Date   

Re: Can I please have some help with my speaking rate?

 

That should have been version and build numbers for Windows 10.  I'm using a new-to-me machine and still don't have the touchpad settings worked out yet and end up accidentally inserting stuff in random places if I accidentally move my insertion point.
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Is there a way to get more voices for NVDA?

Lexi
 

I was wondering, is there a way to get more voices for NVDA? Or am I just stuck with the same five voices? I was told to get more voices that I would have to pay for them. Is that true? If it isn't, how would I go by doing it? Please be blessed and be safe. Thank you,
Lexi 


Re: Can I please have some help with my speaking rate?

 
Edited

It would be very helpful to have additional information:

1.  Version of Windows and, if Windows 10, the Build number as well as whether it's Home or Pro.
2.  The version of NVDA, which is most likely 2020.4, but it still helps to know.
3.  The synthesizer and voice being used, which can be found in Speech settings:  NVDA+N,P,S then either down arrow once to speech, or hit an additional S to get there.  Tab over into the pane and look at what settings are set to, including whether the checkbox for Rate Boost is checked (and it probably isn't).

I use only OneCore voices, and have never had an issue with slowness.  That being said, the rate at which many blind users are accustomed to listening to speech makes my head spin, and a rate typical of normal spoken speech is considered very slow indeed by many.
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

 

Hi,

I see. To check which speech synthesizer you are using, press Control+Insert+S, then listen carefully to what NVDA says.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of Lexi
Sent: Sunday, May 16, 2021 8:04 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Can I please have some help with my speaking rate?

 

 

 

I only found a few voices when I went into the NVDA menu. Around four or five of them actually. I chose the Hazel voice. I haven't tried that. Please be blessed and be safe. Thank you,

Lexi 

On Sun, May 16, 2021 at 9:59 PM Joseph Lee <joseph.lee22590@...> wrote:

Hi,

I think I know the answer, but are you perhaps using the default voice that comes with Windows? If so (Windows OneCore), try enabling rate boost (NVDA menu/Preferences/Settings/Speech).

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of Lexi
Sent: Sunday, May 16, 2021 7:56 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Can I please have some help with my speaking rate?

 

 

 

No, I didn't even know that was a thing. I just started using NVDA a few days ago when my Jaws license expired. please be blessed and be safe. Thank you,

Lexi 

On Sun, May 16, 2021 at 9:49 PM Brian Vogel <britechguy@...> wrote:

The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

Gene
 

Slow speech isn’t related to the registry fixing tool.  It won’t do any harm to run it but it won’t help the problem.
 
Gene

-----Original Message-----
Sent: Sunday, May 16, 2021 10:00 PM
Subject: Re: [nvda] Can I please have some help with my speaking rate?
 
See Also:  The Most Basic Troubleshooting Steps for Suspected NVDA Issues

The keystrokes to run the COM Registration Fixing Tool:  NVDA+N,T,R followed by hitting Enter and following the prompts.  Do not let the prompts rattle you, as they can sound dire to the uninitiated, just respond in the affirmative and allow the tool to run.
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

Gene
 

If you are using One Core voices, I’ve seen some sort of discussion before about how you can speed them up.  I don’t recall what was said but if you open the voice dialog, control NVDA key v, and you tab around, do you get to a check box for rate boost?  Check it with the space bar than tab to and activate the ok button.  But before that, set your speech rate much lower than it is.  Then check the rate boost checkbox and activate the ok button.  I don’t know how fast your speech will be so I’m saying to slow it down a good deal first so you don’t have a rate that is so fast that you can’t understand it.  Perhaps twenty or thirty would be a good rate before turning on rate boost.  I believe some synthesizers don’t have rate boost but try it if yours does.
 
Gene

-----Original Message-----
From: Lexi
Sent: Sunday, May 16, 2021 9:56 PM
Subject: Re: [nvda] Can I please have some help with my speaking rate?
 
 
 
No, I didn't even know that was a thing. I just started using NVDA a few days ago when my Jaws license expired. please be blessed and be safe. Thank you,
Lexi
On Sun, May 16, 2021 at 9:49 PM Brian Vogel <britechguy@...> wrote:
The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

Lexi
 



I only found a few voices when I went into the NVDA menu. Around four or five of them actually. I chose the Hazel voice. I haven't tried that. Please be blessed and be safe. Thank you,
Lexi 

On Sun, May 16, 2021 at 9:59 PM Joseph Lee <joseph.lee22590@...> wrote:

Hi,

I think I know the answer, but are you perhaps using the default voice that comes with Windows? If so (Windows OneCore), try enabling rate boost (NVDA menu/Preferences/Settings/Speech).

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of Lexi
Sent: Sunday, May 16, 2021 7:56 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Can I please have some help with my speaking rate?

