounds logical, I think I will join that camp".

  I would agree that recording analogously DOES add pops and clicks (to a point), but I also think, ignoring those things the method sounds better to me. I cannot stand the way they master music today (this being an extreme opposite only for comparison). The amount of compression that is used on any given album is riduculous. For example Rush remastered most of their albums. There was one album (can't remember which atm) that had to be re-remastered because there was so much compression used on the album that everyone complained.

  This, not to be confused with the loudness wars, but usually when you find one, you will find another.

  I, personally believe that 16/44 cd redbook sounds really good. It is all in the recording to my ears. I am not a huge fan of the group or the music, but Fleetwood Mac is a really good example of a group that really care about how it is recorded, and mixed. Dokken is an example of a band that does not. While I love their music, they record, mix, and master their music like crap (just throw some reverb on the mix and call it a day).

  I am not here to say that 24/96 is not better, on paper, but I would still argue that it is redundant to the end user, since we can only hear from 20 to 20 (ish). For myself, I have had my ears tested and can only hear up to (and including) 16khz.

  Sorry for longwinded ramblings>>> Just bored, and was just typing away.. heh

ForSerious

No need to apologize.
I, as well, can only hear the same range, but after doing some ABX tests with 96/24 vs 44/16, what I noticed is that the 96/24 had more clear bass notes. That made me think that maybe it would be worth it to upgrade. Maybe the inaudible sounds affect what I can hear?
I do not have the kind of income that would be able to support high quality music enjoyment. All the gear plus only being able to buy whole albums, would take over forty years.
Anyway, I've come to the conclusion that the difference I heard was my computer switching sound modes. I tried the same pre-recorded ABX testing on my X1 and didn't notice any such difference.

For some reason I have this drive in the back of my mind that wants to know the truth about recorded music quality. I think it may have come from being able to get lost in the music at one point in time, than having a hard time doing it again. It's almost as if my ears learn to pay attention to more and more small details, and so over time, in most low bit rate mp3s—the same ones that I used to love and get lost in—I get annoyed that I can't make out the details clearly, or that they are distorted. I now seem to have a hunger to hear things more clearly that sometimes even attacks CDs that I own.

So if I understand what you are saying, some recordings are just hopeless, And from what I've been reading lately, buying a 96/24 version would be buying the same version as the CD, just it's had a filter thrown on it to fill the upper frequencies.

samandhi

ForSerious replied at 2015-10-4 02:42
No need to apologize.
I, as well, can only hear the same range, but after doing some ABX tests with  ...

  I'm with you there regarding income for upgrading all my music to more than 16/44. I find it very funny that a site like HDTracks sells 16/44 for more than you can go out and buy a cd for. That is just how it is for purchasing online music. Google Play Music is known for streaming "High" quality (higher than most), but if you pay attention to it, it is actually only 320 .mp3s. The new streaming site that Niel Young (and many other artists) started, also has 16/44 and higher, but it is ridiculously expensive by comparison. I believe it is all about marketing that makes you believe (at first) that small .mp3s are good, then they come back and say that you really need to get high def music for really expensive, but it is actually just normal 16/44 redbook; because they have normalized poor quality mp3s online for so long, that it seems like a really good deal.

  The funny thing is that I can buy brand new cds from Walmart for $4.99(ish). So, why the hell would I pay $20.00 for 16/44 on a site like HDTracks. They know what they are doing.

  I also agree with you about "listening" to music. I used to listen to my music 96kbs mp3, using an old iPod Photo, and some Skullcandy Hesh headphones. I didn't really know the difference. But... When I purchased my Sennheiser HD700s (and joined Head-Fi, and learned how to listen), I couldn't abide by those crappy recordings, through a crappy DAP, playing to some crappy headphones. Again, I STILL believe that one cannot really hear a difference in the "so-called" high def music, but there is a point that you CAN (like 96kbs mp3's, or any mp3 from yesteryear, because the algorithym has gotten so much better).

  I DO believe that there are some recordings that make no difference if you are listening through VERY high end headphones, DAP, DAC, AMP, etc.... There is a great site that has some good information on bands and how dynamic their music is recorded. I believe this makes all the difference in the world, regardless of how the end file is formatted (high def, or lower quality). If you are interested in checking it out, here is the site that shows dynamics of the music by band.

  Obviously, the higher the number the more dynamic it is. The lower the number the more narrow it is (and this is where I would argue you can/will hear the difference). It's almost like taking a 96kbs mp3 and re-encoding it to a flac. While it IS a lossless file, you have taken a poor copy and turned it into something else. I guess they say CICO (crap in, crap out)....

inversesandwich

No need to apologize.
I, as well, can only hear the same range, but after doing some ABX tests with 96/24 vs 44/16, what I noticed is that the 96/24 had more clear bass notes. That made me think that maybe it would be worth it to upgrade. Maybe the inaudible sounds affect what I can hear?

