AudioSmile Challange - Can you detect the AudioSmile DEQ?

Which Song Samples are passed through the modified DEQ?

  • X, Y, Z

    Votes: 1 14.3%

  • XX, Y, Z

    Votes: 1 14.3%

  • X, YY, Z

    Votes: 1 14.3%

  • XX, YY, Z

    Votes: 1 14.3%

  • X, YY, ZZ

    Votes: 0 0.0%

  • XX, YY, ZZ

    Votes: 0 0.0%

  • X, Y, ZZ

    Votes: 1 14.3%

  • XX, Y, ZZ

    Votes: 2 28.6%

  • I cannot hear a difference.

    Votes: 0 0.0%
  • Total voters
    7
Tenson

Tenson

Joined
Nov 12, 2003
Messages
5,947
Reaction score
1
Location
Kent, UK
Given that this has proven a bit confusing and messy, Simon has restarted the test using simple ABX and one test track.

So please don't download or vote on this thread.

Go here: https://www.audio-forums.com/as-rediect/showthread.php?t=23393


----------------------------------------------------------------------------------------------------------------------------------------------

Introduction:

Firstly let me say I want this to be fun for everyone! :)

As AudioSmile, I've been modifying the Behringer DEQ for a long time, and regulars on this forum will be sick of me annoying you all about it! I've often said it is as much DAC or ADC as anyone will need, regardless of the level of their system. I believe it to be extremely transparent, but some remain to be convinced. Fair enough, here is your chance to listen for yourself!


How I made the test:

To create this test I have taken 3 sample songs and ripped them bit-for-bit to a wav file. I have then cut out a 2 minute section of each song so as not to annoy copyright people.

One sample for each song is just an accurate direct rip of the CD data. If you play this file on your home system it should sound exactly as the original CD does. In fact I'd suggest burning all these files to a CD-R so you can play them on your Cd player and main system if your computer sound system isn't up to much.

The other sample for each song uses an AudioSmile unit as a DAC to convert the data to an analog signal. That is then taken in to the analog inputs of another modified unit to be converted in to digital again. This is then recorded on the computer. So, this sample is not just going through the DAC, but the ADC as well, which means there is twice the level of conversion going on compared with using the unit just as a DAC or as an ADC. If the AudioSmile DEQ is adding any character of its own, it should really be audible here.

Oh, and the DEQ was operating at 44.1KHz, so I could playback and record from the computer at the same time. The DEQ can work at up to 96KHz.

The challenge:

Can you detect which sample of each song is the direct accurate data rip, and which one has gone through two layers of conversion on the modified DAC / ADC?

I have made the poll hide the names of those who take part, so don't be embarrassed to cast your vote. However, if you are pretty confident by all means post your opinion in the thread.


The Music:

Lynyrd Skynyrd - Sweet Home Alabama

Sample-X
http://www.mediafire.com/?glzqyk5eifz

Sample-XX
http://www.mediafire.com/?jujgtzx4m20


Lamb - Zero

Sample-Y
http://www.mediafire.com/?3ndim5xzdnm

Sample-YY
http://www.mediafire.com/?om3yyt2ynxy


Usher Demo CD - Track 10 (anybody know it's name?)

Sample-Z
http://www.mediafire.com/?j4zl2jyhh1w

Sample-ZZ
http://www.mediafire.com/?anygngaq32y


Edit: I forgot to add a 'I can't hear a difference' option on the poll. Can a mod add that, please?
 
Last edited by a moderator:
Hi Tenson

I know you mean well but there is no such thing as 'ripped them bit-for-bit to a wav file' and 'direct accurate data rip'. Even using EAC which reads the file about 8 times before 'selecting' the best one, you'll get errors. You will get errors in a transfer no matter what you do. There is no such thing as a perfect transfer.

I'll try and listen to the files this weekend.

Rob
 
sq-

it's pretty much known that you can't make a bit perfect transfer in the digital domain. If you could, digital error correction wouldn't exist. And digital error correction isn't perfect which is why they are constantly being modified.
 
Oh and to add to that last statement....

the fact that a disc has to spin at a high rate means 'errors'. Every time you read the disc the errors look different. Digital is not as perfect - this is planet earth.
 
nothing is perfect nor will ever be, from the days when i worked in recording studios most **** ups were done by egeneers being high on drugs therefore loosing high frquecies and screwing the recordings, we can all play with equipment and as to what we think is the best, but what is good for you may be awewfull to me, and likewise,
nando.
 
