Article: And The Winner Is …

[bottlerocket]

And winner is DR9. Sad.

[JeffH]

This is exactly why I still buy CD's for most of the music I buy. I realize there are some great sounding high resolution downloads out there. But, given the current pricing and the lack of any way to really know what you are buying before you buy, I just can't pay the high prices for the downloads. Besides, I still like possessing the music in cd form. Old school I know.

[peter73]

The "Vinyl Exp" mastering has at least 3.81db (5 db with minuscule compression of few peaks) of unused dynamic range. Why is that? Did they just wanted to make it quiet?

[Skeptic]

I love the concept behind the DR DB but it's too bad the algo is so flawed.

What is so flawed about the "DR" algorithm? Other than the name of course - it's a measurement of peak-to-loudness ratio, not dynamic range.

 

This is exactly why I still buy CD's for most of the music I buy. I realize there are some great sounding high resolution downloads out there. But, given the current pricing and the lack of any way to really know what you are buying before you buy, I just can't pay the high prices for the downloads. Besides, I still like possessing the music in cd form. Old school I know.

This is one of the reasons I continue to buy CDs as well. Buying a CD gives me a good quality 16/44 file, it is often cheaper than even a 16/44 lossless download, it gives me a physical item that I can resell if I don't like the album, and it means that I have a physical "backup" if something happens to my digital files.

 

It is rare that the high resolution downloads are so much better that I would rather pay 2-3x the cost of the CD, than buy another album or two with the same money.

 

The "Vinyl Exp" mastering has at least 3.81db (5 db with minuscule compression of few peaks) of unused dynamic range. Why is that? Did they just wanted to make it quiet?

16-bit has around 96dB of potential dynamic range.

Looking at my CD, the track with the highest "Dynamic Range (R128)" value in JRiver has a dynamic range of just under 12 LU.

It may be somewhat higher with the "Vinyl Exp" master but that still leaves 84dB of dynamic range - this is the reason why many people say that 24-bit is completely unnecessary.

 

You don't want the peaks to be pushed right up against 0dB because it can introduce inter-sample clipping on playback - the reconstructed waveform peaks may be higher than the digital samples and clip if there is not sufficient headroom.

[RankStranger]

How does the vinyl sound?

[InfernoSTi]

What is so flawed about the "DR" algorithm? Other than the name of course - it's a measurement of peak-to-loudness ratio, not dynamic range.

 

I suspect that it is this issue that is causing some to think the measuring tool is imperfect: Why the TT Meter doesn't work on vinyl - YouTube

[david.parker83]

The amount of clipping in the Amazon/MP3 Direct versions is gross. I ended up going with the MFIT version for now (on the road, iPhone listening). Since they bump the db of each track down a little bit to avoid clipping, my guess is the CD version will have the same DR as MFIT version but be a tad louder on some tracks.

[InfernoSTi]

OT: why aren't all the comments tracking/showing up in the same place? There is this part of the "thread" and there is another part of the "thread" that has some of the same posts and then they diverge…

 

Computer Audiophile - And The Winner Is …

 

http://www.computeraudiophile.com/f3-article-comments/article-and-winner-%85-19620/index2.html

 

Sorry to really not understand this…and wondering where to actually post.

 

John

[phatoldsun]

How does the vinyl sound?

 

The vinyl sounds awesome and I put at the top of the list Chris has in this article. 180gm.

[RankStranger]

The vinyl sounds awesome and I put at the top of the list Chris has in this article. 180gm.

 

Thanks. I'll look out for that one.

[Phil C]

Maybe I missed this in the article or comments but does anyone know the dynamic range (DR) score for the Morning Phase CD ripped losslessly (AIFF for example) to a computer? I typically buy a CD and rip the music to my laptop and J River so I'm curious. Thanks much.

 

Win 7 laptop/JRiver18 > Belkin Gold USB cable > Berkeley Audio Design Alpha USB> DH Labs D-75 digital coaxial cable > Benchmark DAC1 HDR > Cardas Golden Presence balanced interconnects > BSG Technologies qol Signal Completion Stage > Cardas Golden Presence balanced interconnects > Music Reference RM-200 Mk II tube amp > Cardas Golden Presence spkr cables > Sonus faber Cremona M spkrs

 

Phil C

[phatoldsun]

Maybe I missed this in the article or comments but does anyone know the dynamic range (DR) score for the Morning Phase CD ripped losslessly (AIFF for example) to a computer? I typically buy a CD and rip the music to my laptop and J River so I'm curious. Thanks much.

 

Phil C

 

I'm pretty sure the HD version is the same as CD. I actually suggest buying the CD instead of the HD version to save money. If you have a turntable, then buy the vinyl LP and get the best of both worlds.