 

 

 

No, I didn't even know that was a thing. I just started using NVDA a few days ago when my Jaws license expired. please be blessed and be safe. Thank you,

Lexi 

On Sun, May 16, 2021 at 9:49 PM Brian Vogel <britechguy@...> wrote:

The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

 

See Also:  The Most Basic Troubleshooting Steps for Suspected NVDA Issues

The keystrokes to run the COM Registration Fixing Tool:  NVDA+N,T,R followed by hitting Enter and following the prompts.  Do not let the prompts rattle you, as they can sound dire to the uninitiated, just respond in the affirmative and allow the tool to run.
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

 

Hi,

I think I know the answer, but are you perhaps using the default voice that comes with Windows? If so (Windows OneCore), try enabling rate boost (NVDA menu/Preferences/Settings/Speech).

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of Lexi
Sent: Sunday, May 16, 2021 7:56 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] Can I please have some help with my speaking rate?

 

 

 

No, I didn't even know that was a thing. I just started using NVDA a few days ago when my Jaws license expired. please be blessed and be safe. Thank you,

Lexi 

On Sun, May 16, 2021 at 9:49 PM Brian Vogel <britechguy@...> wrote:

The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

Lexi
 



No, I didn't even know that was a thing. I just started using NVDA a few days ago when my Jaws license expired. please be blessed and be safe. Thank you,
Lexi 

On Sun, May 16, 2021 at 9:49 PM Brian Vogel <britechguy@...> wrote:
The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: Can I please have some help with my speaking rate?

 

The standard first question (or one of them):  Have you run the NVDA COM Registration Fixing Tool?
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Can I please have some help with my speaking rate?

Lexi
 

Hello everyone, may I please have some help adjusting my speech rate? NVDA says it's at 100%, but it's still speaking very slowly. Any help that you all can offer would be greatly appreciated. Thank you all in advance. please be blessed and be safe. Thank you,
Lexi


Re: NVDA running on a budget laptop

Gene
 

What sort of cases are there where noticeable improvements in NVDA might be made by making some of the changes discussed?  As I use it, delays are so slight that I haven’t even given them more than very passing thought.  And those are very slight delays on an eleven year old machine.  I don’t know if they are even noticeable now on today’s faster machines, I mean typical consumer machines.
 
Gene

-----Original Message-----
From: Joseph Lee
Sent: Sunday, May 16, 2021 8:25 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

Hi,

Cython might help to make NVDA code compile under a C++ compiler such as Microsoft Visual C++ compiler, but then it will raise questions such as dealing with C extensions and how to make NVDA’s own (Python) code more efficient. Transforming NVDA into a pure 64-bit program will require that we use 64-bit Python interpreter, but then we run into the same problem that 32-bit NVDA is facing, although in the opposite direction.

As for using CPU instructions such as Advanced Vector Extensions (AVX), it comes down to the workload to be run by Python interpreter. I imagine C++ components could benefit from that if Visual C++ compiler can be told to use AVX instructions, but then it depends on how NVDA performs its tasks internally (AVX is mostly used for calculation intensive tasks such as scientific computing, which is not really the use case for a screen reader unless the screen reader algorithms do require working with many things that require up to 512 bits to store many components). Ultimately, since NVDA is a Python-based screen reader, it will come down to if Python interpreter can even take advantage of newer instructions for executing bytecode on the spot.

More importantly, consider that not all CPU’s support AVX and newer instructions. It took Microsoft until 2012 to require that processors support SSE2 (ten years after its appearance). Although most processors support SIMD (single instruction, multiple data) commands, only some high-end processors support AVX, AVX2, and parts of AVX-512, and tech press articles (and according to some sources, Linus Torvalds) argue that AVX instructions are power-hungry (understandable as millions of transistors are dedicated to vector instructions).

Therefore, the standpoint of NV Access (as communicated by Quentin and others) and lead contributors (including soon to be former ones like myself) is that we should target the vast majority of configurations that are in use at a given time (that’s one of the reasons for delaying end of support for Windows 7; due to a critical issue with Python 3.8 and later, that decision is uncertain at this time); at this time, this means supporting a processor with at least SSE2.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 6:07 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

Hi Joseph,

Would cythonizing NVDA improve this situatuation though? Or making a native 64 bit version to talk to 64 bit apps?  Also, isn't one reason for the CPU usage NVDA not utilizing the newer processor  instruction sets being developed by Intel/AMD?  Are there any new processor extention sets that would make NVDA more responsive if used?

On 5/16/2021 7:21 PM, Joseph Lee wrote:

Hi,

Enes’s claim is understandable, considering that:

  1. Speech synthesizers do require resources such as CPU to translate text into speech using whatever rules manufacturers and users define.
  2. Most NVDA components run on top of a Python interpreter. This means in order to perform screen reading operations, Python interpreter must be willing to process screen reader instructions on top of housekeeping tasks such as garbage collection. Some NVDA components are written in C++ for faster performance and communication with apps, but still Python is invoked for housekeeping operations.
  3. NVDA must talk to many “people” i.e. API’s and apps at once. Although folks spent years optimizing accessibility API’s and communication between apps (in computer science, this whole thing is termed “inter-process communication”), communication with API’s and apps is still an expensive step that involves processors switching between running app and system codes. If it takes some time to send and receive messages between two 32-bit programs, imagine how long it will take for a 32-bit app such as NVDA to get useful info out of 64-bit apps (this involves formatting bits in a way that allows these two programs to eventually communicate through an intermediary called WoW64).

