Huge Controversy Within the Hi-Rez Community

[watchnerd]

   
If you're saying that an up-converted file, 16/44 to 24/96 will sound exactly the same as the 16/44 file, that's the heart of the controversy. 
 

 
No, that's not what I'm saying.
 
I'm saying 16bit vs 24bit doesn't matter if you're not doing any mastering in the digital domain.
 
Think of it like a vinyl rip.  If I do a straight rip, no editing, it doesn't matter if it's 16bit vs 24bit.  Both have enough dynamic range to cover vinyl's needs and if I I'm not doing digital editing I don't need extra bits to throw away.
 
 

 
In your opinion, since you do this on a weekly basis, will one of your finished HiRez 24/96 masters sound exactly the same as your finished Red Book master?
 

 
If they don't, I screwed something up!
 
Seriously, though, I don't even make the 16/44 version in my DAW.  I save a 24/96 master copy.
 
By default, I make I provide a 24/96 master and a LAME V0 MP3 to the orgs I volunteer for.
 
I'll make a 16/44 version in SoX, not in the DAW, if someone asks for it.  But most of the users seem to either want the high res version or the MP3.  The boring middle ground is less popular.
 
 

 
How do they look on a spectrograph?  Do they null each other out?

 
In the audible range, yes, they null.
 
If they don't, it means something got goofed in the down conversion / down sampling, i.e. I effed up the dither setting, usually.
 

   
If you're saying that an up-converted file, 16/44 to 24/96 will sound exactly the same as the 16/44 file, that's the heart of the controversy. 

 
[That isn't really the heart of the controversy...the heart of the controversy is whether or not recordings made in 24bits sound better than recordings made in 16bits that are upsampled to 24bits.]
 
That being said...
 
Up-converting from 16/44 to 24/96 has to sound the same (assuming your DAC isn't broken doesn't have messed up filters) as the original source unless you believe in magic.  You can't magically create new data that wasn't in the original sample.
 
Down-converting/sampling  is a different story.  You can destroy things in a bad down-sample.
 
Prior to SoX, down conversions that weren't exact sample rate integers could be risky, i.e. 24bit/96khz to 16bit/48khz was safe, but 16bit/44.1khz was riskier.  So if the down-conversion is using some old-ass software, it could be flawed.
 
With recent versions of SoX, this isn't an issue.
 
Also, bad dithering choices, or no dither at all, can jack things up when reducing bit depth.

 

[RRod]

   
Losing 4 bits to editing is pretty easy to do.  And, in the case of highly produced studio albums, pretty much unavoidable once you get done combining a bunch of tracks and leveling them.  
 
AAD CDs should be full 16 bits because all the mastering was done in analog so there shouldn't be any bit loss.

 
Actually I meant some of the early DDD CDs, though I'm a classical guy mainly so I assume there's fewer big effects going on to throw bits out of the mix.

 

[watchnerd]

   
Actually I meant some of the early DDD CD, though I'm a classical guy mainly so I assume there's fewer big effects going on to throw bits out of the mix.

 
I listen to a lot classical, too.  Classical can be just as bad as an anything in chewing up bits if they used a lot of spot mics and used digital volume control to mix it.
 
The Telarc stuff from that era was probably the benchmark for large scale DDD classical.  I like that they had liner notes on all the equipment used in each recording.  They were one of the first to move to high-resolution ADCs.

 

[Deftone]

there should be some anti con software that lets you run a music file through it to tell you if its really 24bit and not upsampled.

 

[watchnerd]

  there should be some anti con software that lets you run a music file through it to tell you if its really 24bit and not upsampled.

 
Part of the controversy is that HDTracks is using Bitter to attempt to do this.
 
You can download it for free to play with yourself.
 
Bu it's far from 100% foolproof.

 

[Deftone]

thanks for the link, if i was a high res believer i would need something like this for piece of mind.

 

[spruce music]

You can run tracks thru Audacity which is free.  Once you learn a few tricks it is pretty easy to see if something was upsampled.  Spectrograms and FFT's are the main useful items.  If you see a sharp filter cut off at 22.05 khz with nothing higher it quite likely was Redbook at some point.  Unhappy customers doing this is what prompted HDT to start their QC program of using Bitter and checking for response cut offs. Which is what tripped up Bob Katz.  He indeed upsampled some redbook content.  While his upsampling may or may not have improved the end result vs staying at 44.1 khz, it still was not actually hirez.  That is what HDT caught and called him on.

 

[Roly1650]

You can run tracks thru Audacity which is free.  Once you learn a few tricks it is pretty easy to see if something was upsampled.  Spectrograms and FFT's are the main useful items.  If you see a sharp filter cut off at 22.05 khz with nothing higher it quite likely was Redbook at some point.  Unhappy customers doing this is what prompted HDT to start their QC program of using Bitter and checking for response cut offs. Which is what tripped up Bob Katz.  He indeed upsampled some redbook content.  While his upsampling may or may not have improved the end result vs staying at 44.1 khz, it still was not actually hirez.  That is what HDT caught and called him on.

