1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

The dreaded DAW question revisited!

Discussion in 'Recording' started by John Stafford, Jul 11, 2005.

  1. John Stafford

    John Stafford Well-Known Member

    Okay, this is something I've found quite interesting.

    I've recently been recording into Sonar and Sampitude, but monitoring through the DAC-1 (via the recording software).
    I use AKG K271 headphones that provide good isolation at higher frequencies.

    When I'm monitoring, there is more high frequency than when playing back. I've used a U87 in omni mode because of the extremely hyped top end, in order to make the sound as abrasive as possible, yet when I play back things sound more subdued. Blatantly obvious positioning differences at the recording stage don't seem so apparent on playback.

    I know this is impossible, as I'm monitoring the digital signal, which should be exactly the same as what goes to disk.

    I'd be very interested to hear if anyone else notices the same thing happening, or if I'm going quietly insane.

    One other thing, while the DAC-1 is a wonderful box, and I really like it, I've noticed that the differences in my mics are less apparent through it than when using my previous setup. I think there is a plausible explanation in that the DAC-1 can probably handle asymmetrical waveforms that can freak lesser converters out, and for some reason there are more of these produced by some of my mics than others -which I can verify in software. Still this is only a partial explanation, and I would love to hear what others have to say on this matter.

    Maybe my mother dropped a lot of acid before I was born -It was the sixites after all :wink:
    Cheers
    John
     
  2. Zilla

    Zilla Active Member

    You do not know that it is impossible. This is digital theory, and theory only. What you do know is what your ears are informing: that there is a sonic difference. In this particular situation, the differences are equivalent to monitoring a tape machine on input vs. playback. What is going through the tape machine electronics is different than what is captured on tape. Likewise, in your daw the audio data that is routed through the software/card drivers is (apparently) handled differently than how the data pulled of the HD is handled. I don't think any DAWs have the equivalent of a confidence head.

    Many people have claimed that the DAC-1 makes their digital equipment sound more similar, supposedly because the unit's jitter correction/upsampling reduces errors that otherwise skew the sound. My feeling is that it is just not that transparent a unit, and that it is homogenizing and masking sound differences. I have found this quality is consistent in devices with 5532's implemented in their design, like the DAC-1. That being said it is a good performer considering the price, and perfectly fine recordings can certainly be made with it.
     
  3. Thomas W. Bethel

    Thomas W. Bethel Well-Known Member

    I did some really intensive listening test when we were thinking about buying our DAC-1 (by the way there is a good review of the unit in this month's (July) Sound on Sound) I listened to about 12 different D to A converters and came to the conclusion that NONE of them sounded the same and that each had its own "sound" The one that I liked the best was the DAC-1. From my earliest days in recording and mastering I have been looking for a device to enable me to hear EXACTLY what the recording engineer was hearing when he recorded the tracks or what another mastering engineer did to the tracks. I was looking for a transparent device and so far the DAC-1 is about as close as I have come to a transparent device. I use to have a couple of Neotek Series I consoles. The are and were the most musical sounding console I have ever heard and or used and the were loaded with 5532s. I understand that some people are "hot rodding" their DAC-1 with different op amps and a beefed up power supply. There is a fellow around here who I trust and his engineering skills and ears are first class. I may ask him to look into the possiblity of upgrading one of our units with the Mods I have seen done and then compare the two.

    Hope this helps
     
  4. Zilla

    Zilla Active Member

    I am not stating that 5532's are unmusical. I am not even saying that I have a general negative opinion of them, per say. My comment was really focused on one characteristic: transparency. In comparison with some of today's hi performance amps (in a simular circuit) the 5532's are much less revealing. I find that they tend to be on the grainy side, which makes it harder for low-level detail to speak easily.
     
  5. Thomas W. Bethel

    Thomas W. Bethel Well-Known Member

    So what are your favorite OP AMPS? for today's electronics?
     
  6. Zilla

    Zilla Active Member

    As with any design problem, knowing your goal is key. I could rattle off a good number of devices, but this would be meaningless without a context; without a stated goal.

    If transparency is most significant, then my favorite op-amps tend to be discrete designs. Being able to choose individual op-amp component elements allows optimization for a given design goal. This has big advantages over integrated circuits (IC) on a single die.