 

One important clarification: let us not equate screen readers to text-to-speech engines (don’t confuse between the two). Although TTS does contribute to screen reading performance (namely rules used to translate text into waveforms or hardware signals), what’s more important in this overall context (screen readers running on a class of hardware) is how efficient screen reading instructions can get, keeping in mind that NVDA is running on top of an interpreter. There are limits as to how power efficient an algorithm can become, as the overall limit is optimizations from the host (in NVDA’s case, not only optimizations from Windows, but also ongoing optimization work done by Python core developers and third-party library/C extension authors). Also, although not addressed here, you can’t just state that making screen readers multi-core aware will bring improvements – it depends on years of effort spent on parallelization and optimizing the screen reader to not only take advantage of at least two cores at once, but also to make itself more efficient when running as a multi-core aware application (power draw from a single core versus multiple cores does play a part here); I have learned the hard way that you will face more unexpected bugs when you make a program run with multiple threads and hardware cores.

Cheers,

Joseph

 

From: nvda@nvda.groups.io mailto:nvda@nvda.groups.io On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 4:58 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:

Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 

 

As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.

 

I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.

 

And I haven’t seen complaints about the performance of NVDA from people using tablets. 

 

Gene

-----Original Message-----

Sent: Sunday, May 16, 2021 7:20 AM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:

Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

 

Hi,

In the long run, pure 64-bit will be the norm, but then enterprises might argue that they still need 32-bit components for some tasks. But then the days of enterprises arguing over upgrade timeline is slowly coming to a close, as evidenced by the fact that there are reports of parts of industries coming to a halt because they didn’t listen to Microsoft’s warnings about outdated Exchange Server installations. The next phase is securing the cloud, and I can foresee a time when all users need to purchase is a thin client with screen readers being streamed from the cloud.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 6:57 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

Hi Joseph,

The 32 bit  problem would be less though I would think. As time has gone on, more and more apps are pure 64 bit, Microsoft is also dropping support for 32 bit windows as well. Most recently, Firefox, Thunderbird, and Adobe Reader have gone pure 64 bit.

On 5/16/2021 8:25 PM, Joseph Lee wrote:

Hi,

Cython might help to make NVDA code compile under a C++ compiler such as Microsoft Visual C++ compiler, but then it will raise questions such as dealing with C extensions and how to make NVDA’s own (Python) code more efficient. Transforming NVDA into a pure 64-bit program will require that we use 64-bit Python interpreter, but then we run into the same problem that 32-bit NVDA is facing, although in the opposite direction.

As for using CPU instructions such as Advanced Vector Extensions (AVX), it comes down to the workload to be run by Python interpreter. I imagine C++ components could benefit from that if Visual C++ compiler can be told to use AVX instructions, but then it depends on how NVDA performs its tasks internally (AVX is mostly used for calculation intensive tasks such as scientific computing, which is not really the use case for a screen reader unless the screen reader algorithms do require working with many things that require up to 512 bits to store many components). Ultimately, since NVDA is a Python-based screen reader, it will come down to if Python interpreter can even take advantage of newer instructions for executing bytecode on the spot.

More importantly, consider that not all CPU’s support AVX and newer instructions. It took Microsoft until 2012 to require that processors support SSE2 (ten years after its appearance). Although most processors support SIMD (single instruction, multiple data) commands, only some high-end processors support AVX, AVX2, and parts of AVX-512, and tech press articles (and according to some sources, Linus Torvalds) argue that AVX instructions are power-hungry (understandable as millions of transistors are dedicated to vector instructions).

Therefore, the standpoint of NV Access (as communicated by Quentin and others) and lead contributors (including soon to be former ones like myself) is that we should target the vast majority of configurations that are in use at a given time (that’s one of the reasons for delaying end of support for Windows 7; due to a critical issue with Python 3.8 and later, that decision is uncertain at this time); at this time, this means supporting a processor with at least SSE2.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 6:07 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

Hi Joseph,

Would cythonizing NVDA improve this situatuation though? Or making a native 64 bit version to talk to 64 bit apps?  Also, isn't one reason for the CPU usage NVDA not utilizing the newer processor  instruction sets being developed by Intel/AMD?  Are there any new processor extention sets that would make NVDA more responsive if used?

On 5/16/2021 7:21 PM, Joseph Lee wrote:

Hi,

Enes’s claim is understandable, considering that:

  1. Speech synthesizers do require resources such as CPU to translate text into speech using whatever rules manufacturers and users define.
  2. Most NVDA components run on top of a Python interpreter. This means in order to perform screen reading operations, Python interpreter must be willing to process screen reader instructions on top of housekeeping tasks such as garbage collection. Some NVDA components are written in C++ for faster performance and communication with apps, but still Python is invoked for housekeeping operations.
  3. NVDA must talk to many “people” i.e. API’s and apps at once. Although folks spent years optimizing accessibility API’s and communication between apps (in computer science, this whole thing is termed “inter-process communication”), communication with API’s and apps is still an expensive step that involves processors switching between running app and system codes. If it takes some time to send and receive messages between two 32-bit programs, imagine how long it will take for a 32-bit app such as NVDA to get useful info out of 64-bit apps (this involves formatting bits in a way that allows these two programs to eventually communicate through an intermediary called WoW64).

 

One important clarification: let us not equate screen readers to text-to-speech engines (don’t confuse between the two). Although TTS does contribute to screen reading performance (namely rules used to translate text into waveforms or hardware signals), what’s more important in this overall context (screen readers running on a class of hardware) is how efficient screen reading instructions can get, keeping in mind that NVDA is running on top of an interpreter. There are limits as to how power efficient an algorithm can become, as the overall limit is optimizations from the host (in NVDA’s case, not only optimizations from Windows, but also ongoing optimization work done by Python core developers and third-party library/C extension authors). Also, although not addressed here, you can’t just state that making screen readers multi-core aware will bring improvements – it depends on years of effort spent on parallelization and optimizing the screen reader to not only take advantage of at least two cores at once, but also to make itself more efficient when running as a multi-core aware application (power draw from a single core versus multiple cores does play a part here); I have learned the hard way that you will face more unexpected bugs when you make a program run with multiple threads and hardware cores.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 4:58 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:

Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 

 

As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.

 

I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.

 

And I haven’t seen complaints about the performance of NVDA from people using tablets. 

 

Gene

-----Original Message-----

Sent: Sunday, May 16, 2021 7:20 AM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:

Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

Gene
 

I don’t know what the usage would be on a typical consumer machine these days but my main ;point is that many people seem to think that when you are buying a computer, you need to worry about NVDA.  While I can’t talk about bottom of the line machines, and I suspect it would work well with Eloquence or E-speak or other undemanding synthesizers, I don’t thinkk most people would have to worry. 
 
You may have helped substantiate what I’m saying but what I’d be curious to know is how much usage is shown for a current or recent machine, let’s say a laptop or desktop in the five or six hundred dollar range.
 
Also, if you use demanding voices, I wonder how responsiveness changes as the speed of the computer increases.
 
Gene

-----Original Message-----
Sent: Sunday, May 16, 2021 8:15 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

Hi Gene,

I have a top of the line mobile processor in this notebook. An AMD R7 Ryzen 4800H, with 8 cores, 16 threads, that can boost to 4.4 GHZ on all 8 cores. On this machine, NVDA CPU usage never exceeds 1% mostly.

On 5/16/2021 7:23 PM, Gene wrote:
Not at that kind of useage.  I don’t knoww at what point lag might result, perhaps eighty percent, certainly at one-hundred but six to ten percent poses no problem.
 
Also, the figure I’m giving is for my computer, about eleven years old and at that time, a moderately powered machine.  I have no idea what the figure would be for modern machines.  Perhaps in moderately powered machines now, the usage would be perhaps three percent. 
 
Gene
Sent: Sunday, May 16, 2021 7:15 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

What you don't realize though is that causes CPU lag,  and throttling, which lags synthesizers when combined with other applications, such as browsers. Even antiviruses don't use that much processing power with realtime protection. This is very much so with synthesizers like eloquence. That is why it makes sense to bvuy the most powerful specs that budget allows to prevent this from occuring.

On 5/16/2021 7:03 PM, Gene wrote:
Six to ten percent out of one-hundred?  Hardly.
 
Gene
-----Orignal Message-----
Sent: Sunday, May 16, 2021 6:57 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:
Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 
 
As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.
 
I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.
 
And I haven’t seen complaints about the performance of NVDA from people using tablets. 
 
Gene
-----Original Message-----
Sent: Sunday, May 16, 2021 7:20 AM
Subject: Re: [nvda] NVDA running on a budget laptop
 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:
Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

enes sarıbaş
 

Hi Joseph,

The 32 bit  problem would be less though I would think. As time has gone on, more and more apps are pure 64 bit, Microsoft is also dropping support for 32 bit windows as well. Most recently, Firefox, Thunderbird, and Adobe Reader have gone pure 64 bit.

On 5/16/2021 8:25 PM, Joseph Lee wrote:

Hi,

Cython might help to make NVDA code compile under a C++ compiler such as Microsoft Visual C++ compiler, but then it will raise questions such as dealing with C extensions and how to make NVDA’s own (Python) code more efficient. Transforming NVDA into a pure 64-bit program will require that we use 64-bit Python interpreter, but then we run into the same problem that 32-bit NVDA is facing, although in the opposite direction.

As for using CPU instructions such as Advanced Vector Extensions (AVX), it comes down to the workload to be run by Python interpreter. I imagine C++ components could benefit from that if Visual C++ compiler can be told to use AVX instructions, but then it depends on how NVDA performs its tasks internally (AVX is mostly used for calculation intensive tasks such as scientific computing, which is not really the use case for a screen reader unless the screen reader algorithms do require working with many things that require up to 512 bits to store many components). Ultimately, since NVDA is a Python-based screen reader, it will come down to if Python interpreter can even take advantage of newer instructions for executing bytecode on the spot.

More importantly, consider that not all CPU’s support AVX and newer instructions. It took Microsoft until 2012 to require that processors support SSE2 (ten years after its appearance). Although most processors support SIMD (single instruction, multiple data) commands, only some high-end processors support AVX, AVX2, and parts of AVX-512, and tech press articles (and according to some sources, Linus Torvalds) argue that AVX instructions are power-hungry (understandable as millions of transistors are dedicated to vector instructions).

Therefore, the standpoint of NV Access (as communicated by Quentin and others) and lead contributors (including soon to be former ones like myself) is that we should target the vast majority of configurations that are in use at a given time (that’s one of the reasons for delaying end of support for Windows 7; due to a critical issue with Python 3.8 and later, that decision is uncertain at this time); at this time, this means supporting a processor with at least SSE2.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 6:07 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

Hi Joseph,

Would cythonizing NVDA improve this situatuation though? Or making a native 64 bit version to talk to 64 bit apps?  Also, isn't one reason for the CPU usage NVDA not utilizing the newer processor  instruction sets being developed by Intel/AMD?  Are there any new processor extention sets that would make NVDA more responsive if used?

On 5/16/2021 7:21 PM, Joseph Lee wrote:

Hi,

Enes’s claim is understandable, considering that:

  1. Speech synthesizers do require resources such as CPU to translate text into speech using whatever rules manufacturers and users define.
  2. Most NVDA components run on top of a Python interpreter. This means in order to perform screen reading operations, Python interpreter must be willing to process screen reader instructions on top of housekeeping tasks such as garbage collection. Some NVDA components are written in C++ for faster performance and communication with apps, but still Python is invoked for housekeeping operations.
  3. NVDA must talk to many “people” i.e. API’s and apps at once. Although folks spent years optimizing accessibility API’s and communication between apps (in computer science, this whole thing is termed “inter-process communication”), communication with API’s and apps is still an expensive step that involves processors switching between running app and system codes. If it takes some time to send and receive messages between two 32-bit programs, imagine how long it will take for a 32-bit app such as NVDA to get useful info out of 64-bit apps (this involves formatting bits in a way that allows these two programs to eventually communicate through an intermediary called WoW64).

 

One important clarification: let us not equate screen readers to text-to-speech engines (don’t confuse between the two). Although TTS does contribute to screen reading performance (namely rules used to translate text into waveforms or hardware signals), what’s more important in this overall context (screen readers running on a class of hardware) is how efficient screen reading instructions can get, keeping in mind that NVDA is running on top of an interpreter. There are limits as to how power efficient an algorithm can become, as the overall limit is optimizations from the host (in NVDA’s case, not only optimizations from Windows, but also ongoing optimization work done by Python core developers and third-party library/C extension authors). Also, although not addressed here, you can’t just state that making screen readers multi-core aware will bring improvements – it depends on years of effort spent on parallelization and optimizing the screen reader to not only take advantage of at least two cores at once, but also to make itself more efficient when running as a multi-core aware application (power draw from a single core versus multiple cores does play a part here); I have learned the hard way that you will face more unexpected bugs when you make a program run with multiple threads and hardware cores.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 4:58 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:

Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 

 

As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.

 

I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.

 

And I haven’t seen complaints about the performance of NVDA from people using tablets. 

 

Gene

-----Original Message-----

Sent: Sunday, May 16, 2021 7:20 AM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:

Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

Gene
 

I’ll read your message more than once.  I think I have sort of an idea of what you are saying.  I should add that the usage on my machine would be lower than I said.  I don’t know how much, but I was reporting the total resource usage that I took immediately after I did things with NVDA.  But that total usage would include Via Voice.  I doubt that would make much difference but I’m stating that for accuracy in what I am reporting.
 
Gene

-----Original Message-----
From: Joseph Lee
Sent: Sunday, May 16, 2021 7:40 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

Hi,

As Enes may have pointed out in a later message, it would have been considered low usage as old as late 2000’s; not much these days:

  • In late 1990’s, CPU speed (clock frequency) was measured in hundreds of megahertz (MHz; hertz = cycles per second, so a CPU can theoretically run hundreds of millions of instructions per second; I’ll explain the actual hardware side of things in a second). This was the days of 32-bit single-core Pentiums.
  • By about 2005, Intel realized that Pentium 4 was power-hungry and wasn’t productive anymore. Folks may recall the “4 gigahertz barrier” – trying to switch transistors on and off up to 4 billion times a second caused processors to eat up power. Meanwhile, AMD convinced the jury (PC makers and enthusiasts) that their dual-core, low-power 64-bit processors were better than top of the line Intel chips.
  • By early 2010’s, computer manufacturers and operating system vendors (along with Intel, AMD, and ARM Holdings) proved that it was somewhat possible to lower the power requirements of processors by bundling many slower processors inside a single chip. This means you can’t simply state that 6 to 10 percent CPU usage is low usage anymore – usage now depends on how many processor cores a program is using, as well as how many instructions from a given program a processor can consume and execute at a given time (in theory, we are talking about up to 16 processor hardware, each performing up to 4 billion or more instructions per second, or in hardware terms, turning on and off many billions of switches in one second).

