|
|
|
I've found that SA-CD.net uses a very user-friendly icon system to categorize titles. If it's a hybrid SACD, it shows the hybrid symbol. If it doesn't show it it's assumed to be an SACD only title. If it's a multi-channel (surround) title, it will show a little symbol for that, too.
What does the DSD symbol mean? Aren't ALL SACD's mastered using DSD? Perhaps only some of them are?
Thanks.
|
|
|
Post by LC October 19, 2004 (2 of 11)
|
|
jonathanlucas said:
What does the DSD symbol mean? Aren't ALL SACD's mastered using DSD? Perhaps only some of them are?
Welcome to SA-CD.net. I'm sure you'll find it a useful site.
All SA-CDs are produced with DSD technology, but many (well, most) were recorded in analgue or PCM digital and then transferred to DSD to be encoded on the playback medium, the SA-CD itself. The DSD symbol denotes SA-CDs whose recorded content was originally captured with DSD (sort of like they used to mark CDs "DDD"). Theoretically, these should sound the best, but there are enough ways to screw up a recording that it is only a very rough guide to how something will sound. Mostly, it gives an idea of the state of technology in the recording industry.
|
|
|
|
|
|
LC said:
All SA-CDs are produced with DSD technology, but many (well, most) were recorded in analgue or PCM digital and then transferred to DSD to be encoded on the playback medium, the SA-CD itself. The DSD symbol denotes SA-CDs whose recorded content was originally captured with DSD (sort of like they used to mark CDs "DDD"). Theoretically, these should sound the best, but there are enough ways to screw up a recording that it is only a very rough guide to how something will sound. Mostly, it gives an idea of the state of technology in the recording industry.
Let me see if I understand you correctly. The titles that are NOT marked with "DSD" involve taking an original recording which was done in analog or PCM digital (which is the process that most CD's are mastered in???) and simply re-encode it so it will play on an SACD.
My question is does that even improve the quality of the music? It seems like that would be similar to taking a bunch of mp3 files and re-encoding them to burn on to a CD which we all know is well and good but it doesn't then make those files CD quality.
According to what you said, however, the original recording format may not matter any way. Perhaps it comes down to WHO is actually doing the producing as far as skill and financial resources goes. Would you agree?
Thanks.
|
|
|
Post by todd October 20, 2004 (4 of 11)
|
|
jonathanlucas said:
Let me see if I understand you correctly. The titles that are NOT marked with "DSD" involve taking an original recording which was done in analog or PCM digital (which is the process that most CD's are mastered in???) and simply re-encode it so it will play on an SACD.
My question is does that even improve the quality of the music? It seems like that would be similar to taking a bunch of mp3 files and re-encoding them to burn on to a CD which we all know is well and good but it doesn't then make those files CD quality.
According to what you said, however, the original recording format may not matter any way. Perhaps it comes down to WHO is actually doing the producing as far as skill and financial resources goes. Would you agree?
Thanks.
Your analogy is faulty. Analog and PCM masters are both considered high resolution (as is DSD). Regular CDs are never as high a resolution as the original masters. CDs are 16bits/44K, whereas PCM sources are likely 24bit/96k or higher. MP3 is a lossy-compressed format and has nothing in common with the formats used for recording. In some sense, analog is the highest resolution possible format.
So converting from analog or PCM to DSD, does _not_ improve the quality, but rather _retains_ the quality.
All things being equal, an SACD produced from a quality source is higher resolution that the CD made from the same source.
However you are right about the production values having a big impact on the quality of the master. Sometimes the CD and the SACD sound the same, and sometimes the SACD mix is poor.
|
|
|
Post by LC October 20, 2004 (5 of 11)
|
|
jonathanlucas said:
Let me see if I understand you correctly.
I basically agree with Todd. SA-CD has two basic advantages (over CD): the sheer amount of data stored, and the sonic advantages of DSD playback. SA-CD will store all of the data from a 24/96 recording and virtually all of the data from a high quality analogue recording. And then you have the advantages of DSD playback, which should offer some improvement even if the source is of low resolution. Tranferring from one format to another may not be elegant, but it isn't necessarily bad. The audiophile label Mobile Fidelity is quite happy recording in analogue and using those masters for CD, LP, and SA-CD. There's less of a case for PCM if you know it will end up in DSD, but I have some excellent sounding 24/96-to-DSD recordings. Although I would expect a label that is serious about SA-CD to be recording in DSD, I'm more concerned with the the overall level of care the recording and production engineers take.
|
|
|
Post by Dan Popp October 21, 2004 (6 of 11)
|
|
Jonathan,
This is a somewhat thorny issue, with over-generalizations a little too easy to make, I'm afraid.
