I am done with HDTracks

[mojave]

I have 3 copies of Elton John's Madman Across the Water: DJM, HD Tracks, and MCA Records. They all differ in DR values, but not as much in Loudness Range measured according to EBU ITU-R BS.1770-2 international standard for measuring audio loudness (also known as R128).

Here are the specifications for TT Dynamic Range Meter: TT Dynamic Range Specifications

Here are the specifications for Loudness Range (R128): Loudness Range: A measure to supplement loudness normalisation in accordance with EBU R 128

JRiver Media Center V19 provides the data for both specifications when doing an audio analysis on your library.

I use use a view to show me albums for an artist and then can drill down to the album level to see the different labels/versions:





Here is how the 3 albums compare:



You can see that even though the HD Tracks has lower DR values, its loudness range is comparable (listed by JRiver as Dynamic Range (R128).

 

[rbbert]

I tend to evaluate compression in two ways. First is the easy one we've discussed in other topics; how does it sound when you keep turning up the volume. IME the more compressed a piece is the lower the comfortable volume, or put another way I can turn up a less compressed piece to a noticeably higher average volume (simply measured by a weighted measurement from an SPL meter) before it gets uncomfortably loud. Second is a little tougher, it involves looking at waveform view in an audio editing program for a "gestalt", often aided by analyzing some of the louder sections for peak, peak rms and average rms volumes.

 

[esldude]

http://www.ohl.to/audio/downloads/

If you look here, and download a zip file called Trackalyzer, it is something I saw in another forum recently. It gives several good plots for analyzing files beyond just a DR number (though it gives that too). You will also need to get Octave for this to work. And it is currently Windows only.

You get plot of RMS level vs time for the track, a level histogram (which will give you an idea how compressed or dynamic it is) and a spectrum graph.

I also find it useful as I assume many know to look at a spectrogram mode. It is easy to see when something has been brickwalled at 22.05 khz from CD and upsampled to be sold as hirez (which HD tracks has been known to do more than once).

 

[rbbert]

...It is easy to see when something has been brickwalled at 22.05 khz from CD and upsampled to be sold as hirez (which HD tracks has been known to do more than once).

I thought we had progressed beyond that type of misinformation? HDtracks sells what the labels give them, and have made efforts (perhaps not perfect) to avoid selling those types of files.

 

[esldude]

I thought we had progressed beyond that type of misinformation? HDtracks sells what the labels give them, and have made efforts (perhaps not perfect) to avoid selling those types of files.

Well, I said they have done it, and they have. They claim to have made efforts. I haven't purchased everything they make to see if that is true or not. I don't see how it is misinformation if it is true information. They have been known to do what I claim more than once.

I consider them guilty of far too little due diligence. Maybe you think just taking what is offered by labels and marketing them in glowing terms of high resolution sound quality without checking is fine for them to do. They agree to obtain and sell recordings at higher prices at various resolutions of remasterings. Yet didn't bother to spend 5 minutes (actually literally less than 5 minutes) to see if what they were sent was redbook sourced? When this became apparent it took them around a year to do more than just offer to credit customers. And they didn't credit all customers of these files, just those who became aware of it and complained. Business is business, and maybe I expect too much, but their handling of that did not engender lots of trust to me personally. I do believe they also still sale some of those earlier offerings that are redbook material at those higher rates. I have pretty much ceased doing business with them for this reason. The whole hi-rez business needs to get together to agree on a standard provenance of such offerings.

If you have other info they have done more then point to it for me.

 

[rbbert]

...If you have other info they have done more then point to it for me.

http://www.whatsbestforum.com/member.php?60-Bruce-B

 

[Bruce B]

Well, I said they have done it, and they have. They claim to have made efforts. I haven't purchased everything they make to see if that is true or not. I don't see how it is misinformation if it is true information. They have been known to do what I claim more than once.

I consider them guilty of far too little due diligence. Maybe you think just taking what is offered by labels and marketing them in glowing terms of high resolution sound quality without checking is fine for them to do. They agree to obtain and sell recordings at higher prices at various resolutions of remasterings. Yet didn't bother to spend 5 minutes (actually literally less than 5 minutes) to see if what they were sent was redbook sourced? When this became apparent it took them around a year to do more than just offer to credit customers. And they didn't credit all customers of these files, just those who became aware of it and complained. Business is business, and maybe I expect too much, but their handling of that did not engender lots of trust to me personally. I do believe they also still sale some of those earlier offerings that are redbook material at those higher rates. I have pretty much ceased doing business with them for this reason. The whole hi-rez business needs to get together to agree on a standard provenance of such offerings.

If you have other info they have done more then point to it for me.

