HDTracks Downloads vs Qobuz/Tidal Streaming

[dontheriault]

I did a "Folger's Taste Test" with a buddy of mine last night and we came to the conclusion that HDTracks downloads just plain sound better than streaming the same music from Qobuz (or Tidal).  In some cases the difference was very stark, in others it was subtle.  In no case though, was the streaming version of a song chosen as having sounded better.  Pink Floyd's DSOTM was like going from FM radio to CD.  

My question... has anyone else done this and if so, am I crazy?  Am I too far into my own head on this.  Apologies if this has been a topic in the past, but maybe my search abilities are rather poor since I could not find a thread that was exactly like this.

[stefano_mbp]

Are you sure they are the same version?

On HDTracks the release date is 2021-10-19 while on Qobuz it is described as the 2011 Remasterd version … 

[DuckToller]

27 minutes ago, stefano_mbp said:

Are you sure they are the same version?

On HDTracks the release date is 2021-10-19 while on Qobuz it is described as the 2011 Remasterd version … 

I agree with Stefano, you can't be sure versions were identical.

However, comparing the streaming signal chain with the download playback chain might show you some differences for the signal path (which you didn't specify in the OP).

If you are a qobuz subscriber,  you can download content encrypted on your HD/phone and play it back on the Qobuz app on your computer or via BubbleUPNP from your phone .

That may help to provide a similar signal chain for content that proves to be similar in format, mastering and publishing.

[dontheriault]

Thanks guys.  I was questioning the version of DSOTM which is why I compared a couple of tracks that I knew were the same version (Jethro Tull Aqualung Steven Wilson Remix and Dave Brubeck's Take Five).  The difference was not nearly as clear, but the HDTracks WAV files always sounded marginally better.

 

I am streaming via a 1GB fiber connection to an Allo USBridge running Volumio (via ethernet).  The HDTracks files are stored on a shared drive that is plugged into my router, so the files are not local to the Allo.  

 

I don't want to go deep into the rabbit hole, but it leads me to believe that FLAC has more loss than I previously believed.  Given that FLAC files are smaller than WAV files, it only stands to reason that some data would be lost.  I just had no idea that the sound quality would be as noticeable as it is.

 

Just this guy's $.02

[stefano_mbp]

6 minutes ago, dontheriault said:

I don't want to go deep into the rabbit hole, but it leads me to believe that FLAC has more loss than I previously believed.  Given that FLAC files are smaller than WAV files, it only stands to reason that some data would be lost.  I just had no idea that the sound quality would be as noticeable as it is.

Flac files are smaller than Wav as they are compressed but Flac compression is totally lossless https://en.wikipedia.org/wiki/FLAC and nothing is lost, this is for sure

[firedog]

25 minutes ago, dontheriault said:

Thanks guys.  I was questioning the version of DSOTM which is why I compared a couple of tracks that I knew were the same version (Jethro Tull Aqualung Steven Wilson Remix and Dave Brubeck's Take Five).  The difference was not nearly as clear, but the HDTracks WAV files always sounded marginally better.

 

I am streaming via a 1GB fiber connection to an Allo USBridge running Volumio (via ethernet).  The HDTracks files are stored on a shared drive that is plugged into my router, so the files are not local to the Allo.  

 

I don't want to go deep into the rabbit hole, but it leads me to believe that FLAC has more loss than I previously believed.  Given that FLAC files are smaller than WAV files, it only stands to reason that some data would be lost.  I just had no idea that the sound quality would be as noticeable as it is.

 

Just this guy's $.02

No data lost with FLAC. Zero. That's the whole point of it. 

Convert a flac file to wave and you will see. 

[dontheriault]

Thanks.  What I find interesting is that when you convert from WAV to FLAC you can set up the level of compression.  