I don't agree with some of the specific comments made above, but I think the overall thrust of the posts is correct. The bottom line for me is that a DSD master will give you one thing on SACD that none of the others can, and that's NO CONVERSION.
If you have an analog master, you have the requisite analog artifacts plus the DSD artifacts on playback plus whatever nasties the conversion process itself introduces. If you have a PCM master, likewise. If you have a DSD master, you have only the artifacts of that one process and no conversion anomalies.
"In theory" that's your winner. One of my favorite aphorisms is, "The difference between theory and reality is that, in theory there is no difference between theory and reality, and in reality there is."
|
|
|
Post by LC October 22, 2004 (7 of 11)
|
|
Dan Popp said:
I don't agree with some of the specific comments made above, but I think the overall thrust of the posts is correct.
I assume you disagree with my statement that PCM or analogue to DSD conversion isn't "necessarily" bad. That was a bit sloppy. All I meant is that the final result will not necessarily be bad or disappointing, and that it might well be better than some, or for all I know, most pure DSD SA-CDs. But I think you have a perfectly good theoretical claim here that needn't be weakened by disparaging "theory." Your claim is that, *other things being equal*, pure DSD is best. I think that pretty well has to be right, unless the term "best" includes some weird subjective preferences for the recording and conversion artifacts you cite.
|
|
|
Post by Dan Popp October 22, 2004 (8 of 11)
|
|
LC said:
I assume you disagree with my statement that PCM or analogue to DSD conversion isn't "necessarily" bad. That was a bit sloppy. All I meant is that the final result will not necessarily be bad or disappointing,
LC,
No, I think there are some cases in which conversion to DSD could subjectively improve the PCM masters slightly - but here I am into "upsampling" and other areas where my knowledge is very scant indeed. I am only speculating that, in 20 years, engineers have found better ways to interpolate the information "between the numbers" of PCM.
I would rather quibble with your statement that DSD can preserve "all" of the information of a PCM hi-res master. That's true in the frequency domain (time), but not in the amplitude domain, as I understand the process.
While the sampling error of DSD may be very small, I think it has to be just less than the value of the Least Significant Bit (which is the only bit in DSD). Any sampling (or re-sampling) is an approximation - a "rounding" if you will - of the real value to the closest value that can be represented. The real value is somewhere between +/- "half a bit."
Therefore some loss must occur in PCM-to-DSD conversion. If (I speculate further) this loss manifested itself as some smearing or smoothing of the digital representation of the waveform, it is a safe bet that many people would declare this to be an improvement, although in absolute mathematical terms it is less accurate.
|
|
|
|
|
|
Dan Popp said:
Jonathan,
This is a somewhat thorny issue, with over-generalizations a little too easy to make, I'm afraid.
I don't agree with some of the specific comments made above, but I think the overall thrust of the posts is correct. The bottom line for me is that a DSD master will give you one thing on SACD that none of the others can, and that's NO CONVERSION.
If you have an analog master, you have the requisite analog artifacts plus the DSD artifacts on playback plus whatever nasties the conversion process itself introduces. If you have a PCM master, likewise. If you have a DSD master, you have only the artifacts of that one process and no conversion anomalies.
"In theory" that's your winner. One of my favorite aphorisms is, "The difference between theory and reality is that, in theory there is no difference between theory and reality, and in reality there is."
Well put and I love that quote! Thanks for the input.
|
|
|
Post by LC October 24, 2004 (10 of 11)
|
|
Dan Popp said:
I would rather quibble with your statement that DSD can preserve "all" of the information of a PCM hi-res master. That's true in the frequency domain (time), but not in the amplitude domain, as I understand the process.
While the sampling error of DSD may be very small, I think it has to be just less than the value of the Least Significant Bit (which is the only bit in DSD). Any sampling (or re-sampling) is an approximation - a "rounding" if you will - of the real value to the closest value that can be represented. The real value is somewhere between +/- "half a bit."
Therefore some loss must occur in PCM-to-DSD conversion.
That's interesting. I think I follow this, although I couldn't really judge whether your understanding is correct/complete. I guess there is no reason why the fact that the total number of bits encoded with DSD is greater than it is with 24/96 should mean that every bit of the latter is represented in the conversion. I have a friend who recently graduated from computer and electrical engineering. I'll try to bug him about this.
|
|
