Just got a Yggdrasil!

[jabbr]

22 hours ago, esldude said:

You are forgetting it doesn't prove it fails to reject the null hypothesis.

 

It definitely would not prove that all DACs sound the same. 

That's the point ... mixing subjective & objective measurements can descend into chaos.

 

if there is a contentious issue, you want very clear and unambiguous measurements.

[esldude]

14 hours ago, PeterSt said:

Guys, help me with some analysis please. What can this mean :

 

DAC                MaxLevel        Avg SPL       Track Length

Original            32399              1012             1:03.54

Ygg                  31814              773               1:00.40

NOS1a             31814              951               1:05.89

Control             31814              778               1:00.40

 

DAC : The DAC used to take the recording from.

MaxLevel : The maximum audio word value, recalculated to 16 bits (the recordings are in 24 bits).

Avg SPL : The average SPL which calculates from the level of each sample in the file, which is averaged (envision that this is used for volume normalization).

Track Length : Obtained because it influences the math when leading/trailing space is present.

 

I don't trust the MaxLevel to be the same for each of the recordings. It would mean that all DACs involved had the exact same output voltage, or that Mani was capable of setting the volume that equal (to the 1/32767th). Also notice that the different filter of each DAC will imply different audio word values to begin with, including the maximum levels (better : especially the maximum levels).

Possibly something in the recording process is soft-clipping.

 

The fact that the SPL of the original is higher is logical. This is because the MaxLevel of the original is close to the maximum allowed (32767 (plus and minus) for 16 bits) and Mani will have set the input level of the ADC so that nothing is "over".

 

It was sorted out (thank you Dennis) that Mani used a rolled off filter for his NOS1a during this recording, with the interesting fact that the SPL is something like 25% higher than the others (this is less than 3dB). However, this can be contributed to the fact that the filter used also aliases beyond 22050Hz at a level of ~ -66dB (up to 24KHz of the recording). The level of this is incorporated in the SPL math.

 

It is also logical that the SPL of the original is inherently higher (although we can not prove that by the given data) because the filter used rolls off slower than both the Ygg and Control DAC. Thus here too more "SPL" is calculated.

 

It was also sorted out that the Ygg's recording is 5.9dB lower in level than the original. This is not what the above table is telling us (with the Ygg as the base, the SPL of the original is close to 3dB more). Do notice that the "math" involved calculates per sample, and does not calculate the dB per sample or something of that kind (which would imply high rounding error).

I mention this only because we may wonder how this math is done and how the data is to be interpreted for real. "Interpreted" means : the level matching. So for example, when my software takes into account the aliasing and other software does not, the level matching is apples and oranges if such software is the base for it. It is more complicated because my software does not take into account aliasing at all BUT in this case it has to because of how the recording implies it. This is quite crucial. Look :

 

 

This is the original (filtered by some production means before it went to CD).

 

 

This is the filtering how the Ygg does it. This is on top of the original, of course.

 

 

How the Control DAC does it (looks to be the same DAC as the Ygg, but alas).

 

 

And how the NOS1a does it per Mani's filter settings of that time.

 

So as you can see, the SPL (or average dB level) is influenced massively by the filtering.

My key point is that in this special case the the math may be regarded as invalid because we play 44.1 but record in 48, which implies that the actually to be discarded part between 22050 and 24000 is now incorporated and this influences the SPL even more. You're not only seeing this when the NOS1a is compared with the others, but also when the Original is compared with Ygg and Control. Never mind that the Original ends at 22050 because it's about what is in the file and what plays. For the original plays more (level) at the end of the spectrum while for e.g. Ygg this is only very low level, which for the NOS1a is a tad more than very low level.

 

So for this particular case it will be very tough to use software means to determine the average level, hence level match properly.

 

With the above all said, what remains is that perceived clipping.

It is just guessing of course, but if soft clipping is in order indeed, all will be moot although it will depend on how often it happens. But when "often" it will influence (SQ) massively.

 

That's all.

 

 

 

 

Firstly I made all files equal to the shortest after lining them up.  I chopped off the excess in the original file when I first starting looking at these a couple days ago.  I haven't continued with the NOS1 file after seeing it differed so much.

 

I did look at them with the amplifying effect in Audacity.  It tells you how much you can amplify a selection without exceeding 0 dbFS.  Both channels of the Control DAC file show they are higher in value by .007 db vs the Yggy file.  So the max sample value should be different for each.

