Jump to content

DAC vs. Stats -- FIGHT


teufelshunde

Recommended Posts

I think 24/192 is a nice to have, at least given the amount of music MOST people likely have in that format. I have two albums in 24/192, nothing else better than 24/96. I copied some hi-rez to a computer I have that has a cheap Diamond sound card that does up to 24/192 out on coax. Using KS or ASIO, they DID sound damn nice. Getting a nice sound card that offers this is a possibility (for laptops, not sure if there is a cardbus or PCMCIA card that does this).

It's one more piece in the chain, but I'd also consider a USB->S/PDIF converter that handles 24/192, and not limit the DAC's I seek otherwise.

Link to comment
Share on other sites

  • Replies 69
  • Created
  • Last Reply

Top Posters In This Topic

whacha coin with anova beefy?

Pretty standard biomedical stuff. Cells with a couple of different treatments, results measured for several parameters and compared to each other. Usually just a bog standard 1-way ANOVA, Dunnett post tests if we have a specific control in mind, Bonferroni if we don't.

I can't see how Bayesian statistics could even remotely be applicable to experiments like this, but then again, I hate stats. My own personal statistical test is: does it stand out like dogs balls? p=yes/no.

Link to comment
Share on other sites

For the OP, if your music is FLAC, then spend $1000 and buy an Oppo BDP-95. You can connect your drive to the unit and a monitor, no computer needed. I believe it's better than the Amarra 4 I had before. It also ays BR, DVD-A, SACD and redbook.

Mike - I've already got an Oppo BDP-95, but it's firmly ensconced in the family room as part of the Home Theatre setup.

I need something for the office where my desktop and NAS are located.

NAD M51?

I think Dinny's on to something here.

Just wondering how much of the M51 price tag is a premium being paid to the HDCP police for HDMI support.

Link to comment
Share on other sites

...

I think Dinny's on to something here.

Just wondering how much of the M51 price tag is a premium being paid to the HDCP police for HDMI support.

Not sure as the HDMI is limited to stereo (or maybe 2.1) tracks.

For the OP, if your music is FLAC, then spend $1000 and buy an Oppo BDP-95. You can connect your drive to the unit and a monitor, no computer needed. I believe it's better than the Amarra 4 I had before. It also ays BR, DVD-A, SACD and redbook.

Interesting Mike. I have decided not to go the Amarra 4 upgrade route. The problem I have with the Oppo is it does not have an SPDIF input.

Edited by morphsci
Link to comment
Share on other sites

The problem with an arbitrary significance level is you are abdicating your respondibility to determine FOR YOUR PARTICULAR EXPERIMENT what a significant result is. This is both a statistical and field dependent process that is too complex to be adequately assessed by a simple apriori alpha (if nothing else it ignores beta and especially in medical trials or tests the risk to the patient of an alpha or beta error is not symmetric). So the preferred methodology in the fields of biology I currently work in is to report your exact p value for your test(s) and then discuss that value in reference to your biological hypothesis. Stating that your value is significant/not significant does not add any additional information.

If you performed two experiments that resulted in the following two p values, exp. 1: p = 0.049, exp. 2: p = 0.051 what would you conclude?

Link to comment
Share on other sites

I can't see how Bayesian statistics could even remotely be applicable to experiments like this, but then again, I hate stats. My own personal statistical test is: does it stand out like dogs balls? p=yes/no.

Do you treat your effects as fixed or random? Many biological effects should be treated as random and thus you have entered, willingly or not, the realm of Bayesian statistics. I do not align myself with either camp BTW.

Link to comment
Share on other sites

Well through this argument I have read more about hypothesis testing than I had ever before and I think I am going to have to learn some more for an upcoming paper. Any suggestions as to reading? I am assuming wikipedia probably isn't the best.

Well I do not know if this will be strictly relevant for you but the first edition of this was what led me to start questioning typical Popperian/Fisherian hypothesis testing methodologies.

http://www.amazon.com/Model-Selection-Multi-Model-Inference-Information-Theoretic/dp/1441929738/ref=sr_1_1?ie=UTF8&qid=1329078313&sr=8-1

Link to comment
Share on other sites

If you performed two experiments that resulted in the following two p values, exp. 1: p = 0.049, exp. 2: p = 0.051 what would you conclude?

Whether it is right or not, physiology universally uses significance at p<0.05. Anything just above 0.05 is usually fixed with more 'n'.

Link to comment
Share on other sites

p = 0.051 = "Shit, I should have done a power analysis before data collection!"

I haven't read the Burnam and Anderson book entirely, but I'm probably the only one studying biology who hasn't - seems that way sometimes anyway.

Still, it all depends on the study. Some of the most interesting stories in biology come down to a simple "does this have an effect or not?" question, where traditional hypothesis testing is entirely appropriate. I do, however, agree with morphsci's point about stating the p-value, rather than just stating that it was > or < 0.05.

Edited by acidbasement
Link to comment
Share on other sites

... Anything just above 0.05 is usually fixed with more 'n'.

Exactly why it has no real meaning, other than to make you feel good.

p = 0.051 = "Shit, I should have done a power analysis before data collection!"

My conclusion is that the results of both experiments are probably the same and "Shit, I should have done a power analysis before data collection!"

Still, it all depends on the study. Some of the most interesting stories in biology come down to a simple "does this have an effect or not?" question, where traditional hypothesis testing is entirely appropriate.

Yep. Which is why I am not a strict Frequentist, Bayesian or Model selectionist.

Link to comment
Share on other sites

Based on the recommendation of our local Godfather of DACses (Dinny) - I'm going w the NAD M51.

If it sucks, I'll send it back.

Plus Audio Advisor is recommending the Cardas Clear Light Balanced XLR interconnects - also going back if they don't deliver the sonic goodness.

Edited by teufelshunde
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.