Last edited by a moderator:
come on chaps ...don't be childish

little children say the c word just to get attention

tenson has gone to the time trouble expense of putting this together ....
the least you can do is have a listen

if you don't agree with his methodology then give a cogent critic...

if you wish to persuade people to go elsewhere then continue in your infantile pooh throwing ...

remember young children can read these posts too ...why not use a brain cell or two before you knee jerk your abuse ?

just remember one day you might just pop by to sling your load and the forum won't be here.

Oh yes ................. I don't bother with these comparatives as they have to be played through my poor computer sound system ...I can't be arsed to transfer to rom or other media, so don't think I'm just sticking up for the poster !
 
I thought you had a Squeezebox?

Bit accurate rips of CDs are obviously possible. A CD reader knows if it hasn't got a correct read, good rippers directly use this information.

Paul
 
Can you tell the difference when you don't know and can't see what it is you're listening to?
I can't.
 
I feel it is time to recount the old Ivor Tiefenbrun chestnut, wasn't he unable to to tell the difference between his own turntable and the PCM copy, and that was with one of the first Sony machines.
 
Paul Ranson said:
I thought you had a Squeezebox?

Bit accurate rips of CDs are obviously possible. A CD reader knows if it hasn't got a correct read, good rippers directly use this information.

Paul
Click to expand...

me ?

yes not attached to same puter .... or web or hifi
 
robM said:
Hi Tenson

I know you mean well but there is no such thing as 'ripped them bit-for-bit to a wav file' and 'direct accurate data rip'. Even using EAC which reads the file about 8 times before 'selecting' the best one, you'll get errors. You will get errors in a transfer no matter what you do. There is no such thing as a perfect transfer.

I'll try and listen to the files this weekend.

Rob
Click to expand...

Hi Rob,

If that is true, and I don't believe it is, it shouldn't change the effectiveness of this game! Playing the direct rip file should still sound just like playing the original CD on your home system. If nothing can make a perfect read, then your home system won't either, and my EAC rip is likely to be closer to the original data than any CD transport. The comparison still holds, so come on, get cracking! :)

I think it is possible to make an accurate rip because EAC can CRC check the ripped data against an internet database. Thus I do at least know my data is the same as that which the majority of other people are getting from their CDs. Also, I have in the past tried taking numerous digital transfers from my CD transport (Cyrus) and doing null tests. They always cancel perfectly, and I mean perfectly bit-for-bit. So the data read is highly consistent even from that, so I'm sure EAC can manage.
 
robM said:
Oh and to add to that last statement....

the fact that a disc has to spin at a high rate means 'errors'. Every time you read the disc the errors look different. Digital is not as perfect - this is planet earth.
Click to expand...

If you check back a few months you can see two CDs containing the same music playing back via the same (20 year old) Meridian transport. To cut a long thread short, we got a perfect null from them when inverting/combining the files produced. Where are the errors?
 
Hi Tenson

I understand you and I will give it a go.

The AES set a copy limit for transferring masters back about 8 years ago, this was for archiving purposes and is still common practice in backing-up digital. They said you could transfer up to 6 or 8 times before the 'generation losses' were noticeable.
This process is for digital domain only transfers.

You can try it yourself. Transfer to CD. Play that copy and make a copy of that and continue the cycle. It's worse if you add an analogue stage to the copy as analogue noise doubles as well. At some point the copy of the copy will not actually play but before that point the errors on the copy will be large enough as to make that copy useless.

So, although you can make what you think is a perfect copy, logically a copy cannot be exactly the same as the original.

BUT I will give your test a go.
 
Hi Robholt

excerp from a friends company website:

As a first line of defense, CD players employ a sophisticated error correction system designed to accurately restore the missing data and enable exact reproduction of the recorded signal. The system uses redundant information on the disc to reconstruct the lost content. If not for the error correction circuitry, the player would never be able to reproduce the music without sound distortion or alteration -- even when playing new, seemingly perfect discs.

The problem here is that not all players and transports are created the same as we already know. The old Meriadian 206 transport is a lot better than the transport section of most computers because the computer drive is not an 'audio specific' drive and that is forgetting the power supply and vibration problems in a computer which may/ does cause the error rates to increase.

Audio specific drives are very rare now as most companies use CD computer drives. The old Naim CDI used an audio specific drive and Creek are about to come out with (or have done) an audio specific drive. Different drives can give very different bit error rates (BER) due to disk wobble, and mechanical problems.
Yes, the correction systems corrects for these 'defects' but some are better than others, so you see no 2 can be the same.
 
Hi RobM,

I'm struggling to understand how this has any affect on the validity of the test, given that I already know the data I extracted from the CDs is the same as the majority of other users out there on the net.
 

Latest posts

Back
Top