 

Using Audacity Sample Data Export which reports sample values in db the highest value for the Yggy was -5.82315 left channel and -5.88960 right channel.   For the control DAC it was -5.81615 left and -5.89704 right.   So a difference of .007 db for left channel and .00744 db for the right channel between the two DACs.  So I don't know how he matched levels, but he did a damned good job of it.  Yet they are not identical from my information.

 

When I originally level matched I sample rate converted the original file to 48/24 which resulted in some intersample overs. I reduced the original file by .3 db and converted again.  Then compared it to the recorded 48/24 file.  Digitally amplified levels accordingly though I listened to 44.1 vs 48 for auditioning. 

 

Now I was wondering if the two recorded files were of the same DAC myself.  They are certainly very, very similar.  The extremely minor volume differences could have come from recording at different times of day with voltage/temperature drift. 

 

[esldude]

9 hours ago, gmgraves said:

Hey, guys; how many angels can you fella's fit on the head of that pin, anyway?Looks infinite from where I'm sitting!

 

16,777,215 I would say. 

[manisandher]

So in summary:

 

1. some people hear no difference between the original file and the Yggy capture (implying that the Yggy, and my replay/recording chain are essentially 'perfect')

 

2. some people hear a difference between the original file and the Yggy capture, but not between the different captures (implying that the replay/recording chain is not of high enough resolution)

 

3. some people hear a difference between the original file and the Yggy capture, and also between the different captures (implying that the replay/recording chain is of high enough resolution)

 

4. no one (as far as I'm away) hears no difference between the original file and any one of the captures (not necessarily the Yggy file), but does hear a difference between the captures

 

For me, we need scenario 3 or 4 to be the case to conclude anything useful about the Yggy.

 

Thanks to everyone for giving the files a go. And for Dennis & Peter for their respective analyses on the files. The last angel on that pin from me.

 

Mani.

[esldude]

48 minutes ago, manisandher said:

So in summary:

 

1. some people hear no difference between the original file and the Yggy capture (implying that the Yggy, and my replay/recording chain are essentially 'perfect')

 

2. some people hear a difference between the original file and the Yggy capture, but not between the different captures (implying that the replay/recording chain is not of high enough resolution)

 

3. some people hear a difference between the original file and the Yggy capture, and also between the different captures (implying that the replay/recording chain is of high enough resolution)

 

4. no one (as far as I'm away) hears no difference between the original file and any one of the captures (not necessarily the Yggy file), but does hear a difference between the captures

 

For me, we need scenario 3 or 4 to be the case to conclude anything useful about the Yggy.

 

Thanks to everyone for giving the files a go. And for Dennis & Peter for their respective analyses on the files. The last angel on that pin from me.

 

Mani.

And what's the identy of the control DAC?

[gmgraves]

13 hours ago, esldude said:

 

16,777,215 I would say. 

 

Is that how many posts there are about The ADC mani used to make his test recordings to show that his eggy isn't very good? I knew it was a lot, but...

 

[Andyman]

13 hours ago, esldude said:

 

16,777,215 I would say. 

Oops - just stepped on by mistake. Need a 32 bit pin now...

[Abtr]

I listened to the 3 files (via Foobar2000 WASAPI, with the original at -6db). The original file clearly sounds best. The Yggy and control recordings sound very similar. The trebble (cymbals) of the Yggy recording may lack a little bit of detail and the piano may sound a bit recessed, relative to the control file. Not sure at all if I could distinguish between Yggy and control in a blind test. I think I could identify the original though.. 

[Superdad]

Hi Mani:

 

I listened to all three files yesterday.  But only for about 10 minutes.

 

First I will say that I appreciate that you chose a track with acoustic instruments.  While I have very eclectic musical taste and enjoy all sorts of electric and acoustic instrumentation, I feel that recordings of real instruments in a real space are the most appropriate for making judgements about which component is more musically truthful and correct.

 

That said, I found your selection to be rather torturous.  Not because of the music itself--though that is not my favorite style of jazz--but because of the crammed together perspective of the recording, and the timbre of some of the instruments.  The upright bass was okay--if not terribly detailed--but oh that piano and those high-hats!  Let me guess, the lead name on the record was that of the drummer?

 

While I applaud you for choosing a track with lots of varied and sharp transients (such always make judgements easier), the lack of depth and the irritating nature of the track kept me from listening for very long.  (Not that I needed to--see below.). If you ever decide to do this sort of thing again, I would be pleased to send you a couple of recordings that I think everyone would find more enjoyable and which would still readily reveal--possibly more so--the qualitative differences between components.