In early 2000’s, what was considered low CPU usage was sort of equivalent to hundreds of thousands of instructions per second (at the hardware level, this theoretical limit was not really achieved because many hardware-level operations are performed to run just one CPU instruction). In 2021, low CPU usage usually translates to millions of instructions per second that are performed by at least one processor core (and even more hardware operations such as connecting circuits, turning on transistors, applying voltage to parts of the CPU, turning off unused memory banks (CPU cache), checking power draw amongst cores and so on). At least this is the case for computer programs that can run on bare silicon; processor usage overhead increases when a program must go through an intermediary such as an interpreter, which in turn must follow rules set by an operating system, which in turn must be smart enough to regulate how long a program (operating system code itself, the interpreter, any program that runs on top of an interpreter, all these combined inside a single executable) can run at any given time.

Hope this helps.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of Gene
Sent: Sunday, May 16, 2021 5:11 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

I’ll rephrase my answer.  I’m not a tech but I have always considered such figures to be low usage.  I just ran MP3 Direct Cut and it uses about two percent of computer resources and its just a small simple Mp3 recording program. 

 

We’ll see what those with far more technical knowledge than I have say but I doubt they will say that that is particularly high useage.

 

Gene

-----Original Message-----

From: Gene

Sent: Sunday, May 16, 2021 7:03 PM

Subject: Re: [nvda] NVDA running on a budget laptop

 

Six to ten percent out of one-hundred?  Hardly.

 

Gene

-----Orignal Message-----

Sent: Sunday, May 16, 2021 6:57 PM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:

Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 

 

As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.

 

I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.

 

And I haven’t seen complaints about the performance of NVDA from people using tablets. 

 

Gene

-----Original Message-----

Sent: Sunday, May 16, 2021 7:20 AM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:

Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

tim
 

Your right its not and it is not continual usage like with screen records like OBS or video editing.

Screen readers uses small amounts and ram is mannaged a lot better not.

I maxed my ram out and now run everything in ram. System just runs even faster now.

On 5/16/2021 8:10 PM, Gene wrote:
I’ll rephrase my answer.  I’m not a tech but I have always considered such figures to be low usage.  I just ran MP3 Direct Cut and it uses about two percent of computer resources and its just a small simple Mp3 recording program. 
 
We’ll see what those with far more technical knowledge than I have say but I doubt they will say that that is particularly high useage.
 
Gene
-----Original Message-----
From: Gene
Sent: Sunday, May 16, 2021 7:03 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 
Six to ten percent out of one-hundred?  Hardly.
 
Gene
-----Orignal Message-----
Sent: Sunday, May 16, 2021 6:57 PM
Subject: Re: [nvda] NVDA running on a budget laptop
 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:
Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 
 
As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.
 
I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.
 
And I haven’t seen complaints about the performance of NVDA from people using tablets. 
 
Gene
-----Original Message-----
Sent: Sunday, May 16, 2021 7:20 AM
Subject: Re: [nvda] NVDA running on a budget laptop
 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:
Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 


Re: NVDA running on a budget laptop

tim
 

you can find computers with 1tb ssd all day on amizon and newegg.

Even if you have to get a system with spinning drives cheep. You can get a 1tb ssd drive for under $70 at both of the above places. And just install it there not that hard.

On 5/16/2021 4:25 PM, Shaun Everiss wrote:
To be honest, its near to impossible to get a newer unit with a duel drive system.

I got one of the last ones before covid and to be honest I am unsure where I will go when I eventually need to.


The most you can get is 512gb ssd  storage.

Thats fine but what I really want to avoid is carrying round hdds all over the show.

I can pay for a sync storage online but when on the go the net will not be fully available and anyway, I like physical cds, storage drives as well as digital.

8gb ram is really minimum.

It all depends on what you want in power these days.

Unless you really need to do things like virtual machines or testing of development stuff, almost anything will do.

Look for something at least 6th generation intel to about 7th gen, be aware that 8th gen and others may have speed issues due to all the security put in to get round specter and other stuff the 8tyh generation was most effected.

I don't see any reason to get any higher than the ryzen 5 2nd gen I allready have though.

I have family that can't afford much running of a high grade 3rd gen and that suits them just fine for what they do.

The fact is with covid and shortages all these twin systems just stopped overnight even gaming units.

Lenovo may be your best bet though they have changed their customisation pages so who knows.

I'd stop short of dell but yeah.

The other way you may want to go is run a budget something for the basics and something else for others.

I know for 1400 bucks I can get a desktop with the latest everything and more storage than I'd know what to do with but still.

You may want to search ebay or other sites and get something will work.

I am currently working with a client desperate to update their systems and needing spaciffic needs, a 1tb hard drive or ssd being one of them.

They don't have that much cash and its hard to get 1tb ssds anyway in units and duel drives on modern up to date systems are well just forget it.

On the other hand, if you ever see a hp db1018 or 1049 unit about with the duel drives, that may be your ticket out.

There is the 17qa which did have at one stage a hard drive.

Finally there is the pav power box with 7th gen processer but all of these are old.

Some places may still have pavilian ce units but these may be a dieing breed.

Some units may have duel drive support but who knows.

Point is I suspect with all the chip shortages laptops= business only.

Yeah budget can be a problem.

If you really can afford 300-400 bucks, then you may be getting better value buying a used or x lease product  which has the ability.

Unsure what I will do when I really need another box.

For all its worth the system I currently have now has come on holiday with me and several other things and is lite and easy to control.

Bar a few issues early on its going strongly.

If you are going for something really old though, an a10 or 12 amd is fine but I'd try to get something 2017 or 2018 make if its amd because in 2016 the hd class video cards got stopped due to newer chips support.

I can only wish you luck in this.





On 17/05/2021 6:52 am, Bob Cavanaugh wrote:
I've had a number of techs look at my laptop over the last couple
months. They've all recommended I get some combination of a new
machine with a faster processor, more ram, or an SSD, but I've had
very little to complain about with this machine as currently
configured. Sometimes, there are instances where I open a large number
of windows in my web browser, and the number of windows I can open was
significantly smaller on this machine with Windows 10, 4 GB of ram and
an Intel I3 than on my previous laptop with an Intel I5 and Windows 7
with the same specs. The number of windows I can open at a time has
decreased even more since I switched from IE to Firefox.

On 5/16/21, Brian Vogel <britechguy@gmail.com> wrote:
On Sun, May 16, 2021 at 02:21 AM, Jeff Samco wrote:

to be  confident the laptop won't frustrate me.
-
And right there is the key point.  One can definitely use a "less finely
specced" machine than one's home workhorse when all it's meant for is light
duty emailing, web browsing, word processing, and even streaming, but even
when something with really minimum spec works (and I have an ancient Gateway
with 4GB and old Intel processor that does), "works" is often not enough to
make one feel OK while using it.

You still want the speed and responsiveness you're used to on your regular
machine, but just for lighter loads.  With Windows 10, it's almost certain
that you will not have that experience with under 8GB of RAM, and you
absolutely won't if you've got less than that with a really old processor
and a conventional HDD.   SSDs can and do gain you a lot of zippy-ness when
it comes to tasks that are disk I/O intensive.  Boot times shrink
incredibly.  But for stuff that's not disk I/O intensive, and the intended
uses aren't, your throttling point is typically from the combination of
processing power and RAM, which are intimately related in that a poor
processor with lots of RAM could, in theory, work better than a mediocre (or
good) processor with insufficient RAM.

Windows 10, and all modern OSes, are now built to absolutely maximize their
exploitation of available RAM to make faster performance (from the user
perspective, anyway) happen.

I've never owned anything near to a "flagship machine" myself, because I
don't have the need for that kind of processing power or speed.  I have, at
times past, had to get "the cheapest thing possible" so I've lived with
these long term, and it wasn't fun.

Often even small increases in price on the lower end of the market get you
some pretty substantial performance boosts.

In the end, I'm really not telling anyone what they should buy, because I
can't know that.  Only the individual making the choice knows what is
"minimally acceptable" to them.  But I am trying to identify factors to
consider.  You have to make the choice in the end, and what's right for me,
or any random someone else, may not be right for you.

--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042

Always remember others may hate you but those who hate you don't win unless
you hate them.  And then you destroy yourself.

~ Richard M. Nixon








.



Re: NVDA running on a budget laptop

 

Hi,

Cython might help to make NVDA code compile under a C++ compiler such as Microsoft Visual C++ compiler, but then it will raise questions such as dealing with C extensions and how to make NVDA’s own (Python) code more efficient. Transforming NVDA into a pure 64-bit program will require that we use 64-bit Python interpreter, but then we run into the same problem that 32-bit NVDA is facing, although in the opposite direction.

As for using CPU instructions such as Advanced Vector Extensions (AVX), it comes down to the workload to be run by Python interpreter. I imagine C++ components could benefit from that if Visual C++ compiler can be told to use AVX instructions, but then it depends on how NVDA performs its tasks internally (AVX is mostly used for calculation intensive tasks such as scientific computing, which is not really the use case for a screen reader unless the screen reader algorithms do require working with many things that require up to 512 bits to store many components). Ultimately, since NVDA is a Python-based screen reader, it will come down to if Python interpreter can even take advantage of newer instructions for executing bytecode on the spot.

More importantly, consider that not all CPU’s support AVX and newer instructions. It took Microsoft until 2012 to require that processors support SSE2 (ten years after its appearance). Although most processors support SIMD (single instruction, multiple data) commands, only some high-end processors support AVX, AVX2, and parts of AVX-512, and tech press articles (and according to some sources, Linus Torvalds) argue that AVX instructions are power-hungry (understandable as millions of transistors are dedicated to vector instructions).