All that's true, but what HDTracks hasn't done is go back and re-classify any of their product which is labelled hi-res, but isn't, so their QC system doesn't apply retroactively. This means a potential user has to go hunting around on other websites to find out if the album they are interested in purchasing from HDT is the real deal or not. A mucking fuddle in my book and sufficient inducement for me to keep my money in my pocket, even if it was clear that hi-res was audibly superior to RBCD.

If HDT really wanted to maintain credibility and transparency, they would either a) issue a refund for or b) offer a free download of the right files, (if they have them), to those customers they've sold RBCD to as hi-res. And HDT know exactly who bought what and when, so it's not a major excercise, they have every purchasers email address on file. What they're doing now is basically disingenious and dishonest, QC system or no QC system.

 

[castleofargh]

I feel like the only reason those guys aren't sued all day long and shut down until they do it right, is that the law must stop at some ultra basic definition of highres. like, if the resolution of the file is highres then it's highres. or more logical maybe, if it was recorded/processed/sold at more than 16/44, then it's highres. so with that brilliant system, 16/48 is highres and 24/44 is highres. therefore they are not lying when they sell a 24/44 upsampled to 24/192 as being highres.
 
there must be something like this going on, for everybody to knowingly do what they do and often not even taking out the files when we tell them. they must know they risk nothing legally.

 

[watchnerd]

  I feel like the only reason those guys aren't sued all day long and shut down until they do it right, is that the law must stop at some ultra basic definition of highres.

 
There is no statute law definition of high resolution audio.
 
Heck, until recently there wasn't even a consensus definition of 'labeling standards' for high res.  And even though that standard has been proposed, I've never seen it used.

 

[icebear]

ROFL, there isn't even agreement about labeling food correctly. GMO salmon will be on the shelves at your trusted fish monger and no need to be labelled as such.
There are more important issues than customers who can't decide if something sounds better and is worth paying more money for than RBCD. They pay for numbers or LED's ligting up on their DAC's and feel cheated if some analysis tells them it can't sound all that good because it was only up-sampled.
 
OK, I completely agree that is cheating but only of folks who can't listen with their ears and make an honest judgement for themselves.
Infected by marketing BS they wanted to believe highrez is the only true audiophile nirvana.
Sorry folks you have been had

 

[watchnerd]

Here are the proposed definitions, taken from the article "Music Industry Titans Unite to Standardize High Resolution Audio."
 
 
 

1. MQ-P: From a PCM master source that is 48kHz/20 bit or higher (typically 96/24 or 192/24 content).
2.  

1. MQ-A: From an analog master source.
2.  

1. MQ-C: From a CD master source (44.1 kHz/16 bit content) in which legacy CD masters are employed as source material and upsampled to high resolution.
2.  

1. MQ-D: From a DSD/DSF master source (typically 2.8 or 5.6 MHz content).

 
"MQ-C" is the silly one at the heart of controversy.

 

[Roly1650]

ROFL, there isn't even agreement about labeling food correctly. GMO salmon will be on the shelves at your trusted fish monger and no need to be labelled as such.
There are more important issues than customers who can't decide if something sounds better and is worth paying more money for than RBCD. They pay for numbers or LED's ligting up on their DAC's and feel cheated if some analysis tells them it can't sound all that good because it was only up-sampled.

OK, I completely agree that is cheating but only of folks who can't listen with their ears and make an honest judgement for themselves.
Infected by marketing BS they wanted to believe highrez is the only true audiophile nirvana.
Sorry folks you have been had

But it's not a case of not being labelled, which is one thing, this is a case of being deliberately mislabelled, which is something else entirely. And why does it matter they find out after the sale that what they paid for isn't what they got? That makes it worse, not better, doesn't it? It's totally besides the point if a difference is or isn't audible. However it gets sliced and diced, the consumer paid for something they didn't get, through deliberate misrepresentation, dictionaries define that as fraud. I'm not sure why anybody would think that fraud is ok.

 

[RRod]

  MQ-P: From a PCM master source that is 48kHz/20 bit or higher (typically 96/24 or 192/24 content).

1.  

 
Are there no PCM master sources >= 20-bit at 44.1kHz?

 

[watchnerd]

   
Are there no PCM master sources >= 20-bit at 44.1kHz?

 
In theory, there could be, but all the old 20bit ADCs that I'm aware of were operating at 48khz.  I don't know of any 20bit ADCs that are natively 44.1khz.
 
Plus I guess they're also saying that even if there were, they aren't high resolution.