 

 

Anyway, I had no trouble hearing the differences between all three files, with the original rip of course being quite a lot better than either of the other two.

While your ADC did an admirable job of preserving the "sound" of the instruments, both of the versions run through it lost "realness" and "feeling." (And yes, I always level-matched using my stepped attenuators--after first using a meter to determine which steps were appropriate.)

 

As for the differences between your "Contol DAC" file and the "Yggy" file:  Yes I heard some, but I found that the nature of the track selection (particularly the somewhat unnatural aggressiveness of some of the instruments) did not lend itself to identifying aspects of one of these two files that were clearly more musically correct.

I am 98% certain that I would quickly have a preference between the Yggy (never heard in my system BTW) and whatever was your "control" DAC where I to hear them in person--not through the "filter" of your ADC chain.  (Especially if I could use my own recordings and my own room/system.)

 

I doubt if any of the above shines any more light on your test or the debate going on in this thread.  But since I made the effort, I figured I'd share my results.

 

Best,

--Alex C.

[JoeWhip]

This thread frankly went off the rails awhile ago. Nevertheless, with my experiences yesterday, I figured I would post one last time. I was involved in a recording project yesterday at one of the best studios in the area. We used a fantastic sounding Steinway, one of the best I have ever heard. It was tuned for the session. The recording was vocal with jazz piano. I know the sound of the vocalist well as he is my son. We recorded using pro tools at 24/96. I played the piano a bit myself beforehand. I brought the files home and listened to the 24/96 files as well as a 16/44.1 version created for digital distribution. So, I heard the unamplified piano and vocal in the studio, the live mic feed and the playback. Listening at home through the Yggy, the file sounded just like the live mic feed and the control room playback. The piano had a rich sound with natural decay. It was stunning just like hearing it live. The piano was not thin or hard sounding at all, quite the opposite. Same with the vocals. Both the 24/96 and the 16/44.1 files were superb with the 24/96 sounding just a tad better. This is a great DAC. If you don't like the sound, fine. But it is not screwing up the sound. Far from it. There is a ton of work going on to try to dump on this DAC. I find this to be ridiculous and a waste of time proving nothing. Time to close the thread. Bye.

[GUTB]

Listened to Yggdrasil in the Schiit room at AXPONA. Vidar and Freya setup playing on a Song3 tower speakers. Very resolving, up there with the better DACs at the show -- W4S DAC2v2SE, Optologic, etc. The Schiit setup was a little thin on harmonics and shallow soundstage.

[semente]

7 hours ago, GUTB said:

Listened to Yggdrasil in the Schiit room at AXPONA. Vidar and Freya setup playing on a Song3 tower speakers. Very resolving, up there with the better DACs at the show -- W4S DAC2v2SE, Optologic, etc. The Schiit setup was a little thin on harmonics and shallow soundstage.

 

Your constant references to shallow soundstage and thin harmonics probably mean that you really enjoy euphonic distortion and should get a DAC with a valve output stage, or even an analogue source.

I wonder is this can be DSP'ed to taste...

 

R

[mourip]

13 hours ago, GUTB said:

Listened to Yggdrasil in the Schiit room at AXPONA. Vidar and Freya setup playing on a Song3 tower speakers. Very resolving, up there with the better DACs at the show -- W4S DAC2v2SE, Optologic, etc. The Schiit setup was a little thin on harmonics and shallow soundstage.

 

I suspect that the issue was more with the associated equipment and tweaking. With a good front end the Yggy has a deep, wide, detailed, natural presentation. It will sound as good as the associated equipment will allow. This is my own observation as I have made changes/upgrades in my own system. I do use a tube preamp and amp but they are not classically "tubey".

 

The DACs I had before it were a Ws4 DAC2 DSD SE and an Auralic Vega. These were thin and uninvolving by comparison...in my own system.

[GUTB]

14 hours ago, semente said:

 

Your constant references to shallow soundstage and thin harmonics probably mean that you really enjoy euphonic distortion and should get a DAC with a valve output stage, or even an analogue source.

I wonder is this can be DSP'ed to taste...

 

R

 

It sounds like you've never heard an actual soundstage. You probably think a good soundstage is hearing a left and right and a center.

[semente]

4 minutes ago, GUTB said:

 

It sounds like you've never heard an actual soundstage. You probably think a good soundstage is hearing a left and right and a center.