Album details - Dynamic Range Database

[apple2k]

I am literally dumbfounded - Where is the outrage at HD tracks?

 

Blame Beck? Really? I blame HD Tracks. My CD of this album sounds as good if not better than the HD download, it was not advertised as 'HD' 'HiRes' '24bit' and cost 9.99 not 17.98

 

I bought the Flaming Lips "Yoshimi Battles the Robots' from HD Tracks which was a 5.1 mix ported to stereo, completely unlistenable, I complained to HD Tracks and they told me nothing was off with the file. The album is no longer on their site. If you bought a BMW and it had the wrong engine in it, but BMW said it was the supplier's fault - 'it looked like a straight 6 cylinder but only had 4 pistons inside' - would you accept it? No. The catalog on HD Tracks is minimal compared to most, there is nothing impeding their ability to limit their sources to genuine HD (24 bit master) sources but their own will. Without a willingness & commitment to test their material - behavior that represents quality - they're an unchecked, unregulated and meaningless intermediary between the music business and you. Not a brand with integrity.

 

When will the audiophile community stand up for itself and demand transparency? I sit here in disbelief as I read earlier posts suggesting 'it's so hard for HD Tracks to know the source" ..... it took me 30 sec into the Flaming Lips download to know 1. my stereo was broken 2. something was seriously wrong with the file. It took me 2 minutes into Beck to know nothing is special about these files. "HD" is an unregulated format, there is nothing legally that defines it.

 

I read editorials from the audiophile print press about the coming HD revolution. Where is it? Where is an Airport express or Apple TV with the hardware to process 24 bit files? Same with Sonos. Those are the only two installed user bases that matter to the music industry. The 24bit audio revolution isn't coming. I wish this wasn't the case, but wishing isn't enough.

 

Linn, NAIM, Sound by B&O don't have these quality issues, because they're committed to it and cannot afford the consequences of being wrong. HD Tracks isn't in the same boat.

[junker]

The main issues are that it appears to only take 10000 samples for the entire track. From these 10k sample set it then only selects the loudest 20% of these samples to determine the RMS value. Is high-res material with many more samples getting penalized here? It's not the RMS of the track, but rather the RMS of the 20 loudest peaks. And then finally, it compares these 20 loudest passage RMS values to the peak (from the 10k samples? What if the true instantaneous peak are elsewhere?).

 

 

So my concern is that what if the true RMS value is much lower than that indicated by the so-called "RMS" value. Then the value it derives is much lower than it is in actuality. It would also penalize long tracks. Let's say I played some 20 minute Wagner track and in this track it crept along but had a very loud passage every minute or so. Well this loud passage every minute would be considered the RMS value despite the other 80% of the track being quiet.

 

Also, I also see are that the offline check is for 16/44.1 so how is it getting the DR from hi-res material. Is 24 bit material getting penalized here? Is it downsampling it?

 

To me it seems like an old algorithm were the writer was concerned about memory, computing power, and bandwidth constraints. It's appears to use fatally flawed heuristics in that it is is biased against long tracks, it is not designed with hi-res in mind, only takes 10k samples, and from this sub-set only derives 20 loud passage "RMS" peaks, compares this to a single peak, and spits out a single DR number.

 

Fantastic idea, but nothing like taking the RMS value of every sample in 24 bit space, and comparing this to the absolute peak.

 

I'd be very interested in what others think. Maybe I'm reading it wrong, but I think there are too many shortcuts that people aren't aware of. Jud made a point earlier that it still might be useful for comparing different formats of the same material as a self-contained set. I think I agree with him here, but I don't know if a DR6 in and of itself means anything.

 

What is so flawed about the "DR" algorithm? Other than the name of course - it's a measurement of peak-to-loudness ratio, not dynamic range.

[firedog]

Sorry, most of your post makes no sense. HDT (except for Chesky) isn't a manufacturer like the companies you mentioned, only a reseller. They sell the files the labels give them. If you think the file "sounds bad" it isn't their fault. "Sounds bad" is a matter of taste. If the file is actually defective they are responsible, and as far as I know they replace defective product or give credit towards another purchase. They have done it for me.

 

Unfortunately, there is no guarantee a 24 bit download will sound better than a CD, it could actually have inferior mastering and sound worse. That doesn't mean it isn't "hi-res". How many cases in the last year do you know of where they sold something as 24 bit that wasn't?

 

Most places I know, when you buy a CD, you get whatever is on the CD. If you don't like the sound, but the CD itself isn't defective, its just too bad for you. I have lots of bad sounding CDs in my collection. HDT is essentially the digital equivalent of a record store. Nothing more. If I order a CD online and don't like the SQ when I get it, I understand that it isn't worth it to pay to send it back and try to get a refund.