 

I am going to download the same version of two albums from Qobuz and HDTracks and do a comparison and then compare both to a Qobuz stream.  Believe me, I don't want the Qobuz stream to sound inferior... I don't want to spend more money on downloads than I have to.  I was just taken aback a bit doing some anecdotal back to back listening our of curiosity.  

[stefano_mbp]

6 minutes ago, dontheriault said:

What I find interesting is that when you convert from WAV to FLAC you can set up the level of compression

The compression (lossless) is the main feature of Flac, compression level is from 0 to 8 (9 levels) where the level 5 is the most efficient in terms of encoding time/reduced dimension.

It can happen, depending on dac mainly, that wav sound can be perceived better than flac and this can be explained with less cpu stress of the dac during the decoding process … but this is a very controversial topic

[dontheriault]

That sounds like the case.  I was not expecting any difference at all and upon hearing a difference (maybe perception bias...?) I want to dig in further.  In the end, I will listen to what my ears tell me, but just to iterate, I had an audio friend over and he had the same impression.  I did some blind back  to backs and he was always able to identify the HDTracks track that was playing.   We went through 5-6 before I thought maybe there was something to it.  

I understand the science and measurements, but the ears always tell us the most.  I have had numerous Schiit DACs and while the Bifrost 2 doesn't measure as well as the Modius per ASR, there is no question that the BF2 sounds better 100% of the time on 100% of the tracks.  Some things just cannot be fleshed out by a line graph.

[stefano_mbp]

5 minutes ago, dontheriault said:

That sounds like the case

You can try yourself without using streaming services: take a flac file and convert it to wav (or viceversa) then play them one after the other …

[firedog]

31 minutes ago, dontheriault said:

Thanks.  What I find interesting is that when you convert from WAV to FLAC you can set up the level of compression.  

 

I am going to download the same version of two albums from Qobuz and HDTracks and do a comparison and then compare both to a Qobuz stream.  Believe me, I don't want the Qobuz stream to sound inferior... I don't want to spend more money on downloads than I have to.  I was just taken aback a bit doing some anecdotal back to back listening our of curiosity.  

You need to set it up so you don't know which one is being played and make sure the volume is set the same for all playback. 

Otherwise it's not a good test.

[PeterG]

I did this 2 or 3 times years ago, the local music was always better.  Streamer was a Bluesound Node 2, local music a Naim Uniti Core, both through a Yggy.  I was not hyper disciplined about this.  OTOH, I did try a variety of music and it was not really a close comparison.  I moved the Node to my secondary system, and it's mostly gathered dust.  My dealer was not surprised, and opined that streaming imposes a lot of additional activity on the front end.  I lost interest at that point and decided if/when I wanted to try again, I would go to a high end streamer, such as Aurender or Naim

 

[dontheriault]

I had a Node and went to the Allo USBridge Signature since it was an upgrade but didn't cost me all the outdoors.  I am streaming both local and Qobuz through that unit to a Bifrost 2/64 via balanced outputs to the balanced inputs of a Freya + and then to monoblocked Aegirs.

I always wished the Node was better.  Super slick interface and rock solid stability.  The lack of a USB port is what turned me off.  I did not a slight bump in quality by going to Wifi rather than ethernet and I know that there is also a power supply upgrade that supposedly makes it better as well.... but I never ventured down that path.

[feelingears]

+1 for local files > streaming.

[One and a half]

I was expecting Qobuz to sound at least the equal of local files, but nowhere near it. Concluded noise was wrecking the image and tonal response, but which culprit was responsible was too much of an effort to find. Sticking with local files for some time to come.

[kennyb123]

On 1/4/2023 at 6:01 AM, dontheriault said:

My question... has anyone else done this and if so, am I crazy?  Am I too far into my own head on this.  Apologies if this has been a topic in the past, but maybe my search abilities are rather poor since I could not find a thread that was exactly like this.

Your not crazy at all.  I found that network optimization moved things in the direction of leveling the playing field.   I pretty much use Qobuz to figure out what downloads I want to buy.  

