|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
50.164.204.218
In Reply to: RE: You're thinking of digital in analogue terms posted by Chris from Lafayette on February 09, 2015 at 11:57:27
Sometimes, more resolution is just redundant resolution, as when, say, 16/44.1 files are placed in 24/192 (or, for that matter, 2 X DSD) containers, and then some listeners pretend that the resulting files are hi-rez.
I know there were a fair number of SACDs and DVD-As made from 48k PCM masters, and a few 44.1k (e.g. Nora Jones, JT, Donald Fagan). But as far as I know there is no way to tell whether the PCM master was 16-bit or 24-bit from a spectrogram.
So I'm not sure the accusation of Redbook upsampling can be supported, even if it might be true in some cases. Mark Waldrep has been spreading FUD about it, but as far as I know he hasn't any evidence. Do you know of any that I've missed?
To my ears, a SACD sourced from a 24/48k PCM recording can still offer a nice improvement over the CD. Just last night I was listening to Alison Krauss Forget About It, a 24/48k recording. The SACD is much nicer sounding than the CD, even after converting it from DSD to 24/88.2k for playback in my current system. I would love to have the original 24/48k for comparison.
Follow Ups:
Yes - I agree that you can't tell the bit depth from a spectrograph. But spectrographs are still useful to see actual frequency response, as well as to check whether the uncorrelated noise dusting (typical of the DSD files used on SACD's) is present.
I think Mark Waldrep's point is that for a recording to be considered hi-rez, every stage along the production process should be hi-rez, from microphones to finished files. He just had a column a couple of days ago where he praised other companies for their commitment to this hi-rez production methodology, including 2L, Naxos, Linn, Chesky, Pentatone, Channel Classics (as well as, of course, his own AIX label). Over on one of the other discussion boards here, Archimago linked to some tests he had done with some PONO downloads comparing them to other source files in terms of dynamic range and frequency extension. His PONO downloads, although touted as hi-rez, were sometimes more compressed (in both dynamics and frequency) than files ripped from CD's. Other posters have had better luck with PONO downloads - we need to get info on a critical mass of these downloads before we can say whether the majority of their hi-rez downloads are actually hi-rez. Basically, I think Waldrep is just worried that the term hi-rez audio is going to get diluted and cheapened.
I also agree with you that "a SACD sourced from a 24/48k PCM recording can still offer a nice improvement over the CD" - no question about it. Still, I'd rather avoid that last conversion (or transcoding or whatever it is) to DSD, so I look for the availability of hi-rez PCM downloads for recordings mastered in PCM. (It would be the same with DSD: I would want my download to stay in the same format as the master.) In any case, I've got plenty of SACD's derived from PCM masters, and I still enjoy them very much (despite my general principles!).
Part of the "disappointment" with 16/44 CDs or files compared to DSD or higher bit depths/sample rate PCM can be attributed to the sample rate (down)conversion, dithering and noiseshaping choices made by the mastering engineer. There are many ways to get to the final result, and poor algorithms or inappropriate choices in filter steepness and phase characteristics lead to rather nasty artefacts which show up as "harshness" and all the bad things one reads constantly about CD audio.
When upsampling or format converting, you may get a more subjectively pleasing result, but it can't undo the original damage (including embedded jitter from the ADC stage).
On my reference system, I can produce 16/44 files that are virtually indistinguishable from the high resolution masters (which are indistinguishable from the source). Close A/B comparisons can show up the differences, but played in isolation they are sufficiently close that guess work would be required unless one was intimately familiar with the source to know what to listen for.
Regards Anthony
"Beauty is Truth, Truth Beauty.." Keats
I have often wonders about whether it wound not be better to take a higher bit rate master and rather than crunch the numbers back to redbook just do a playback transfer through a great DA and then into another great AD to 16/44.1 and compare and use the one that sounds better.
The decimation programs may be better now, but I thought that there we compromises if the higher bit rate was not a multiple of 44.1. so it may be that if the tracking masters were 2496 if might be better to do it manually? Just curious. I would be through a DCS stack it would be pretty special.
Jim Tavegia
If the high resolution master is an integer multiple of the 44.1kHz, then it is a trivial process to convert to 44.1kHz by dropping the unwanted samples (i.e the inverse of oversampling). The problem isn't so much that the bandwidth is limited, but the cumulative effect of filter characteristics used throughout the recording-playback chain as well as jitter induced "distortion". The additional problem is introduced in the dithering process when the bit depth is reduced.
My own experiments show that a high resolution master is demonstrably better than 16/44. Of that, in my experience, there is no question. It is the equivalent viewing an upscaled DVD image then compaing to a true 1080p image. On superficial examination at the "normal" viewing distance, there is little to choose between them. However, move closer to the screen and the differences become more obvious. The same applies to noise-shaping and dithering - modern algorithms are based on a psycho-acoustic model (such as iZotope MBIT+) and these are optimised based on expected playback volume. If the playback volume deviates significantly from the expected level, then differences may be noted by the listener. In other words, the mastering engineer needs to understand the requirements of the listener with respect to the musical content. Classical piano is highly revealing of dithering and noise-shaping "errors" especially if the piece has a lot of pianissimo sections. If the noise shaping is too aggressive, then spurious artefacts may be heard on very low amplitude level notes in the midband. I have many commercial CDs that exhibited this and I originally thought there was something wrong with my playback system! It wasn't until I experimented in my own recordings that I realised that I could induce the same "distortions" on playback as well as eliminate them depending on the noise-shaping and dither amplitude chosen.
My point was really to highlight that early CDs were disappointing largely due to the technical limitations of the A/D equipment in terms of filter characteristics and timing precision as well as poor linearity of the DACs and reconstruction filter implementation, combined with simpler dithering. More advanced dithering schemes combined with noise-shaping characteristics based on psycho-acoustic models now provide a much better result, but are still no substitute for the high resolution master if ultimate quality is desired.
Regards Anthony
"Beauty is Truth, Truth Beauty.." Keats
I agree with Ted, your DAC most likely has a sweet spot. My Hilo likes PCM at 24:192 and my Lampi likes DSD128. Download enough different samples of the same thing, or play with up converting software like HQPlayer and find your DAC's preferred format. Then download that format or upsample/downsample to it.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: