Dynamic Range...?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Tried to ask this question in another tread ( http://www.diyaudio.com/forums/showthread.php?postid=865346#post865346 ) but never got any answers regarding this.
(I also tried to search but I didn' find anything relevant info regarding this.)

Would I hear a lot of difference if I change an AD-converter with 107dB dynamic range (A-veighted) with one that has only 94dB...? :confused:
(The prior one is obsolete, and I am in big need of a replacement ADC)

Regards
Lyra
 
I don't have answers. Only questions.

How about the rest of your system?

Do you expect to get 94 db out of a CD?

Last I looked an analogue tape has about 65 db. Since all recordings are transfered to analogue tape and then back to a digital storage format, anything beyond 65 db should be smoke screens and mirrors. Welcome to the sideshow act known as "high-end audio." P.T. Barnum smiles.
 
phn said:
I don't have answers. Only questions.

How about the rest of your system?

Do you expect to get 94 db out of a CD?

Last I looked an analogue tape has about 65 db. Since all recordings are transfered to analogue tape and then back to a digital storage format, anything beyond 65 db should be smoke screens and mirrors. Welcome to the sideshow act known as "high-end audio." P.T. Barnum smiles.


:) I'm not sure if this nessecarily are relevant here...this equipment (Lexicon PCM90...the predecessor for this: http://www.lexiconpro.com/pcm91/details.asp ) is meant used on live productions etc., and I don't know if I need more headroom here...but nothing would be better than if this doesn't matter...

Lyra
 
phn said:
Last I looked an analogue tape has about 65 db. Since all recordings are transfered to analogue tape and then back to a digital storage format, anything beyond 65 db should be smoke screens and mirrors. Welcome to the sideshow act known as "high-end audio." P.T. Barnum smiles.

Come on. That's for normal cassatte tapes. Good studio tapes probably have some more dynamic. Also, lots (if not most) of studio's can do digital recording and mixing by now.

Still, if you find modern recordings with more than 30 dB of dynamics, you'll be glad :whazzat: . It's really is something to cry about :bawling:

like this:

http://ccrma.stanford.edu/~kglee/m420/hw7/fig7_4c.png

Edit: Image is a bit big, link if better ;)
 
4real, any studio can. But it sounds like cr*p. I wouldn't be surprised if my two Groundhogs CDs were culled from a digital master. It doesn't sound very nice. The problem with digital masters is that they have too much detail.

So what does an analogue master tape have? 70 db? More?

Indeed, modern mass-market CDs are compressed too high. They aren't even stereo. They are basically dual mono, using two almost identical channels. At least they will sound "good" on that old mono FM or AM receiver.
 
phn said:
The problem with digital masters is that they have to much detail.

How can you have to much detail :confused:

So what does an analogue master tape have? 70 db? More? [/B][/QUOTE]


I think even a few dB more, but that still doesn't say anything about the recored audio of course. It's not more than a tecnical spec, just like with the DAC.
 
How can the low-resolution transformers of the SET result in clear and detailed highs? How come audioholics' measurements of vinyl show close to 90 db when people say vinyl doesn't have that kind of dynamics?

I know, answering a question with a question isn't very bright. But I just don't have THE answer.

All I know is what people that work with these things tell me. I have never transfered a digital master to an analogue tape in my life.
 
Recording dynamic range

There may still be some people using tape for mastering but at the AES and NAMM shows there is no tape in sight. Just 24 bit digital which has 144 db of dynamic range. Here in LA, other than a few boutique houses, I don't know anyone who uses tape any more. Even my technicians with home recording studios are using 96KHz digital at 24 bits.

Then it gets down sampled to 44.1K and 16 bit depth for audio CD.

Only releases of old material done before digital became more or less standard are mastered from tape. Because there is no digital master though it is usual practice to digitize the tapes as well and there are ways to recover some signal down in the noise floor though these are of questionable value


You get 6db of dynamic range for each bit in a sample. Thus 16bits in an audio CD, gives you 96db of dynamic range.

When doing 24bit work, is is possible to get less than 144db if the A/D and D/A converters are not up to the task. Many converters only have analog sections capable of 100 to maybe 115db of dynamic range. Getting it perfectly quiet down to the last of the 24 bits is very difficult as you are getting down to extremely low noise levels.

Going to say a 96db converter from a 110db converter basically means a little more grunge at very low listening levels and a little more low level hiss.


Do CDs made from analog masters sound different than all digital ones. Yes, they do but that is not the subject at hand in this thread.
 
Yes, they do but that is not the subject at hand in this thread.

The thread started to drift because of my first reply. For that I'm sorry.

The CD might be 96 db, but only on paper.

If recording studios don't use tape anymore, then it seems I guessed right--the Groundhogs CDs were culled from digital masters and they, subsequently, suck. All the more reason to stay with vinyl.
 
Topic Drift - Tube Amps

Tube amps can sound good because the spectral balance of the distortions is different than on other amplifiers. Some distortions actually sound good to the ear. After 30 years in the business I still forget if it is the even order harmonics or the odd order ones that sound good and the others sound bad.

However, that "warm" sound we remember from tube amps that measured so bad is because of that different balance in the ratios of the different orders of the harmonics.

This is why, as I think everyone in the DIY audio field knows, measurements are not the final determining factor whether something sounds good or bad. While things that measure good generally sound good, to a point, and things that measure bad generally sound bad, this cannot be held to be an absolute and sometimes you get good measuring stuff that sucks and bad measuring stuff that sounds great.

The human brain, in processing what the ear picks up, is in this area as well as in many other things, capable of being illogical.

By the way, the tube amps do their cancallation of the nasty harmonics through having an output transformer. In essence tube amps have a balanced output stage and in the output transformer you get cancellations of some things and re-inforcements of others. Some transistor amps used to have output transformers but because of the large cost of output transformers, they disappeared in the 60s. And just putting a transformer on the output of a transistor amp is no help.

The overall configuration of the amplifier circuitry is totally different. However, if someone is willing to pay the added cost of the transformers, the configuration for a balanced output with an output transformer is well known and perfectly possible to build. Either at tube voltages, using existing tube output transformers, or at low voltage at a transistor friendly impedance level.

I could probably throw together a better schematic in a few hours than the one I have now as my 15W PA amplifier posted in the Audio section at my web site, www.schematicsforfree.mattsoft.net

Don't build that one. It has no feedback and the sound quality will suck.
 
True.

Sorry for unintentinally having hijacked the thread.

I do think there's a difference. That is, what we want on the recording end isn't necessarily what we want on the listening end. Microphones don't work like the human ear any more than a camera lens works like an eye. A lens has very limited focus, unlike our eyes. Same with microphones. What the latter means in reality I don't know. But I'm not the one to jump "hi-rez" bandwagon just because the marketers tell me there's where the action is. I do believe in measurements. But I also know what I like when I hear it.
 
Don't worry, I know. It's the even harmonics that appear to sound good btw ;) I don't like it, and so do many others (and just as many do like it I guess).

But to much detail, as descibed above, is not possible. You cannot have more then there is. If you persive to much, then there is most likeley something else wrong.

LP's often sound better because the mixes are far better then those on CD. To bad though :(
 
Lyra said:
Tried to ask this question in another tread ( http://www.diyaudio.com/forums/showthread.php?postid=865346#post865346 ) but never got any answers regarding this.
(I also tried to search but I didn' find anything relevant info regarding this.)

Would I hear a lot of difference if I change an AD-converter with 107dB dynamic range (A-veighted) with one that has only 94dB...? :confused:
(The prior one is obsolete, and I am in big need of a replacement ADC)

Regards
Lyra


I assume you referring to the performance of the CS5330A against that of the CS5389. A point or two to note.
The CS5389 was a flagship device and the CS5330A is a low end device. I would have thought it best to the replace the CS5389 with something of a similar ranking.
The CS5389 has differential inputs. The CS5330A does not. You will have to add differential to single ended converters to your board.
If it is your own PCM90 you are repairing then it probably doesn't matter what you stick in there but if it was mine and I sent it in for repair, I would not be best pleased to find my 3.0 V8 had turned into 1.0 straight 4.
 