 

It's possible. I also have a bit of trouble hearing soundstage in live performances (of classical music) with my eyes closed.

 

R

[Jud]

3 minutes ago, GUTB said:

 

It sounds like you've never heard an actual soundstage.

 

Please try to avoid coming off like someone we can't have a decent, interesting conversation with.

 

Soundstage depends on the recording.  If you have equipment that imparts a 3D soundstage to every recording, it isn't being faithful to the music.  I find I eventually tire of such "one-trick ponies."  Do you have a different opinion?

[Ralf11]

I find soundstaging to be nice and fun, but not essential.  I think a lot of what panel speaker people worry about re spacing from walls is the collapse of the soundstage, but they still sound good in other ways if slammed right up to the rear wall.

 

Here are some fun facts relating to wall reflections and soundstaging (when present in the recording to begin with):

 

 

[esmit]

Indeed, it looks rather vague.

5 minutes ago, Ralf11 said:

I find soundstaging to be nice and fun, but not essential.  I think a lot of what panel speaker people worry about re spacing from walls is the collapse of the soundstage, but they still sound good in other ways if slammed right up to the rear wall.

 

Here are some fun facts relating to wall reflections and soundstaging (when present in the recording to begin with):

 

 

 

[esldude]

6 minutes ago, Ralf11 said:

I find soundstaging to be nice and fun, but not essential.  I think a lot of what panel speaker people worry about re spacing from walls is the collapse of the soundstage, but they still sound good in other ways if slammed right up to the rear wall.

 

Here are some fun facts relating to wall reflections and soundstaging (when present in the recording to begin with):

 

 

Your insert there is too small to read.

 

I can't speak for other panel speaker people, but spacing from the wall has an effect on the lowest octave the speaker can produce.  It is the criteria I use to choose the spacing.    The right spacing reinforces the bottom where the speaker is rolling off and might dampen where it has a room related peak.  A smooth lower couple octaves is much nicer than a lumpy one or one mostly missing two octaves. 

 

As for slamming right up to a wall all my panel speakers sounded horrible.  If you have to be close to a wall you can sometimes get away with it if you angle them steeply so the backwave bounces off at a severe angle. 

[mansr]

5 minutes ago, Jud said:

 

Please try to avoid coming off like someone we can't have a decent, interesting conversation with.

That ship has sailed.

5 minutes ago, Jud said:

Soundstage depends on the recording.  If you have equipment that imparts a 3D soundstage to every recording, it isn't being faithful to the music.  I find I eventually tire of such "one-trick ponies."  Do you have a different opinion?

Proper soundstage reproduction depends mostly on speaker placement. DACs and amps make very little difference in this regard assuming they are reasonably competent pieces.

[mansr]

7 minutes ago, esldude said:

Your insert there is too small to read.

 

I can't speak for other panel speaker people, but spacing from the wall has an effect on the lowest octave the speaker can produce.  It is the criteria I use to choose the spacing.    The right spacing reinforces the bottom where the speaker is rolling off and might dampen where it has a room related peak.  A smooth lower couple octaves is much nicer than a lumpy one or one mostly missing two octaves. 

 

As for slamming right up to a wall all my panel speakers sounded horrible.  If you have to be close to a wall you can sometimes get away with it if you angle them steeply so the backwave bounces off at a severe angle. 

The dipole character of panels makes rear reflections much more influential than with box speakers.

[crenca]

23 hours ago, GUTB said:

Listened to Yggdrasil in the Schiit room at AXPONA. Vidar and Freya setup playing on a Song3 tower speakers. Very resolving, up there with the better DACs at the show -- W4S DAC2v2SE, Optologic, etc. The Schiit setup was a little thin on harmonics and shallow soundstage.

 

Did they have the Song3's right up against (the front) wall?  

[Ralf11]

esl - it is the back reflection enhancing the bass in the room - wave superposition effects

[crenca]

1 hour ago, semente said:

 

It's possible. I also have a bit of trouble hearing soundstage in live performances (of classical music) with my eyes closed.

 

R

 

Ha, know what you mean.  I have more than a few classical recordings that have a much larger "soundstage" than I ever hear at a live perfomance.  However, I am not a front row kind of guy, so maybe you have to be right up their more.

[mansr]

1 minute ago, crenca said:

Ha, know what you mean.  I have more than a few classical recordings that have a much larger "soundstage" than I ever hear at a live perfomance.  However, I am not a front row kind of guy, so maybe you have to be right up their more.

There's a reason the first few rows are often cheaper.