[zoltan]

@dontheriault If your streamer does the decoding of flac files it will stress the CPU and take away some juice from the power supply too. This was mentioned by somebody earlier. If you build a system where these two processes are done separately (like: Roon server + Roon Ready, HQ Player + HQ Player NAA) this will not not happen as the server is usually a very powerful PC/Mac/Linux bases machine and the streamer transport is separated by the network. I have a fanless PC running Roon+HQ Player and my SOtM SMS-200 ultra is used for streaming from there. Local files are stored on the server and both Qobuz and Tidal are integrated under Roon. 
I have hundreds of CD rips from XRCD, FIM, Analogue Production, Gold CDs, RVG editions, Reference Recordings from CDs I owned and also HDtrack files. CD rips are all AIFF (uncompressed) format. 
It is very hard to identify if two releases ( masters) are the same or not so very difficult to do A/B testing but when I was able to do that, there was absolutely no difference in the sound quality between the local files (uncompressed) and Qobuz (flac). My system is about USD80K system (not bragging, just for reference) and reveals practically any small change. Maybe my ears don't :-)  
Try to build a two machine playback system. This is what Roon recommends too. 

[dontheriault]

@zoltan Thanks for the reply.   Your experience is consistent with most people's experiences here.  I did a blind test with a friend who has a good ear and I would say he could guess the HDTracks files about 50% of the time, and frankly that could have just been luck of the draw.  So far, only the Pink Floyd tracks are the only ones that seem to yield a difference.  Whatever the case, I just did a DAC upgrade and to be honest, I am not sure I really care anymore.  I think I solved all of my experience problems with this DAC, so I am going to just enjoy whatever source I have at my disposal (within reason... I am still an audio snob after all).  

 

Thanks to everyone that replied.  I learned a fair amount in this thread.

[davide256]

On 1/4/2023 at 11:16 AM, dontheriault said:

Thanks guys.  I was questioning the version of DSOTM which is why I compared a couple of tracks that I knew were the same version (Jethro Tull Aqualung Steven Wilson Remix and Dave Brubeck's Take Five).  The difference was not nearly as clear, but the HDTracks WAV files always sounded marginally better.

 

I am streaming via a 1GB fiber connection to an Allo USBridge running Volumio (via ethernet).  The HDTracks files are stored on a shared drive that is plugged into my router, so the files are not local to the Allo.  

 

I don't want to go deep into the rabbit hole, but it leads me to believe that FLAC has more loss than I previously believed.  Given that FLAC files are smaller than WAV files, it only stands to reason that some data would be lost.  I just had no idea that the sound quality would be as noticeable as it is.

 

Just this guy's $.02

I don’t buy music from HDTracks anymore. With Qobuz anything i buy can be re-downloaded at a later date and mediocre recordings don’t  improve by buying a download. My HDtrack recordings are pretty hit or miss for engineering quality,

Qobuz makes it far easier to identify well recorded music that will be worth the cost of local play  

[Jud]

1. You can buy downloads from Qobuz (Tidal also I'm sure, but I happen to be a happy Qobuz subscriber).

 

2. When I bought from HDTracks, they gave me quite a hassle about re-downloading if there was an error on my end  - or theirs, of which there have been some doozies, e.g., the Who's Tommy missing three tracks.  I can't think of another download vendor that does this (unless they've changed), so why go through the bother?

 

3. As has been mentioned, if you really think the download will sound better, a streaming+download service offers the opportunity to listen and determine whether it's a download you really want to buy, rather than guessing from 30 second snippets.

[Jud]

On 1/4/2023 at 10:20 AM, firedog said:

You need to set it up so you don't know which one is being played and make sure the volume is set the same for all playback. 

Otherwise it's not a good test.

 

Since echoic memory (that is, our memory for sounds) lasts about 4 seconds, even after all that it's not a good test.