    But if we are limited to IC's, one nice thing is that you can pop similar devices in a out of a circuit. This easy experimentation allows you to measure, listen and choose to your liking. Its probable that swapping out the 5532 in the DAC-1 may make it more revealing, but there is no guarantee. By the way, I hope that I am not giving the impression that I think poorly of the DAC-1. I think it is a brilliant piece of gear. You would have to spend a lot more money to beat its performance and elegant packaging. But at that cost point and packaging size, one can expect some limitations. It just isn't going to be as revealing as some higher-end equipment, like a Lavry gold, for example.
     
  7. John Stafford

    John Stafford Well-Known Member

    Zilla,
    First of all thank you for your reply. I must admit that I'm quite surprised to have noticed this difference. If I remember correctly, I was given some information by Echo Audio about the way Windows handles audio data, and that it can play around with data unless it is told not to. It can upsample and downsample and perform various other conversions unless it is told not to. I have a little PCMCIA card made by Echo with a driver that allows you to tell Windows not to mess with the data. Maybe this is some sort of clue?

    So I'm not the only one who has come to this conclusion! I'm glad to know this. I was very surprised when I listened to some of my recordings expecting to hear huge differences between my mics. Your explanation about the 5532 is fascinating. The more I learn, the more I realise how little I know!

    John
     
  8. John Stafford

    John Stafford Well-Known Member

    Yes indeed this is helpful. I have no doubt that I chose the right box -just think of the competition! Of course the performance for the price is amazing. Having said all that, the thought of super-charging the box is tantalizing to say the least. I think this is something I will have to look into. I'd love to hear an improved version -that would be a very interesting prospect!

    If you have any mods done PLEASE let us know. I still love my DAC-1 :wink:

    John
     
  9. Thomas W. Bethel

    Thomas W. Bethel Well-Known Member

    For a long time I have noticed that certain DAW software sounded different. If I playback something in Wavelab and then switch to Sound Forge (using exactly the same settings and exactly the same monitoring setup) I noticed that the signal sounds slightly different (not better not worse just different.) I have also noticed this with other DAW programs. Now logic dictates that 1s are 1s and 0s are 0s and that everything in the digital world of computers should sound the same but I do hear a difference. Why? I have NO idea.

    Maybe someone who is more into the programming of DAWs can answer the question.

    Until then I will just assume that it is a "normal" occurance and one I should be aware of but not concerned with.
     
  10. DavidSpearritt

    DavidSpearritt Well-Known Member

    Here we go again. Its simple to test. Capture the stream coming out of both programs, same WAV, same sound. There is no plausible alternative. You simply have to construct a tight experiment and stop wondering. There must be some settings different, unknown, not thought about, not checked.
     
  11. Cucco

    Cucco Distinguished Member

    Interestingly enough, I tried a bit more experimentation today.

    I used wavelab to sum to mono one of the demo tracks from the Mytek website. It is a 24/96 stereo recording of a piano.

    I summed it to mono without dither, conversion, etc.

    I did the same in Cubase and in Sequoia.

    The Wavelab and Sequoia are identical. The Cubase was not. This tells me that either Cubase is using some funky means of summing (putting something in the bus that I didn't ask for, etc.) or it simply is applying some type of dither, etc to the signal before or during processing.

    FWIW, the remaining audio from the Cubase version was quite polarized containing mainly extremely low frequencies (below 80 Hz) or extremely high frequencies (above 5 kHz). There were other frequencies mixed in, but the majority were in these outer bands. This leads me more to the dither conclusion. (applying dither where it need not be.) Oh, and yes, I checked to make sure that I did not inadvertantly turn on any type of noise-shaping, etc.

    J.
     
  12. FifthCircle

    FifthCircle Well-Known Member

    Ugh... Not this subject again...

    In our last go-around, I asked the developers of the Samplitude and Sequoia DAWs this exact question. If everything is 32 bit float, what accounts for the differences that folks hear?

    Basically the answer was that there are certain things in the math that are absolute. However, there are ways of dealing with errors and rounding that is not absolute and there is a certain amount of voodoo to it. These differences appear much more substantial with processing with plugins than with just basic summing.

    --Ben
     
  13. DavidSpearritt

    DavidSpearritt Well-Known Member

    If there was "voodoo" that really made a difference, everyone would be onto it. Also, someone would have experimented and documented it. Some of the firms are playing on customer ignorance by stating that their program has a certain je ne ce quoi therefore it sounds better.

    Don't worry I have finished with this topic. :)
     

Share This Page