Re: Recording dynamic range

dmfraser said:
....Going to say a 96db converter from a 110db converter basically means a little more grunge at very low listening levels and a little more low level hiss.

Thanks :)
If this is true...then this was really good news....!
In the environment this reverb usually are used there are never low listenings levels...well allmost never :wave2s: :hypno2:

Lyra
 
Re: Re: Dynamic Range...?

rfbrw said:



I assume you referring to the performance of the CS5330A against that of the CS5389. A point or two to note.
The CS5389 was a flagship device and the CS5330A is a low end device. I would have thought it best to the replace the CS5389 with something of a similar ranking.
The CS5389 has differential inputs. The CS5330A does not. You will have to add differential to single ended converters to your board.
If it is your own PCM90 you are repairing then it probably doesn't matter what you stick in there but if it was mine and I sent it in for repair, I would not be best pleased to find my 3.0 V8 had turned into 1.0 straight 4.

Thank You for all the reply's !

-It's not mine privately, it's my firm's.
-There are none AD-converters available...at least as far as I can find that could do this job better!?...18-bit is a REALLY obsolete standard !?
(Think I could use a 16-bit converter together with some glue logic (something like this: http://www.geocities.com/nonospcm1704/ but with a "combined LR-serial out) to make the signal fit the 18-bit requirement of the rest of the box, but I can't find any 16-bit ADC's either that has any better dynamic range, and by using an 20, or 24-bit's ADC I think I just unleashes a LOT! of other problems, when trying to make this <word> fit into 18-bit.)
-I am aware of the non balanced inputs, but this is of no major consern to me beacause the balanced signal is first made unbalanced, an then made balanced again just before it meets the ADC. Don't ask me why they have done it this way :confused: ...but this is how it is :)

PS: If some of You guys (...and girls, if there are some in here :) ) know where I can get my hands on some of the "original" CS5389-KP, would You please give me ha hint !?

Lyra
 
Re: Re: Re: Dynamic Range...?

Lyra said:


Thank You for all the reply's !

-It's not mine privately, it's my firm's.
-There are none AD-converters available...at least as far as I can find that could do this job better!?...18-bit is a REALLY obsolete standard !?
(Think I could use a 16-bit converter together with some glue logic (something like this: http://www.geocities.com/nonospcm1704/ but with a "combined LR-serial out) to make the signal fit the 18-bit requirement of the rest of the box, but I can't find any 16-bit ADC's either that has any better dynamic range, and by using an 20, or 24-bit's ADC I think I just unleashes a LOT! of other problems, when trying to make this <word> fit into 18-bit.)

As I have made clear before, if the data from the CS5389 is I2S or Left/MSB- justified, then using a greater wordlength creates no problems other than the negative effects of truncation. It is a straight forward replacement. No logic, glue or otherwise, required.
As for the effects of truncation, I have suggested the still current but not recommended for new designs, CS5396. It offers multiple wordlength options all properly dithered.
The TAS3103A as per macboy's suggestion is even more readily available than the CS5396 and only costs around 8 euros. A CS5361 will cost about the same. Combine the two and away you go.
Both of the above suggestions offer a solution more in keeping with a device like the CS5389.



I am aware of the non balanced inputs, but this is of no major consern to me beacause the balanced signal is first made unbalanced, an then made balanced again just before it meets the ADC. Don't ask me why they have done it this way :confused: ...but this is how it is :)

Lyra

Unless it is done that way for gain control and ADC input filtering.
 
My thoughts was that a truncaded 20->18bit signal could! be worse than a "low end 18-bit ADC", or a 16-bit signal...!???
As far as I know (they told me...) the engineers at Lexicon have tried to replace CS5389 with CS5390...the closest chip available today with no luck...
CS5390 is 20-bit...and pin to pin compatible with CS5389 as far as I know.

When it comes to CS5396: what a £$&ck is psychoacoustic output ???...don't understand this part.....yet...:confused:

As far as TAS3103A concerned, I just might try this appoach....
A lot of unneccesary functions though...but can be matched with whatever ADC I want to choose...big advantage !!:D
....does there exist a smaller:smash: chip like TAS3103A, but with not 4, but 1 signal input and fewer options besides converting the signal ?

Lyra
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.