Let me see you check 40-50 track of one classical album in 5 minutes. To correctly and diligently check for upsampling or whatever, it may take an hour per album. Don't tell me all I need to check is one track because it has been found that labels put different mixes on an album and 1-2 tracks out of 15 can be upsampled.
HDtracks had no knowledge of 24/44.1 files on SACD's. They trusted the labels and took what they said at face value. Everyone praised the sonority of these files until this was found out, they started back tracking and stumbling over their words. This is why a subjective database wouldn't be worth the bandwidth it takes up.
It took me the better part of a year to check their back catalog. You want to put blame someone, blame me.

Besides, I haven't been able yet to hear with my eyes, so plots of DR values and such have no meaningful value what so ever.

 

[Bruce B]

I have 3 copies of Elton John's Madman Across the Water: DJM, HD Tracks, and MCA Records. They all differ in DR values, but not as much in Loudness Range measured according to EBU ITU-R BS.1770-2 international standard for measuring audio loudness (also known as R128).

Here are the specifications for TT Dynamic Range Meter: TT Dynamic Range Specifications

Here are the specifications for Loudness Range (R128): Loudness Range: A measure to supplement loudness normalisation in accordance with EBU R 128

JRiver Media Center V19 provides the data for both specifications when doing an audio analysis on your library.

I use use a view to show me albums for an artist and then can drill down to the album level to see the different labels/versions:

Here is how the 3 albums compare:

View attachment 14341

You can see that even though the HD Tracks has lower DR values, its loudness range is comparable (listed by JRiver as Dynamic Range (R128).

And this sounds how??

 

[esldude]

Let me see you check 40-50 track of one classical album in 5 minutes. To correctly and diligently check for upsampling or whatever, it may take an hour per album. Don't tell me all I need to check is one track because it has been found that labels put different mixes on an album and 1-2 tracks out of 15 can be upsampled.
HDtracks had no knowledge of 24/44.1 files on SACD's. They trusted the labels and took what they said at face value. Everyone praised the sonority of these files until this was found out, they started back tracking and stumbling over their words. This is why a subjective database wouldn't be worth the bandwidth it takes up.
It took me the better part of a year to check their back catalog. You want to put blame someone, blame me.

Besides, I haven't been able yet to hear with my eyes, so plots of DR values and such have no meaningful value what so ever.

I assume you are referring to working with the master tapes (perhaps analog) of such a recording. Does HDtracks get that or do they get a digital track to distribute? If they get a digital track to distribute (which they must at some point as that is how they sell them) why can they not take a few minutes to look at them?

While some of their albums have been mixed with some upsampled tracks, others have been whole albums of redbook material. Got some good explanation why someone paying near $20 for a 192/24 album should be okay with that? Got some explanation as to why 192/24 from a redbook source would sound better than a lower (and cheaper) resolution from HDtracks?

 

[rbbert]

Can you give an example of a 24/192 file from HDtracks that is upsampled Redbook?

 

[esldude]

Can you give an example of a 24/192 file from HDtracks that is upsampled Redbook?

Just off the top of my head, there was a long thread about the Talking Heads Speaking in Tongues album here on WTB. Bruce was involved in that and had lots of input. Showed graphs of what he supplied and Amir confirmed it wasn't what HDtracks was selling. There are more than just this example. Google hdtracks and upsampled you will find plenty to read about.

I having read of these issues won't purchase from HDtracks unless someone has confirmed them as genuine hi-rez. The few I purchased from them were hirez of some sort. Having read of their in my opinion relatively poor handling of this, and difficulty at times to deal with them about refunds on such, I eventually decided I didn't care to do any business with someone you have reason to be so leery of.

 

[rbbert]

Speaking in Tongues is not available in 24/192, and I haven't heard of any 24/192 tracks being sold that are upsampled Red Book. So I'll ask the question again.

 

[esldude]

Speaking in Tongues is not available in 24/192, and I haven't heard of any 24/192 tracks being sold that are upsampled Red Book. So I'll ask the question again.

Oh so it was 24/96, does that really alter the idea of trusting you get what you pay for?

 

[rbbert]

Oh so it was 24/96, does that really alter the idea of trusting you get what you pay for?

First, you should say what you mean. Second, I've had a lot of dealings with HDtracks, purchased a lot of stuff from them and have had some major disappointments. Some they have dealt with satisfactorily, some not, about on a par with most retailers; better than some, worse than others, but overall about average in that respect. However, I've also received some of the best sounding recordings I own, in all genres of music, that aren't readily available any other way. So as I posted earlier in this thread, it pays to do your research before you buy, just as you would before you buy anything.

 

[TJE]

To me the important point in the article is that at least according to the author, the lossy MP3 with less dynamic compression sounded clearly better than the "high res" file with more dynamic compression. I haven't made the comparison with this album but would expect to agree with this, subjectively. I'd be curious if anyone disagrees, and finds the "high res" version does not sound worse than the less compressed version. My bias is that people tend to over-emphasize the benefits of higher bit depth or sampling rate when that's not the most important limiting factor in sound quality for digital files.

I also suspect this is one of many examples where an audiophile's preference for vinyl vs. CD of "the same album" is actually a preference for the vinyl master, which has less dynamic compression than the CD master.

 

[mojave]

I have 3 copies of Elton John's Madman Across the Water: DJM, HD Tracks, and MCA Records. They all differ in DR values, but not as much in Loudness Range measured according to EBU ITU-R BS.1770-2 international standard for measuring audio loudness (also known as R128).

Here are the specifications for TT Dynamic Range Meter: TT Dynamic Range Specifications

Here are the specifications for Loudness Range (R128): Loudness Range: A measure to supplement loudness normalisation in accordance with EBU R 128

JRiver Media Center V19 provides the data for both specifications when doing an audio analysis on your library.

I use use a view to show me albums for an artist and then can drill down to the album level to see the different labels/versions:

Here is how the 3 albums compare:

View attachment 14341

You can see that even though the HD Tracks has lower DR values, its loudness range is comparable (listed by JRiver as Dynamic Range (R128).

And this sounds how??

On Thursday evening I spent an hour listening back and forth to the DJM and HDTracks version of Indian Sunset. I chose it because I like the song, it has a lot of dynamics, and the difference in DR values was the greatest (14 vs 8).

I used R128 volume leveling so that both would be the same volume. This applied -5.2 dB to the HDTracks. I also changed the album name so both songs had different album names. Otherwise R128 volume leveling thinks the songs are part of the same album and that the volume level differences are intentional so it won't change volume. I used the playback range tag in JRiver to compare various smaller portions of the song.

The most immediate noticeable difference is that the DJM has a much higher noise floor. So much so that I would choose the HDTracks album over it for that alone. By noise floor I refer to the hiss heard during what are supposed to be silent portions. The HDTracks version had slightly more detail and nuances. The DJM had a longer reverb tail. I couldn't tell any difference between them regarding dynamics. Overall I preferred the HDTracks.

The TT Dynamic Range uses an algorithm that I linked to above. On Friday I decided to see what the actual crest factor was of the two songs. I used the Voxengo Span VST plugin since it shows max crest factor along with a lot of other data.

The max crest factor of Indian Sunset for HDTracks is 15.5. For the DJM it is 14.7. Interesting. I also looked at the average RMS levels of each song. This is in relation to 0 dBFS and doesn't relate to the actual listening volume. The HDTracks also has higher peak to RMS levels for the portion I checked (3:00-3:45). Again, interesting.

I would say that the TT Dynamic Range number isn't a very good metric for determining sound quality - at least with this song. I wonder if the high/loud noise floor is contributing to the increased number for the DJM version?

 

[TJE]

My understanding is that dynamic compression brings up "nuances" including noise floor. These results are screwy, including what you note about noise floor and crest factor. It doesn't look like this is a comparison of more vs. less compression, just two versions of the song that have several differences in how they were produced.

 

[esldude]

A post about being ripoffs from HDTracks. The recording is the 96/24 Beck Album Morning Phase

http://www.ultrahighendreview.com/another-rip-off-from-hdtracks/

And a spectrogram of one track.

http://www.stereophile.com/content/beck-track-11

And here you can read comments by JA and what Chesky had to say. (Basically they have to sell files that they are given without modifications). These are in the comments section to this short article.

http://www.stereophile.com/content/becks-imorning-phasei-and-vinyl-experience

 

[rbbert]

I just bought two "new" releases on sale, Keith Jarrett "Arbour Zena" and John Abercrombie Quartet "39 Steps". Both sound great, are uncompressed and by spectrographs are true 24/96 and 24/88.2 respectively. FWIW.

 

[esldude]

I just bought two "new" releases on sale, Keith Jarrett "Arbour Zena" and John Abercrombie Quartet "39 Steps". Both sound great, are uncompressed and by spectrographs are true 24/96 and 24/88.2 respectively. FWIW.

It is worth plenty to me. I don't buy HDTracks until someone confirms. The Beck is one I would have gotten too. Maybe David Chesky should discount for how many of the tracks are low rez on the selling price.

Not familiar with the Abercrombie Quartet, but know of Keith Jarrett. So thanks.