 

What is a good test is to use well recorded mono, for example the Beach Boys' Pet Sounds (or create mono from well recorded stereo - you can do this with Audacity), then use Paul Kane's DeltaWave Comparator ( https://deltaw.org ) to listen to both tracks simultaneously.  Switch ears and see whether your preference (if any) switches too.

 

This unfortunately won't work for streaming vs. download, nor for FLAC vs. WAV (though it will for original WAV vs. FLAC that's expanded to WAV).

 

One other thought about eliminating any nagging doubts regarding compression (even if it's all in your head, why be bothered?): AIFF is an available uncompressed download format from Qobuz, and it's more easily amenable to good metadata tagging than WAV files are.

 

 

[cpcat]

So qobuz doesn’t offer WAV downloads?    AIFF is fine though as long as dbpoweramp does conversion from AIFF to uncompressed FLAC?  I’m pretty sure it does, will have to check.
 

Currently evaluating this,  as have recently adopted Roon and am trying qobuz again.  HDTRACKS has been my go-to for downloads til now.

 

On 1/18/2023 at 1:05 AM, Jud said:

 

Since echoic memory (that is, our memory for sounds) lasts about 4 seconds, even after all that it's not a good test.

 

What is a good test is to use well recorded mono, for example the Beach Boys' Pet Sounds (or create mono from well recorded stereo - you can do this with Audacity), then use Paul Kane's DeltaWave Comparator ( https://deltaw.org ) to listen to both tracks simultaneously.  Switch ears and see whether your preference (if any) switches too.

 

This unfortunately won't work for streaming vs. download, nor for FLAC vs. WAV (though it will for original WAV vs. FLAC that's expanded to WAV).

 

One other thought about eliminating any nagging doubts regarding compression (even if it's all in your head, why be bothered?): AIFF is an available uncompressed download format from Qobuz, and it's more easily amenable to good metadata tagging than WAV files are.

 

 

 

[Jud]

15 hours ago, cpcat said:

So qobuz doesn’t offer WAV downloads?

 

 

I don’t know, they might. I have all my stuff in AIFF, so that’s what I’m familiar with.

 

15 hours ago, cpcat said:

   AIFF is fine though as long as dbpoweramp does conversion from AIFF to uncompressed FLAC?  I’m pretty sure it does, will have to check.

Don’t use it so wouldn’t know. If it doesn’t there are other apps that do (on Windows, MediaHuman is free and I think it works well).

[cpcat]

16 minutes ago, Jud said:

 

I don’t know, they might. I have all my stuff in AIFF, so that’s what I’m familiar with.

 

Don’t use it so wouldn’t know. If it doesn’t there are other apps that do (on Windows, MediaHuman is free and I think it works well).

Thanks Jud.

Sorry if I misunderstood.  According to this WAV is available along with AIFF, ALAC, and FLAC:  https://help.qobuz.com/en/articles/10167-what-are-the-different-audio-formats-available-for-download

 

[Apollo]

AIFF and WAV anf FLAC with zero compression offer exactly the same bitperfect quality.

 

Initially AIFF was exclusively for MACs but was very quickly supported on Windows.

WAV was originally only on Windows (and still is AFAIK)

The other difference was that AIFF supports the inclusion of metadata, which WAV did not.  Nowadays, WAV does support Metadata, but not as extensive as AIFF.

FLAC was originally supported on Windows and Linux, nowadays supported on MAC  as well. 

Most ripping applications did not offer choice of compression level with FLAC, today almost all do.

 

At the time I ripped all my CD's (+/-2000) in (compressed) flac by means of a ripping robot.

I converted everything to losless AIFF by means of dbPoweramp batch convertor, as this format was already supported on all platforms that I was considering to use (no Linux at that time).

 

dbPoweramp can perfectly convert to FLAC with 0 compression without any issue, if that would be necessarry for you. (I cannot think of any reason why AIFF would need to be converted, but to each his own).

 

I would never choose for WAV.  Not supported on MAC (AFAIK) and mimited support of metadata.