 

And yes, it is hard for HDT to know the source. Often the labels won't tell them, and in some cases even the labels don't really know, as their back catalog is a mess ("where are those tapes?").

 

Some of the other download sites only sell product they produce themselves, so the situation is different.

 

All of that said, I think HDT should give us whatever provenance info they have about every download. They also should allow user reviews, so we can better judge if a given download is worth buying.

 

 

I am literally dumbfounded - Where is the outrage at HD tracks?

 

Blame Beck? Really? I blame HD Tracks. My CD of this album sounds as good if not better than the HD download, it was not advertised as 'HD' 'HiRes' '24bit' and cost 9.99 not 17.98

 

I bought the Flaming Lips "Yoshimi Battles the Robots' from HD Tracks which was a 5.1 mix ported to stereo, completely unlistenable, I complained to HD Tracks and they told me nothing was off with the file. The album is no longer on their site. If you bought a BMW and it had the wrong engine in it, but BMW said it was the supplier's fault - 'it looked like a straight 6 cylinder but only had 4 pistons inside' - would you accept it? No. The catalog on HD Tracks is minimal compared to most, there is nothing impeding their ability to limit their sources to genuine HD (24 bit master) sources but their own will. Without a willingness & commitment to test their material - behavior that represents quality - they're an unchecked, unregulated and meaningless intermediary between the music business and you. Not a brand with integrity.

 

When will the audiophile community stand up for itself and demand transparency? I sit here in disbelief as I read earlier posts suggesting 'it's so hard for HD Tracks to know the source" ..... it took me 30 sec into the Flaming Lips download to know 1. my stereo was broken 2. something was seriously wrong with the file. It took me 2 minutes into Beck to know nothing is special about these files. "HD" is an unregulated format, there is nothing legally that defines it.

 

I read editorials from the audiophile print press about the coming HD revolution. Where is it? Where is an Airport express or Apple TV with the hardware to process 24 bit files? Same with Sonos. Those are the only two installed user bases that matter to the music industry. The 24bit audio revolution isn't coming. I wish this wasn't the case, but wishing isn't enough.

 

Linn, NAIM, Sound by B&O don't have these quality issues, because they're committed to it and cannot afford the consequences of being wrong. HD Tracks isn't in the same boat.

[junker]

I think the HDTracks version sounds good with lots of texture. Can't wait for the vinyl to arrive tomorrow to compare.

[Synfreak]

Junker,

please re-read the informations again, and think about it.

How would you calculate a RMS value out of "every sample"?

 

And, your quote:

The main issues are that it appears to only take 10000 samples for the entire track

is simply wrong.

 

The DR-meter is one tool to use, and uses a weighting algorythm to make things easier to read and compare for the everyday user.

If you are not convinced about the precision of it, just use another one - there are many out there ...

[4est]

In reading this, the clear winner is Beck (and probably more so his label). Just think of all of the extra copies they will be selling as others search for the best one. I love Beck's music, but this is a sad conundrum for me. I do not want to support this concept, and yet I want the music.

[bottlerocket]

I am guessing Beck is an audiophile and that he has a computer so the best forum for outrage may well be this one. I agree with Firedog and CA that HDTracks is just another CD store that happens to offer hi-res also so there is not much they can do except not offer the music at all. Chris, if you want an off the charts story for your website, get Beck's response to this! His response is the only thing missing from an otherwise great column.

[4est]

Beck may not even have a say in it.

I am guessing Beck is an audiophile and that he has a computer so the best forum for outrage may well be this one. I agree with Firedog and CA that HDTracks is just another CD store that happens to offer hi-res also so there is not much they can do except not offer the music at all. Chris, if you want an off the charts story for your website, get Beck's response to this! His response is the only thing missing from an otherwise great column.

[ChrisG]

An interesting post today at realhd-audio.com regarding the recording and mastering process. Exactly That The Artist Wanted You To Hear… | Real HD-Audio

[matt123]

Just listened to both the vinyl and MP3 versions once through. On the plus side, I like this album more with each listen, and the last track "Waking Light" is just awesome.

 

The Vinyl is actually louder than the mp3 through my system with the volume knob in the same place. I know that doesn't mean a whole lot with all of the differences between the two sources, but still, its very rare and it sums up my overall opinion.

 

On the vinyl, the bells, drums and acoustic guitar sound more real. There are nice echoes and reverb effects in the background. However, it just sounds loud as if the levels are pushed up in the mix. One some tracks, I like this better, on the more ambient ones - like "Cycle" and "Morning" at the start of the album, I prefer the MP3, but miss the more realistic sound on the vinyl.