That doesn't even make any sense, higher sample rates are only going to affect frequencies above 22.05kHz. As for bitdepth, even with a 24bit recording the DAC will only be able to resove it to 17-18 bits, unless someone invents a room-temperature superconductor.

samandhi

inversesandwich replied at 2015-10-6 00:34
That doesn't even make any sense, higher sample rates are only going to affect frequencies above 22 ...

  Hmmmm! Could be considered splitting hairs, but it seems as though they have in fact done that very thing (if one can trust that the article itself is not a total lie). The splitting hairs is due to this statement, that talks about for how long it becomes a room-temperature superconductor: “It was only a few millionths of a millisecond,”

The Article.

ForSerious

I've heard that in a 16 bit recording about 8-10 bits are used for what you hear, about 2 for what you aren't supposed to hear, and the rest is just padding.
24 bit just adds more nothing to the padding.

samandhi

ForSerious replied at 2015-10-6 05:30
I've heard that in a 16 bit recording about 8-10 bits are used for what you hear, about 2 for what y ...

  3 bits are the noise floor. If good techniques are used 16 bit can sound better than 24 bit. It is not the bit depth that makes a recording sound better, though you have more room for error with 24 bit. Here is a quote from an article I was reading:

But! Its not that 24 bits of data makes the sound better. It actually does not. What is does is give your audio more room to breathe in the numeric realm of digital audio. Remember, we are talking about numbers, calculations, not analog waveforms. With 24 bits of data demarcing your recording medium, its is possible to record extremely dynamic music, with very quiet soft passages and extraordinary loud passages. Quiet passages will be less likely struggling to stay above the noise floor on your system. One can record with no compression. You can record at lower levels, with more headroom. This ensures that the occasional peak is not truncated at the top and it will give converters some room the breathe. Because you are not pushing the limits of your bandwidth, your instruments will sound clearer, and the vocals may sound "cleaner", the song will mix better and there will be less noise. So its not that 24 bit recordings sound better. In fact they may sound just as bad or worse than 16 bit. But 24 bits gives the recordist a noise floor and headroom to create an excellent recording. Its a tool, and in the right hand, it can blow you away, audio wise.

  And, here is a GREAT article that explains bits in the digital recording/playback world.

ForSerious

I've done some music editing in my time, and what I've noticed is that when I mix two sounds together, they look louder in the editor, but do not sound louder. If I don't first set it to 24 bit, and sometimes, turn the over all volume down, it will clip at the mixed section. Of course I dither the final product back to 16 bit. I totally understand the need for 24 bit depth, but yeah: Useless in playback.

When I first got Imagine Dragons' Night Visions album, I was so sure that they had the recording volume up too high, and that was why it clips so much. I did the same thing when I was about 15 doing some of my own recording from my favorite computer game soundtracks. What an incompetent thing to do!

I've learned now that the clipping on that particular CD comes from the most extreme dynamic range compression on the market today, and I've added my voice to a re-release petition. Not sure if we'll ever get it, but we can hope.

inversesandwich

As said, 24bit is only useful in the digital domain, it gives you 144dB of dynamic range and makes tracks much easier to master.

samandhi

ForSerious replied at 2015-10-7 02:28
I've done some music editing in my time, and what I've noticed is that when I mix two sounds togethe ...

  Also, Death Magnetic from Metallica was so overdone that it clipped frequently. I remember reading that it was done on purpose, and that the group (at least Lars) thinks it makes it sound better. Fortunately for those that really did like the album, the Guitar Hero version (for some reason) was NOT mastered this way.

  There have been many "remasters" (if one can truly call them that) done from these GH tracks, but there are really 2 that have done such a fantastic job that it could be passed off as professional, and/or the original (or at the very least what the album SHOULD have been). I have the DECEIFER remix version, and let me tell you it totally is what Metallica should have released to begin with.

  There is some discussion about it here but if you do a search for one of those 2, you will find it, and can listen and compare for yourselves.

  BTW, here is an article of Lars being an as*hole (as usual) when questioned about it. You know, the standard Lars?! heh

ForSerious

I wonder if all artists get to choose if such things happen to their recordings, or if sometimes the engineer just takes the liberty of doing it and calling it professionalism. I hope we haven't gotten to that point.

samandhi

ForSerious replied at 2015-10-9 02:12
I wonder if all artists get to choose if such things happen to their recordings, or if sometimes the ...

  That is a good question. One that I would also like to know the answer to. I think that, more and more, the artists have less and less knowledge on anything related to music (other than playing it, which I would debate that might also be true). With the advent of auto-tune, and now being able to even use it on the fly at a concert, I think the talent pool for artists has gone downhill IMO (on a general scale, though there ARE exceptions).