Therefore, the standpoint of NV Access (as communicated by Quentin and others) and lead contributors (including soon to be former ones like myself) is that we should target the vast majority of configurations that are in use at a given time (that’s one of the reasons for delaying end of support for Windows 7; due to a critical issue with Python 3.8 and later, that decision is uncertain at this time); at this time, this means supporting a processor with at least SSE2.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 6:07 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

Hi Joseph,

Would cythonizing NVDA improve this situatuation though? Or making a native 64 bit version to talk to 64 bit apps?  Also, isn't one reason for the CPU usage NVDA not utilizing the newer processor  instruction sets being developed by Intel/AMD?  Are there any new processor extention sets that would make NVDA more responsive if used?

On 5/16/2021 7:21 PM, Joseph Lee wrote:

Hi,

Enes’s claim is understandable, considering that:

  1. Speech synthesizers do require resources such as CPU to translate text into speech using whatever rules manufacturers and users define.
  2. Most NVDA components run on top of a Python interpreter. This means in order to perform screen reading operations, Python interpreter must be willing to process screen reader instructions on top of housekeeping tasks such as garbage collection. Some NVDA components are written in C++ for faster performance and communication with apps, but still Python is invoked for housekeeping operations.
  3. NVDA must talk to many “people” i.e. API’s and apps at once. Although folks spent years optimizing accessibility API’s and communication between apps (in computer science, this whole thing is termed “inter-process communication”), communication with API’s and apps is still an expensive step that involves processors switching between running app and system codes. If it takes some time to send and receive messages between two 32-bit programs, imagine how long it will take for a 32-bit app such as NVDA to get useful info out of 64-bit apps (this involves formatting bits in a way that allows these two programs to eventually communicate through an intermediary called WoW64).

 

One important clarification: let us not equate screen readers to text-to-speech engines (don’t confuse between the two). Although TTS does contribute to screen reading performance (namely rules used to translate text into waveforms or hardware signals), what’s more important in this overall context (screen readers running on a class of hardware) is how efficient screen reading instructions can get, keeping in mind that NVDA is running on top of an interpreter. There are limits as to how power efficient an algorithm can become, as the overall limit is optimizations from the host (in NVDA’s case, not only optimizations from Windows, but also ongoing optimization work done by Python core developers and third-party library/C extension authors). Also, although not addressed here, you can’t just state that making screen readers multi-core aware will bring improvements – it depends on years of effort spent on parallelization and optimizing the screen reader to not only take advantage of at least two cores at once, but also to make itself more efficient when running as a multi-core aware application (power draw from a single core versus multiple cores does play a part here); I have learned the hard way that you will face more unexpected bugs when you make a program run with multiple threads and hardware cores.

Cheers,

Joseph

 

From: nvda@nvda.groups.io <nvda@nvda.groups.io> On Behalf Of enes saribas
Sent: Sunday, May 16, 2021 4:58 PM
To: nvda@nvda.groups.io
Subject: Re: [nvda] NVDA running on a budget laptop

 

I'm sorry. 6-10% CPU power is alot of system resources.

On 5/16/2021 11:50 AM, Gene wrote:

Nothing I’ve seen convinces me that NVDA itself uses a lot of computing power.  nor have I seen this with screen-readers in general to the small extent I’ve checked.  Its using the newer synthesizers that uses a lot of computing ;power.  If you want to use the newer voices, I can’t comment on the minimum specifications to get good performance but in the old days, I had machines that today would be laughably underpowered, running Windows 95 and Windows 3.1 and Via Voice, very similar to Eloquence, ran well.  This was in a 166MHZ, not GHZ, Pentium machine and in an even older and less powerful machine running Windows 3.1. 

 

As for NVDA using a lot of computing power, if I monitor use when I’m typing text with carachter echo on in the Windows Task manager, I get low numbers.  I just checked and while moving up and down the list in task manager, then pressing f5 to refresh the screen, I get a 6 percent CPU reading.  When typing in this e-mail message, alt tabbing immediately to the task manager and refreshing the screen, I get a 10 percent usage reading.

 

I’m not saying there won’t be variations, but those figures are close to what I generally get when I test doing these things.

 

And I haven’t seen complaints about the performance of NVDA from people using tablets. 

 

Gene

-----Original Message-----

Sent: Sunday, May 16, 2021 7:20 AM

Subject: Re: [nvda] NVDA running on a budget laptop

 

I would do a minimum of 8gb of ram, and a current gen i5/r5. That is as low as you should go. Even with those specs, NVDA is heavy on CPU usage.

On 5/15/2021 7:36 PM, Brian Vogel wrote:

Personally, I would not even consider running Windows 10 with less than 8 GB of RAM.  Nor would I consider a Celeron processor, for anything, these days.

I'd invest a bit more for additional memory and a better processor.  You might also consider a refurbished business-class laptop, which can be had at very reasonable prices (or at least could prior to the pandemic - everything's getting more expensive as supply is constrained).
--

Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042  

Always remember others may hate you but those who hate you don't win unless you hate them.  And then you destroy yourself.

       ~ Richard M. Nixon

 

3701 - 3720 of 87641