 

The MP3 has a lot nicer sense of spaciousness, though the cymbals, drums and guitar sound...like an MP3 and are a lot less realistic. The echo/reverb in the mix is not as apparent.

 

I tend to think of my analog set up as spacious and airy - so this is generally not what I hear when comparing digital vs. analog versions of albums (recognizing its by no means apples vs. apples.) I keep comparing the vinyl of Morning Phase to the MOFI vinyl reissue of Sea Changes, which is absolutely fantastic. The one thing I keep comparing is the glockenspiel on both albums - the Sea Changes mix has it just right. Sounds like a bell (!) and has a wonderful sense of air around it. Maybe in 10 years, we'll get a MOFI reissue of Morning Phase to match :-( But, for now, I wish we could have the MP3 mix on vinyl or high res.

 

Makes me wonder even more which file/mix was used to cut the vinyl and where the heck the "EQ Vinyl Experience" file came from.

 

Analog source: VPI Scoutmaster > Soundsmith Zephyr cartridge > Sutherland Ph3d pre-amp

Digital: Vortex Box > Sonore Rendu > Naim DAC

Naim Naim XS integrated, Totem Hawk speakers

 

 

Just received the vinyl + "Vinyl Experience" download version in the mail today. Haven't had time to listen to either in detail. Will try to report back after I do.

 

But, my main question now is which file was used to cut the vinyl!

 

Thank you Chris for this report. I hope you get some answers, and if you do have the opportunity to speak with someone in the know, please do ask about the source used to cut the vinyl.

[junker]

The square root of the sum of the square of the amplitudes of every sample?

 

Junker,

please re-read the informations again, and think about it.

How would you calculate a RMS value out of "every sample"?

[InfernoSTi]

Here are two posts from another thread that comment on the DR issue and some of the possible concerns.

 

snip… The DR meter takes into account only the loudest 20% of a track. The idea is to avoid an increase in the DR value due to long low-level intros and such that would counterbalance the effects of severe compression and limiting in the calculation of the DR value, and it's the latter that the tool is supposed to reveal. The name "dynamic range meter" is a little unfortunate since it doesn't measure what most people would call dynamic range. In short, it measures micro-dynamics, not macro-dynamics.

 

Yes, ITU BS.1770-2 specifies a -10 LU "gate function" so that quiet passages are ignored. What the DR number measures the dynamic range / compression of the loud passages - it does not describe the "musical" dynamic range of the whole piece of music.

 

There is a pretty good paper by Thomas Lund (of TC Electronic), with interesting graphs...

 

Perhaps this is useful in the discussion…

 

John

[junker]

The DR-meter is one tool to use, and uses a weighting algorythm to make things easier to read and compare for the everyday user.

 

Yes, Adobe Audition seems to be a much better tool. DR-Meter is great tool - or rather has the potential to be- but people just need to be aware of the serious limitations of "Official" DR values from DR-Meter. It could be misleading.

 

It gives the vinyl experience a DR rating of 9 and HD Tracks a rating of only 6 despite the fact that the HD Tracks version has a whopping 42dB more of DR.

 

[TABLE=width: 446]

[TR]

[TD][/TD]

[TD=colspan: 2]HD Tracks[/TD]

[TD=colspan: 2]EQ The Vinyl Experience[/TD]

[/TR]

[TR]

[TD][/TD]

[TD]L[/TD]

[TD]R[/TD]

[TD]L[/TD]

[TD]R[/TD]

[/TR]

[TR]

[TD]Possibly Clipped Samples[/TD]

[TD]7[/TD]

[TD]5[/TD]

[TD]0[/TD]

[TD]0[/TD]

[/TR]

[TR]

[TD]Avg. RMS Amplitude[/TD]

[TD]-13.32[/TD]

[TD]-13.17[/TD]

[TD]-22.2[/TD]

[TD]-21.98[/TD]

[/TR]

[TR]

[TD]DR[/TD]

[TD]76.05[/TD]

[TD]71.73[/TD]

[TD]48.1[/TD]

[TD]49.14[/TD]

[/TR]

[TR]

[TD]DR Used[/TD]

[TD]61.35[/TD]

[TD]61.85[/TD]

[TD]38.7[/TD]

[TD]40.9[/TD]

[/TR]

[TR]

[TD]Loudness[/TD]

[TD]-10.34[/TD]

[TD]-10.48[/TD]

[TD]-19.37[/TD]

[TD]-19.45[/TD]

[/TR]

[TR]

[TD]"Official" DR Value[/TD]

[TD=colspan: 2]DR6[/TD]

[TD=colspan: 2]DR9[/TD]

[/TR]

[/TABLE]